Anthrax SET Protein: A Potential Virulence Determinant that Epigenetically Represses NF-κB Activation in Infected Macrophages

Toxins play a major role in the pathogenesis of Bacillus anthracis by subverting the host defenses. However, besides toxins, B. anthracis expresses effector proteins, whose role in pathogenesis are yet to be investigated. Here we present that suppressor-of-variegation, enhancer-of-zeste, trithorax protein from B. anthracis (BaSET) methylates human histone H1, resulting in repression of NF-κB functions. Notably, BaSET is secreted and undergoes nuclear translocation to enhance H1 methylation in B. anthracis-infected macrophages. Compared with wild type Sterne, delayed growth kinetics and altered septum formation were observed in the BaSET knock-out (BaΔSET) bacilli. Uncontrolled BaSET expression during complementation of the BaSET gene in BaΔSET partially restored growth during stationary phase but resulted in substantially shorter bacilli throughout the growth cycle. Importantly, in contrast to Sterne, the BaΔSET B. anthracis is avirulent in a lethal murine bacteremia model of infection. Collectively, BaSET is required for repression of host transcription as well as proper B. anthracis growth, making it a potentially unique virulence determinant.


INTRODUCTION
Human pathogens express an extensive repertoire of effectors that modulate the host's transcriptional machinery to overcome the immune response (1). Strategically, most effectors aim to stabilize the NF-κB/IκB (Nuclear factorkappa B/Inhibitor of kappa B) complex for precluding nuclear localization and transcriptional activation of NF-κB. Yersinia enterocolitica has a versatile effector, YopJ, a cysteine protease that inhibits ubiquitination of IκB as well as TRAF6, and acety l ates IKKα, wh i ch p r ev e nt IκB phosphorylation causing cytoplasmic retention thus avoiding nuclear translocation of NF-B to achieve its transcription function (2)(3)(4). Recent studies show that serine/threonine acetylation of TGFβ-activated kinase by YopJ inhibits innate immune signaling and activates caspase-1 to induce intestinal barrier dysfunction (5)(6). In Shigella flexneri, OspG is a serine/threonine kinase that inhibits the host's ubiquitin conjugating enzymes, Ubc5 and Ubc7, thereby, blocking ubiquitination-mediated IκB degradation, and subsequent, NF-B activation (7). Further, OspF in epithelial cells during shigella infection inhibits histone H3 phosphorylation that precludes the recruitment of NF-B to IL-8 promoter (8). Among viruses, the HIV trans-activator Tat protein recruits multiple co-activators CBP and PCAF, which direct acetylation-mediated chromatin remodeling that activates HIV transcription and replication (9)(10). Similarly, HBx protein of hepatitis B virus interacts and stimulates DNA helicase subunits of TFIIH (11). Thus, it is clear that human pathogens mainly target signaling pathways that alter host gene expression to promote their growth and replication. However, no effector has yet been characterized that directly modifies human chromatin to inhibit the inflammatory responses.
The etiologic agent of anthrax, B. anthracis is the most notorious pathogen within the Bacillus genus. Like all Bacillus species, B. anthracis is a Gram-positive spore-forming organism found in soil environments (12)(13). Unlike most other Bacillus species, respiratory, gastrointestinal, or cutaneous entry of B. anthracis spores into mammals can result in a rapid systemic infection and death (14)(15). A growing number of studies indicate that B. anthracis uses its toxins to down-regulate the host innate response. These toxins are comprised of three individual components including protective antigen (PA), lethal factor (LF) and edema factor (EF). PA interacts with EF to form edema toxin (ET), and LF to form lethal toxin (LT). ET functions as a calmodulin-dependent adenylate cyclase resulting 3 in the elevation of intracellular cAMP levels and LT is a zinc-dependent metalloproteinase that inactivates mitogen-activated protein kinases (MAPKKs or MEKs) (16)(17)(18)(19)(20). Recently, it has been shown that ClpX, the regulatory ATPase subunit of the ClpXP protease is linked to degradation of cathelicidin antimicrobial peptides, a major effector of the host innate response (21). However, given the severe outcome of anthrax disease, there is a need to identify and elucidate the biological functions of new effectors, which contribute to B. anthracis pathogenesis. Rationally, many studies have focused on dissecting the proteome of B. anthracis based on their immunogenicity to identify new vaccine candidates (22). Though this strategy is valid, it could potentially overlook proteins that help in evading the immune system, which could play a crucial role in B. anthracis pathogenecity.
Dynamic control of histone methylation is fundamental to epigenetic regulation of gene expression during development, X-inactivation, stem-cell pluripotency, cancer, and inflammation (23)(24). Published studies demonstrate that a battery of transcriptional co-factors in eukaryotes possess an evolutionarily conserved protein module, Suppressor of variegation-Enhancer of zeste-Trithorax (SET) domain, which methylates lysine residues on chromatin-associated proteins including histones and transcription factors (25)(26)(27). However, of the large family of SET domains, a small subset is encoded by human microbial pathogens, including B. anthracis, B. cereus, and Chlamydia trachomatis, (28)(29)(30). In eukaryotes, several lines of evidence support that site-specific methylation of histone proteins by a SET domain has potential to direct gene activation or silencing (31)(32). The presence of a SET protein in B. anthracis (BaSET), which lacks chromatin, suggests its role in altering the host's chromatin upon infection. Moreover, the role of BaSET in the B. anthracis life cycle also remains to be investigated. To address these questions, we have characterized the function of BaSET to understand its role in augmenting B. anthracis survival in an infected host.

EXPERIMENTAL PROCEDURES Phylogenetic analysis of SET proteins from human pathogens
The non-redundant database of protein sequences (nr) was searched with the BLASTp program (33) (http://blast.ncbi.nlm.nih.gov/Blast.cgi) using the sequences of Leishmania major (gi 157876484), Giardia lamblia (gi 29250071), Chlamydia trachomatis (gi 166154954), B. anthracis (gi 229603772), and B. cereus (gi 52140458) as queries. Multiple sequence alignment and the associated phylogenetic tree of the protein SETdomains were computed with the software Tcoffee (34)(35) . The colors used in the sequence alignments have been based on the Clustal coloring scheme (http://ekhidna.biocenter.helsinki.fi/pfam2/clustal_ colours). Homology models of SET proteins from Leishmania major and B. anthracis were built with MODELLER version 9.10 (36) using Homo sapiens histone H3K4 methyltransferase (PDB code 2W5Y) and the Rubisco LSMT SET-domain proteins (PDB code 2H21) as templates (37)(38) .

Cloning, purification and in vitro histone methylation assay
The full length BaSET gene was synthesized and cloned into pGEX4T-3 (GE health care) to express GST-BaSET, which was affinity purified using glutathione beads. The GST tag was then cleaved using thrombin at 4°C. Later, the cleaved-GST was removed upon reloading the glutathione column. To perform the histone methylation assay, the BaSET was subjected to buffer exchange (39). Flag-tagged BaSET (Flag-BaSET) was cloned into pcDNA3.1 (Invitrogen) that constitutively expresses Flag-BaSET in transient transfection using Fugene 6 (Roche) as well as into tetracycline-inducible pcDNA4-TO vector (Invitrogen). Based on the manufacturer's instructions, HeLa Trex cells (Invitrogen) were transfected with pcDNA4TO-Flag-BaSET and stabilized with blasticidin and zeocin; later, the cells were induced with tetracycline.

4
RAW 264.7 (mouse macrophage) cell lines were purchased from Invitrogen, ATCC and System Biosciences. These cells were grown in Dulbecco's minimum essential medium (DMEM) or RPMI supplemented with 5% (v/v) penicillinstreptomycin solution (Mediatech) and 10% FBS (USA Scientific) and incubated in a 5% CO 2 atmosphere at 37°C. Cells were sub-cultured with 0.25% (w/v) trypsin in the presence of 0.02% (w/v) EDTA. For the luciferase assays to evaluate the transcriptional effects of BaSET on the activation of Gal-UAS promoter by Gal4-ERα and Gal4-CBP, the 293T cells were stabilized with hygromycin. Similarly, to study the effect on NF-κB target genes, HEK293T cells were stabilized with IL-6, IL-8 and VCAM promoters using hygromycin. The transfection efficiency of Flag-BaSET, Gal4-ERα and Gal4-CBP were monitored by their co-transfection with renilla luciferase vector. Later, firefly lucifearse activity was normalized with luciferase values from renilla. Further, HEK293T cells were also transduced with lentivirus containing NF-κB response element cloned in tandem with a Green fluorescence protein (GFP) and luciferase reporter genes using the Transdux reagent (System Biosciences). To enhance the homogeneity of transduced 293T cells for developing an effective assay, the cells were treated with TNFα and sorted with FACS Star. Later, these stable 293T NF-κB_RE cells were treated with TNFα (25ng/ml), IL-6, IL-8, and IL-1 for 24 h and then the luciferase activity was determined in presence and absence of normalized Flag-BaSET with renilla luciferase values (39)(40).
Total RNA was isolated using an RNeasy Plus Mini Kit (Qiagen) from HeLa and RAW 264.7 cells transfected with Flag-BaSET or infected with Sterne and BaΔSET bacilli, which were also treated with TNFα. The total RNA isolated was non-degraded and free from protein and DNA contamination. Complementary DNA (cDNA) synthesis was performed using affinity script QPCR cDNA synthesis kit (Agilent). The resulting cDNA served as a template for comparative quantitative Real Time PCR (Brilliant II faster SYBR Green QPCR master mix -Agilent) with validated real-time gene specific primers for TNFα (Gen e lin k), C-Fos, C-Jun, IL-4, IL-5 and VEGF (MC Lab). PCR data were analyzed and Ct values were normalized with housekeeping gene (beta-actin) and calculated using the method 2 - . Results were expr essed in the form of fold expression (39)(40).

Antibody development, immunoprecipitation (IP) and immunoblotting (IB)
The polyclonal antibody to the purified BaSET protein was produced in rabbits (Alpha Diagnostic). The anti-BaSET antibody strongly recognized submicromolar quantities of purified BaSET protein (1:10,000) in an immunoblot and ELISA (Supplemental Figures 3a and b,). The antibody did not cross-react with other cellular proteins within HeLa cells, E. coli, or the BaSET Sterne knockout (Supplemental Figs. S3C, S3D and S4A) but reacted specifically with the BaSET protein when complemented in BaΔSET strain (Fig. 5A). The immunoprecipitation (IP) and immunoblots (IB) were performed as described previously (10,39,41). Briefly, the Flag-BaSET was transfected into HeLa Trex and HEK293T cells. After 24 h, cells were harvested and incubated with lysis buffer on ice for 2 h with occasional vortexing every 25 min for 15 s and then centrifuged at 14,000 rpm for 5 min at 4°C. After protein estimation, cell extracts were mixed with anti-Flag M2-coupled Agarose beads or anti-Flag Agarose beads (Sigma), respectively, overnight. The overexpressed Flag-BaSET were immunoprecipitated overnight. Beads were washed extensively in IP lysis buffer (20 mM Tris pH 8.0, 150 mM NaCl, 1 mM EDTA, 1 mM EGTA, 1% Triton X-100, and 50 mM NaF) and separated by 4-20% gradient SDS-PAGE. For direct IB, 30-50 μg lysates and/or nuclear extracts were subjected to SDS-PAGE, transferred onto nitrocellulose membranes, blocked with 5% milk or BSA/TBST (20 mM Tris, 150 mM NaCl, and 0.05% Tween-20), and blotted with the recommended dilution of primary antibody overnight. After washing with TBST, HRP-labeled secondary antibodies (anti-mouse or anti-rabbit from Bethyl Laboratories) were added for 60 min. at room temperature. The blots were then washed with TBST and subjected to autoradiography after development using the ECL kit (Pierce).

Extracellular secretion of BaSET protein
B. anthracis Sterne was electroporated with a Green fluorescence protein (GFP)expressing plasmid in order to be used as a control for cell lysis and release of cellular contents into media. A 50 mL culture of Sterne and Sterne/GFP was inoculated (1:100) and grown at 30˚C and 200 rpm in BHI and BHI/Kan/Spec respectively. Cells were harvested at 8 and 24 h post inoculation by centrifugation at 4,000 rpm, 4C for 20 min. The supernatant was filtered through a 0.22 µm sterile filter. Samples were then concentrated down from 50 mL to ~2 mL using a 50 mL 3,000 kDa Millipore untracentrifugation tube. Cell pellets were then re-suspended and lysed using PlyG lysin. Cell lysate was sterile filtered (0.22 µm). Supernatants and cell lysate samples were frozen at -80˚C until immunoblots were performed.

Time course of macrophage infection of Sterne
The Raw 264.7 mouse macrophage cell line was grown in MEM media with 10% fetal bovine serum (FBS) until reaching a confluency of approximately 90% (42)(43). The media was aspirated and cells were washed thrice with MEM media without FBS. Cells were inoculated with Sterne spores at a 1:1 ratio in MEM media with light centrifugation (200 rpm, 10 min.) and incubated at 37˚C and 5% CO 2 for one hour. After the one-hour infection period, the media was removed and the cells were washed with MEM media to remove excess spores that had not been phagocytozed. MEM media containing 10% FBS and 20 µg/mL of gentamycin was added to the plates which were placed back at 37 C under 5% CO 2 . Cells were harvested at 0, 0.5, 1, 2, 4, 6, 8, and 12 h post-infection by scrapping off macrophage and centrifugation (1000 rpm at 4˚C, for 20 min.). Cell pellets were frozen at -80˚C followed by immunoblot with anti panKme3 antibody (Immunochem, Canada) as well as direct immunoblot for presence of BaSET in the following post-infected 0, 1, 2, and 4 samples were performed.

Cellular effects of B. anthracis and BaΔSET on macrophages
The growth and viability of the macrophage cell line (Raw 264.7) was studied by BRDU incorporation (Calbiochem) and oligonucleosomal detection (Roche) using ELISA in 96-well plates. Cells (1X10 5 ) were seeded into each well 16 h before inoculation with spores. Spores from wild type Sterne and BaΔSET bacilli were then added to the wells containing the macrophage. However, before inoculation BRDU was added and cells were allowed to grow for 24 h. Subsequently, cells were fixed and permeabilized.
In the cell lysates, the anti-BRDU antibody detected the incorporated BRDU in the dividing cells, and anti-histone antibody detected the fragmented oligonucleosomes in the cells undergoing apoptosis. Finally, after secondary antibody incubation, ELISA was developed and to record the respective readings.

Growth analysis of B. anthracis and BaΔSET
All overnight cultures of B. anthracis were grown at 30˚C with aeration in BHI and, for the BaΔSET mutant, in BHI supplemented with 50 µg/mL kanamycin and 250 µg/mL spectinomycin. The OD 600 of each overnight culture was normalized and diluted 100-fold in 50 mL (BHI) in a 250 Erlenmeyer flasks. The flasks were incubated at 30C and shaken at 200 rpm during which time samples were removed every hour for 15 h and a final sample at 24 h for OD 600 analysis. The experiment was performed in duplicate.

Expression of BaSET during in vitro culture (growth) of B. anthracis
For analyzing BaSET expression in B. anthracis, under in vitro conditions, 50 mL cultures (BHI, Difco) were grown of Sterne. Cells were washed twice with 1X PBS and then resuspended in 1 mL of 1X PBS. To lyse the cells, 10 µL of PlyG lysin was added. Subsequently, the culture was incubated at 37˚C for 30 min. Then, the cells were centrifuged at 4,000 rpm, at 4˚C for 20 min. and the supernatant was frozen at -80˚C until IP and IB were performed similar to mammalian cells as described in the above section.

Light and transmission electron microscopy (TEM)
B. anthracis Sterne and BaΔSET 50 mL cultures (BHI in a 250 Erlenmeyer flask) were inoculated (1:100) from an overnight culture and grown at 37˚C with aeration (200 rpm) for periods up to 24 h. At indicated time points, culture samples were removed and examined by light microscopy using a Nikon eclipse E400 microscope (1000x). TEM images were taken for Sterne and BaΔSET at 15 h (5000x), when the greatest difference in morphology had been observed by light microscopy.

Creation of BaΔSET complement mutant (cBaΔSET)
The BaSET gene was cloned into a xylose-inducible plasmid (pWH1520) from B. megaterium (MoBiTec) and confirmed by sequence analysis and PCR and named (pSET1). To increase quantities of pSET1, dH5-α cells were transformed with pSET1. Purified pSET1 was then transformed into SCS110 (DAM-, DCM-) (Agilent Technologies) to have unmethlyated DNA for electroporation into B. anthracis. Unmethylated DNA was prepared by a Qiagen Midi-kit the day of the electroporation into the BaΔSET mutant. BaΔSET mutant cells were prepared for electroporation by inoculating 1:100 from an overnight culture into a 500 mL culture of BHI/Kan/Spec (1:100) and grown for 4 h at 30˚C. Cells were washed 7 times with filtered ddH 2 O and resuspended in electroporation buffer (2mM phosphate buffer pH 7.4, 10% sucrose and 15% glycerol). Electroporation was performed under the following conditions: 200Ω, 25µF, 2.5kV in a 0.4 cm electrode gap (Bio-Rad, cat#165-2088). Colonies were PCR confirmed using BaSET specific primers (GTACCCGGGATGATGAAGTAAAAACG and GTAGGATCCCTTCCTCCTTATCAAAC) and by 16sRNA to confirm it was B. anthracis. The complimented BaΔSET mutant was termed cBaΔSET.

Assessment of BaSET expression in cBaΔSET
To assess the expression of cBaΔSET, 50 mL cultures in 250 Erlenmeyer flasks were grown in BHI/Tet in the presence or absence of 0.3% xylose or glucose. Four samples of cBaΔSET were prepared 1) uninduced, 2) with glucose present to suppress expression of the BaSET gene and test for leaky expression, 3) xylose added at time 0 of inoculation. Samples were grown for 8 h at 30C at 200 rpm and were harvested and washed two times with 1X PBS buffer and were resuspended in 500 µL 1X PBS and lysed using PlyG lysin. Lysed cells were then centrifuged and the supernatant was filtered using a 0.22 µm filter (Millipore). Samples were kept frozen at -80˚C until immunoblot analysis.

Growth and morphological changes in cBaΔSET
To determine the effects of BaSET gene expression in cBaΔSET at different time points during growth of B. anthracis, we monitored the growth optically (OD 600 ) and morphologically by light microscopy. A 50 mL culture of BHI or BHI/Tet respectively for Sterne or the cBaΔSET was grown in a 250 mL Erlenmeyer flask at 30 C and 200 rpm for 24 h. Xylose was also present in the Sterne culture medium. For cBaΔSET, xylose was present in the culture at the time of inoculation (time 0). Time points for the growth curve were taken every hour until 15 h and a final time point at 24 h. For time points at 3, 5, and 15 h, light microscopy images were taken for purposes of comparison (1250x).

Comparison of cell lengths for Sterne and cBaΔSET mutant
Cell lengths were determined and an average length in µM was calculated. The lengths of 50 B. anthracis cells for Sterne, cBaΔSET induced at 0 h, and cBaΔSET induced at 4 h were tabulated by printing out images and measuring the length of the cells in mM. All images were taken at a magnification of 1250x using a Nikon Eclipse E400 microscope. This was done for time points 0, 3, 5, 15, and 24 h. An image of 10 µM segments was taken and printed in order to convert measured lengths in mM into µM. The conversion was 23.5 mM = 10 µM.

Sporulation of Sterne and BaΔSET
A 500 µL aliquot from Sterne or BaΔSET overnight cultures was plated onto Leighton-Doi sporulation medium (LD) (45); and LD supplemented with kanamycin and spectinomycin as above. The plates were incubated at 37C for two days in the dark. Cells were then recovered, suspended in 5 mL of ddH 2 O and vortexed for 1 7 min. to break up any remaining agar. The samples were then serially diluted in ddH 2 O and 20 µL aliquots were plated onto BHI plates. The serially diluted samples were also heated at 80C for 20 min. and 20 µL of each dilution were plated on BHI plates.
Plates were incubated at 30C overnight. This experiment was conducted in triplicate. CFU comparison was performed between heated and unheated samples to determine sporulation efficiency.

Spore germination of Sterne and BaΔSET
Spore germination was monitored by decrease in optical density at 595 nm and by spore resistance to heat treatment (44,(46)(47).
To measure the decrease of OD 595, spores were heat activated at 65˚C for 20 min. then re-suspended in 4˚C germination buffer to an OD 595 of 0.6. After the addition of germinants (50 mM L-alanine and 1 mM inosine), the OD 595 was monitored using either a Spectra Max M5 or Plus spectrophotometer (Molecular Dynamics) using 15 second intervals for 60 min. at 30˚C. Before and after each reading the plate was shaken for 3 seconds to prevent spores from settling. Spore germination was monitored by loss of heat resistance. Briefly, spores were diluted to ~4x10 5 spores/mL and were heat-activated by incubating at 65˚C for 20 min. Spores were then diluted 1:200 in germination buffer (10 mM Tris-HCl/NaCl 50 mM L-alanine, and 1 mM inosine). At each time point (2,5,10,15,20, and 30 min.), 100 µL aliquots were removed and incubated at 65˚C for 20 min. to kill any germinated spores and then 50µL aliquots were plated on BHI agar and incubated overnight at 30C. Additionally, a 50 µL aliquot at the zero time point was plated without heat treatment to get an initial titter for germinated spores.

Mouse infection: Sterne vs BaΔSET
For the inoculum, overnight cultures of B. anthracis Sterne and BaΔSET were gro w n as described above, diluted 1:100 in BHI and grown for 3 and 5 h, respectively. Cells were harvested, washed twice with 50 mL of sterile PBS (pH 7.4), and adjusted to a density of ~1x10 8 (OD 600 of 1.0). Infections were performed as previously described (48). Briefly, 4-5 week-old female C57BL/6 mice (5 per group) were infected (i.p) with ~1x10 8 bacilli (200 µL in PBS), and survival was monitored for 14 days. These experiments were repeated twice, and a third set of infected mice was euthanized 3 h post-infection. The heart, liver, spleen, and kidneys were excised from the third set, washed with 70% ethanol and sterile PBS (pH 7.4), homogenized in PBS and plated on BHI plates to determine bacterial viability. The colony forming unit per gram of tissue (CFU/gram) was calculated by the following equation (Colony forming unit)X(Dilution factor)/Grams of Tissue. It was assumed that the limit of detection was one CFU.

Phylogenetic analysis of SET proteins in B. anthracis and other human pathogens
To examine the functional and evolutionary relationships between the SET proteins in major human pathogens, including Leishmania major (gi 157876484), Giardia lamblia (gi 29250071), Chlamydia trachomatis (gi 166154954), B. anthracis (gi 229603772), and B. cereus (gi 52140458), a phylogenetic analysis and homology alignment were performed. Our results confirmed earlier data that the SET protein module, though wide-spread in mammals, has limited distribution in plants and human pathogens, including bacteria and tropical parasites (29). However, the combined analysis of homology alignment and phylogenetic distribution revealed that these modular domains could be broadly distinguished into two groups, namely Zn 2+binding and non-Zn 2+ -binding, based on the presence of the conserved post-SET domain that coordinates Zn 2+ ion (Supplemental Fig. S1A). Based on homology alignment, the Zn 2+ -binding SET domains have a classical SET domain signature motif NHSCxPN/NHSxxPN as well as a CxCxxxxC motif in the C-terminal region of the protein, which is known as the post-SET-domain (Fig. 1). These conserved cysteine residues accompanied by a fourth cysteine in the SETdomain signature motif coordinates a Zn 2+ ion. The Zn 2+ -binding post-SET domains are mostly present in mammals but also in human parasites including Leishmania major and Giardia lamblia (Fig. 1). In contrast, the majority of bacterial 8 members does not contain this cysteine-rich motif and hence does not bind Zn 2+ (Fig. 1). This is seen in Zn 2+ -independent SET proteins of chlamydia trachomatis, B. anthracis, and B. cereus. A detailed structural analysis of the homology models is described in the supplementary section (Supplemental Figs. S1B and S1C).

Gene regulatory functions of BaSET
To confirm the ability of BaSET to modulate transcription, in the three sections below, we performed several in vitro histone methylation assays. We also determined the effects of BaSETmediated methylation on promoter activation and on the expression of NF-κB target genes.

BaSET is a histone H1 methyltransferase
To investigate the biochemical functions of BaSET, the BaSET gene was cloned to be expressed as a glutathione-s-transferase (GST) fusion protein in E. coli (BL21). Subsequently, the GST-BaSET protein was affinity purified and thrombin digested to remove the GST-tag (Supplemental Figs. S2A-D). Using apomyoglobin as a calibrant, the purified BaSET was confirmed to be 16 kDa by mass spectrometry (Supplemental Figs. S2E and S2F), which exists as a dimer in solution based on analytical ultracentrifugation (Supplemental Fig. S2G).
Using the purified protein, we examined whether BaSET is enzymatically capable of methylating histone proteins. Towards this goal, histones H1, H2A, H2B, H3 and H4 were individually purified in a bacterial expression system to avoid any possibility of pre-existing methylated-lysine residues prior to being used as a substrate. A non-radioactive histone methyltransferase assay using S-adenosyl methionine (SAM) as a cofactor followed by immunoblotting with pan-tri-methylated (pan-Kme3) antibody showed that BaSET specifically tri-methylates histone H1 and not the other histones (H2A, H2B, H3 and H4) (Fig. 2A). The pan-Kme3 antibody is capable of detecting tri-methylated lysine moieties independent of its site and position. Notably, we obtained similar results, thereby, confirming histone H1 as the major target, when equimolar concentrations (1.0 µM) of histone proteins were subjected to methylation reaction by BaSET (Supplemental Fig. S2H). To test whether BaSET methylation activity is dependent upon concentration and time, we performed an in vitro methylation assay with increasing concentrations of BaSET as well as incubation time. The results show that by keeping concentrations of both histone H1 (500 nanogram) and SAM unchanged, but increasing the concentration of BaSET (15 µM to 110 µM) the level of histone H1 tri-methylation was enhanced and BaSET achieves maximum potential to methylate H1 by 30 min (Figs. 2B and  C). Finally, mass spectrometry on methylated H1 revealed that BaSET methylates 8 lysine residues (Supplemental Figs. S2I (i) and S2I (ii)). Amino acid composition revealed that thirty percent of the residues in histone H1 are lysines, which are all surface exposed.
Since BaSET methylates H1 protein, it was important to investigate whether it also methylates a chromatin template i.e. polynucleosome. Thus, we extracted chromatin from nuclei of 293T cells, digested with micrococcal nuclease (MNase) and subjected the polynucleosomes to an in vitro methylation reaction. Post-methylation, the reaction mixtures with or without BaSET were subjected to immunoprecipitation (IP) with anti-histone H1 antibody and immunoblotted with pan-Kme3 antibodies. The result shows that histone H1 within polynucleosomes was methylated by BaSET (Fig. 2D).
To test whether BaSET can methylate histone H1 on cellular chromatin, human embryonic kidney 293T (HEK293T) cells were transiently transfected with increasing amounts of the Flag-BaSET plasmid. After twenty-four hours, the transfected cells were harvested and nuclear extracts prepared. Subsequently, equal amounts of nuclear proteins were subjected immunoprecipitation with anti-histone H1 antibody followed by immunoblotting with pan-Kme2 and pan-Kme3 antibodies. Results revealed that dosedependent expression of BaSET led to enhancement of di-and tri-methylation levels on histone H1 (Fig. 2E).

A potential transcriptional repressor, BaSET inhibits NF-κB activation
Previous studies revealed that methylation of lysine residues on histone proteins in the vicinity of a promoter could cause gene activation or repression (25). Therefore, to determine whether the BaSET-mediated H1 methylationaffects promoter activity, we tested atleast four different luciferase-based reporter systems. All luciferase experiments were performed in stabilized 293T cells generated with either pGL4based luciferase vectors using hygromycin as a selection marker or by the transduction of 293T cells with lentivirus, which were sorted using GFP as a marker. Of the four systems, two were yeastderived Gal-based heterologous systems, namely, a beta-estradiol (β-EST)-dependent Gal4-estrogen reporter alpha (Gal4-ERα) and a Gal4-CREBbinding protein (Gal4-CBP). The third luciferase assay included the treatment of a single NF-κB reporter element with a battery of proinflammatory cytokines, and the fourth was the treatment of multiple NF-κB target genes by only one cytokine, i.e. TNFα.
To examine whether BaSET modulates gene activity, 293T cells were stabilized with pGL4-Gal Upstream Activation Sequence (Gal-UAS), which is in tandem with luciferase gene. Later, these stable 293T cells were transfected with Gal4-ERα alone or together-with increasing concentrations of Flag-BaSET. The Gal4-ERα contains the yeast Gal4 DNA-binding domain fused to an ERα ligand binding domain (Gal4-ERα) that can induce Gal-UAS when activated by β-EST. After transfection, cells were treated with 1 0 nM of β-EST for 24 h. Then, cells were harvested and lysed to quantitate the luciferase activity. The data demonstrate that Gal4-ERα alone resulted in an 800-fold increase in the Gal-UAS activity. However, increasing concentrations of BaSET expression could effectively repress the β-EST/Gal4-ERα-mediated activation of the Gal-UAS promoter (Fig. 2F).
CBP serves as a ubiquitous co-activator that activates many transcription factors (49). Acetylation of histone proteins within the chromatin of a promoter is one of the key mechanisms through which CBP co-activates downstream genes (50)(51). To directly evaluate the impact of BaSET -mediated histone H1 methylation on CBP co-activation function, we transfected Gal4-CBP into 293T cells stabilized with Gal-UAS.
The results obtained after estimation of luciferase activity show that Gal-CBP alone activated Gal-UAS by 1400-fold. Subsequently, this CBP dependent activation of the UAS promoter was inhibited by increasing expression of the Flag-BaSET, thereby, suggesting that H1 methylation by BaSET possibly counters CBP-mediated gene activation (Fig. 2G).
Based on the above observations, the BaSET protein appears to be a potential transcriptional repressor, prompting us to speculate that B. anthracis could be using BaSET to repress NF-κB transcriptional activation. To test this, HEK293TN cells were transduced with lentivirus containing NF-κB response elements (NF-κB_RE) adjacent to GFP and luciferase, whose induced expression will reflect NF-κB activation. Treatment of transfected 293TN-NF-κB_RE cells with pro-inflammatory cytokines, including TNFα, IL-6, IL-8, and IL-1 showed enhanced luciferase activity, indicating NF-κB activation. Subsequently, transient transfection of Flag-BaSET led to inhibition of NF-κB activation upon treatment with each pro-inflammatory cytokines (Fig. 2H). Overall, the data support our conclusion that BaSET could effectively reduce the activation of the NF-κB_RE in a dose dependent expression.
Next, we wanted to examine the effect of BaSET on diverse NF-κB target gene promoters. Hence, 293T cells were transfected with pGL4.14 vector containing IL-6, IL-8 and VCAM promoters and stabilized using hygromycin. The NF-κB-RE plasmid served as a positive control, pGL4.14 vector as a normal control, and RPL10 represented a promoter of a housekeeping gene. These NF-κB target gene promoters were transfected to be stabilized into 293T cells, and later were again transfected with BaSET and renilla luciferase. Subsequently, the luciferase data demonstrate that upon TNFα treatment, IL-6 was activated 9-fold, IL-8 10-fold and VCAM 8-fold (Fig. 2I). Similar to the above results, BaSET expression led to repression of IL-6 by 3-fold, IL-8 by 2-fold and VCAM by 3-fold (Fig. 2I). All firefly values were normalized with renilla luciferase and were performed in triplicate, including technical and biological repeats.
For a better understanding and summarizing above results, using purified BaSET protein as standard (data not shown), we estimated that 1.0 µg pcDNA-BaSET vector when transfected into ~one million 293T cells expressed 250 nanogram of BaSET protein, which was detectable in 50 µg of cell lysate. Thus, 250 nanogram of BaSET has ability to repress activation of promoter systems, including β-EST/Gal4-ERα/Gal-UAS by ~1.3 fold, Gal4-CBP/Gal-UAS by ~2.5 fold, and NF-κB RE by ~1.8 to 2.5 fold across various cytokines, as well as, NF-κB target genes respectively. This ability of BaSET to repress was twice enhanced with 2.0 µg of BaSET plasmid. Although 3.0 µg of plasmid does not enhance either BaSET expression or methylation, it still continues to repress activation, indicating accumulation of stable repressor complex or heterochromatinization.

BaSET downregulates expression of NF-κB Targets
To examine the ability of BaSET to inhibit endogenous NF-κB target genes, Flag-BaSET was expr essed in stabilized HeLa Trex cells after being cloned into a tetracycline (Tet) inducible pcDNA4TO vector. After transfection of empty pcDNA4TO and Flag-BaSET-cloned pcDNA4TO, HeLa cells were stabilized with blasticidin, which was later induced with Tet. Fig. 2J shows Tetinducible expression of BaSET, which is detected by IP, lanes 1-5) and direct immunoblot (lanes 7-10). Subsequently, we selected limited but heterogenous NF-κB target genes, including TNFα, a pro-inflammatory cytokine, c-Fos and c-Jun, which are early genes containing NF-κB binding sites in their promoters, and vascular endothelial growth factor (VEGF), a downstream target of NF-κB that supports angiogenesis. After TNFα treatment, there was a 180-fold enhancement of TNFα, 20-fold enhancement for c-Fos, no change in c-Jun and a 10-fold increase in VEGF transcripts. Induction of HeLa Trex cells with one microgram of Tet, which induced the expression of Flag-BaSET, led to between 30 to 50-fold reduction in the levels of TNFα, c-Fos, and VEGF transcripts (Fig. 2K). Collectively, the data demonstrate a unique role of BaSET in directing chromatin-mediated methylation of histone H1 that leads to inhibition of NF-κB functions.

BaSET is secreted by B. anthracis and is localized to the nucleus of infected macrophages where it methylates histone H1
Since BaSET is an enzymatically active histone H1 methyltransferase, we determined whether it is secreted by B. anthracis. To address this, a GFP encoding plasmid was electroporated into the Sterne strain. Both GFP-Sterne and their growth media were harvested at 8 and 24 h. After cell lysis, both lysate and media were subjected to immunoblotting with anti-BaSET antibody. The results revealed that BaSET is intracellular at 8 h, but at 24 h, the BaSET protein is found in the growth medium, raising the possibility for secretion. To determine whether the presence of BaSET in the medium is the result of secretion or bacterial lysis, we probed the same blot with anti-GFP antibody. The data show that GFP remained intracellular at 8 and 24 h, indicating that BaSET in the medium is mainly due to secretion (Fig. 3A). However, at 24 h a small amount of GFP could be detected which is likely the result of a small amount of bacterial lysis. Taken together, the mechanism by which BaSET translocates outside the bacillus remains currently unclear, since analysis of the BaSET sequence does not reveal an obvious secretion signal and the upstream and downstream regions of the gene give no further clues.
A critical step during the B. anthracis infective cycle involves its phagocytosis by macrophages as well as its ability to challenge host defenses. Published studies have indicated that toxins alter signaling pathways in B. anthracis pathogenicity, which favor its persistence (16,(19)(20)52). However, the presence of a histone methylating BaSET protein in B. anthracis raises the question as to whether secreted BaSET can be detected within the nucleus of infected macrophages. To address this, cytosolic and nuclear extracts were prepared from a macrophage cell line (RAW 264.7), which was infected for 2 and 4 h with B. anthracis. Using histone H3 as a nuclear and actin as a cytosolic marker, the direct immunoblot data indicate that BaSET is localized inside the nucleus of RAW cells (Fig. 3B).
We next questioned whether the presence of BaSET in the nucleus of infected RAW 264.7 cell line is time-dependent. To address this, we prepared nuclear extracts from uninfected and infected RAW 264.7 cells ranging from 0 to 12 h post-B. anthracis infection. The immunoblot data show that at one-hour post infection, the BaSET protein first appears in the nucleus and the levels peak between 2 and 4 h followed by rapid decline (Fig. 3C). This suggests that the expression of BaSET may be tightly regulated, likely indicating a restricted role during early stages of infection.
To confirm, whether the BaSET protein methylates macrophage chromatin, nuclear extracts were prepared and analyzed from uninfected and B. anthracis-infected RAW cells at 2 and 4 h post-infection. Histone H1 was then immunopreciptated with a polyclonal antibody and followed by immunoblot with polyclonal pan-Kme3 antibody, hence the presence of heavy chain. The data indicate that the presence of BaSET protein in the nucleus resulted in increased levels of methylated H1, suggesting that BaSET facilitates transcriptional repression by altering chromatin modification (Fig. 3D).
Having demonstrated the biochemical consequence of secreted BaSET upon macrophage infection, we determined the effect of BaSET protein on the growth and viability of macrophages. Towards this goal, we engineered a BaSET deletion mutant of Sterne (BaΔSET) (Supplemental Fig. S4A, lane 2). The RAW cell line was incubated with spores generated from Sterne and BaΔSET bacilli (Supplemental Figs. S3E and S3F). The growth characteristics of infected cells were determined by BRDU incorporation and later detecting the cells undergoing division by ELISA using anti-BRDU antibody. The results show that the RAW cells infected with BaΔSET grow twice as fast as Sterne (Fig. 3E). This difference correlated with evaluation of apoptosis, in which RAW cells infected with Sterne showed ~2.5 times more apoptosis than RAW cells infected with BaΔSET.
These results were obtained by quantitative ELISA that detects the presence of oligonucleosomes using anti-histone antibody representing cleavage of the integrated genome by activation of apoptotic pathways (Fig. 3F).
Insight that BaSET is a transcriptional repressor and BaΔSET does not alter the growth of RAW cells prompted us to examine whether BaSET could impact NF-ĸB targets. Therefore, total RNA was extracted and cDNA was prepared from RAW cells that were treated with TNFα (positive control) and infected with Sterne and BaΔSET bacilli. The data show that compared to BaΔSET, Sterne downregulated the expression of TNFα ~3-fold and IL-6 ~2-fold, as well as known NF-ĸB targets, including c-Fos 2-fold and c-Jun ~4-fold. IL-4 and IL-5, which represent Th2 cytokines, remained unaffected (Fig. 3G).
Finally, we wanted to determine if differences in cellular responses (Figs. 3E and F) and activation of NF-κB target genes (Fig. 3G) are in-part due to the lack of H1 methylation by BaΔSET bacillus. Therefore, after 4 h, histone H1 was immunoprecipitated from the nuclear extracts that were prepared from RAW cells infected with Sterne and BaΔSET. Later, probing of immunoprecipitated histone H1 with pan-Kme3 antibody revealed that only RAW cells infected with Sterne had elevated level of H1 methylation, but not cells infected with BaΔSET (Fig. 3H).

Effects of BaSET deletion (BaΔSET) on growth and morphology of B. anthracis
To determine if BaSET plays a role in the growth of B. anthracis, we compared the growth curves of Sterne and BaΔSET with aeration. The results show that the lag phase of the BaSET mutant is extended by about 1.5 h (Fig. 4A). After that, though the growth curves of the two organisms remain the same, the BaΔSET always remains behind. To investigate whether the extended lag phase in the BaΔSET mu t ant is a direct result of the missing BaSET protein, we examined the expression of BaSET in Sterne for 24 h by immunoblot. Surprisingly, the results show that BaSET begins to be expressed at 6 h after inoculation (Fig. 4B), suggesting that SET 12 has a role later in the growth cycle during late log and stationary phases of growth.
We next determined if the growth rate differences with BaΔSET mutant had an effect on cell morphology. Over a 24 h time course, phasecontrast microscopy presented a similar, yet offset, series of morphological changes which involved a filamentous phase for both the Sterne strain (most obvious at 3 h) and the BaSET mutant strain (most obvious at 15 h). It is possible that these morphological changes seen in the BaΔSET mutant were delayed because of its slower growth rate compared to Sterne. For example, at 0 hours both Sterne and BaΔSET displayed similar morphologies, which correspond to those published previously for B. anthracis (Fig. 4C, panel a and f). At 3 h however, when Sterne is in mid-log phase, it has an elongated morphology whereas BaΔSET cells, which are in late stationary, look similar to the initial inoculum (Fig.  4C, panel b and g). By 5 h (around the time BaSET is expressed) Sterne has septated and returned to its original length and the BaΔSET mutant became elongated (Fig. 4C, panel c and h). While the transition between elongation and septation in Sterne lasted for about 2 h, for the BaΔSET mutant, elongation persisted from 5 h to >15 h, >5-times that of Sterne. (Fig. 4C, panel d  and i). By 24 h BaΔSET had septated and returned to its original morphology similar to Sterne (Fig. 4C, panel e and j). When CFU counts were made for both mutant and Sterne at 24 h, they were found to be the same indicating that chain length and bacillus separation had normalized for both strains.
Finally, transmission electron microscopy (TEM) was used to examine the cellular ultrastructure of the BaΔSET mutant compared to wild-type. The 15-hour time point was selected due to the greatest difference in morphology between Sterne and the BaΔSET mutant by phase contrast microscopy. No gross morphological defects were observed in the BaΔSET mutant compared to wild-type (Figs. 4D and E). However, we see that while some elongated forms of Sterne have clearly visible septa, that have not yet separated, the markedly elongated forms of BaΔSET exhibit few if any septa (Figs. 4D and E) supporting the phase contrast images (Fig. 4C) showing a significant delay in septum formation. Taken together, these findings suggest that BaΔSET has an extended stationary phase compared to Sterne, which could be related to the observed delayed septum formation and separation of BaΔSET mutant cells.

Effects of BaSET complementation on the growth and phenotype of BaΔSET bacilli
To further investigate the effects of the BaSET gene on B. anthracis growth and morphology, the BaSET gene was complemented in the BaΔSET mutant under xylose control (cBaΔSET). Immunoblot analysis of BaSET expression in cBaΔSET revealed that xylose could control the expression of BaSET without any apparent leakiness (Fig. 5A). Although induced, the BaSET expression in cBaΔSET remains unregulated unlike the native BaSET protein in Sterne (Figs. 4B vs. 5B). We induced BaSET in cBaΔSET at 0 h, well before the 6 h it is first expr essed in Sterne. Similar to BaΔSET, the cBaΔSET organisms lagged in growth behind Sterne but unlike BaΔSET it managed to catch up by 10-15 h (Fig. 4A vs. Fig. 5C). The constitutive production of BaSET resulted in phenotypic morphological changes in cBaΔSET. Unlike Sterne, which becomes elongated by 3h, the cBaΔSET cells are generally seen as mono-and di-bacilli and significantly shorter than Sterne (7.4 µM vs. 17 µM) (Fig. 5D). By 5 h the cBaΔSET bacilli are still found as singles and doublets but appear somewhat longer (8.7 µM vs. 12 µM) than at 3 h. By 15 h when Sterne has septated into mostly single bacilli, cBaΔSET become shorter in stark contrast to wild type Sterne (5.6 µM vs. 9.1 µM). This length difference remains at 24h (4.1 µM vs. 10.4 µM) (Supplemental Table 1 and Supplemental Fig. S4B). These results strongly suggest that the BaSET gene may in some way be involved in regulating septum formation, since uncontrolled BaSET expression in cBaΔSET results in unregulated cell division and as such a significant reduction in cell size.

BaSET in sporulation, germination and pX01 retention
Given the previous results, we wished to know whether deleting the BaSET gene had an 13 effect on B. anthracis sporulation and germination. In a 24 h sporulation assay, the Sterne strain yielded a culture consisting of 83% heat-resistant endospores, while the BaΔSET mutant yielded a culture consisting of 73%) heat-resistant spores. This difference was not significant, thus, indicating that BaSET is not required for spore formation.
To determine if spore germination was affected by the SET gene deletion, spores of both the Sterne and BaΔSET strains were examined following the addition of germinant. By following the loss of optical density in BaΔSET and Sterne spores, the plots were almost superimposable, with 80% of optical density being lost in 60 min. (Fig.  6A).
To corroborate this finding, spore germination was monitored by loss of heat resistance. The germination rates of BaΔSET and Sterne were again almost superimposable with all spores germinating by 15 min. (Fig. 6B). These findings eliminate the possibility of a major role for BaSET in the formation and germination of heat-resistant endospores.
To rule out the possibility that the creation of BaΔSET do es n ot resu l t in loss of the pX01 plasmid, we isolated pX01 from Sterne and BaΔSET strains at 3, 8 and 15 h post inoculation. The result clearly demonstrated that there was no loss of the pX01 plasmid from either organism at any time point (Fig. 6C).

The significance of BaSET in a mouse virulence model
Considering the secretion of BaSET by intracellular B. anthracis, and the subsequent localization of BaSET to the nucleus, we evaluated the importance of the protein to the pathogenicity of B. anthracis in vivo. We examined the virulence of both the BaΔSET mutant and Sterne strains in a mouse bacteremia model of infection. An intraperitoneal (i.p) route of infection with B. anthracis was chosen that has been shown to induce a rapidly fatal illness in mice (53)(54). Approximately 1.0x10 8 exponential phase Sterne strain cells (harvested at 3 h) and BaΔSET mutant cells (harvested at 5 h) were injected into C57BL/5 mice. These time points were selected due to similarities in OD 600 , growth profile, and morphology of the Sterne and BaΔSET B. anthracis cells. Within 24 h, Sterne-infected mice presented characteristic symptoms of infection followed by death in 4 days or less (Fig. 7A). Mice infected with the BaΔSET mutant, however, exhibited no sign of illness and survived at least up to two weeks after infection.
To investigate the in vivo survival rates of the Sterne and BaΔSET, infected mice were euthanized at 3 h post-infection and the internal organs (including the kidneys, liver, spleen and heart) were recovered, homogenized and tested for bacterial counts. Sterne infected mice were found to have colony forming units in all excised tissues with the most in the spleen, while no colony forming units could be recovered in the BaΔSETinfected mice (Fig. 7B). These results support a major role for BaSET in the virulence of B. anthracis and the loss of this characteristic in the BaΔSET mutant.

DISCUSSION
Multicellular eukaryotes possess a complex set of anti-microbial recognition and defense mechanisms that enable their survival despite continuous interactions with both pathogenic and non-pathogenic microbes. Nevertheless, microbial pathogens have evolved a range of effector molecules, or virulence factors for dominating the host defenses for their growth and propagation. Overall, the consequence of this complex interplay is exhibited in the signaling pathway that leads to NF-κB activation and the downstream inflammatory response. Most bacterial anti-host proteins counter the upstream signaling pathways to negatively impact NF-κB transcription functions. As a result of this inactivation, NF-κB is retained in the cytoplasm of the host, preventing nuclear localization and transcriptional activation. In this study, we show that B. anthracis distinctively utilizes its unique effector, SET protein, to not only inhibit nuclear functions of NF-κB in the host but also to control its own bacterial growth. This is in complete contrast to metazoan SET domains, which have mainly one function, gene activation or repression.
In the present study, we have utilized computational, biochemical, and molecular methods, coupled with a murine bacteremia model, to characterize the functions of a SET protein, from B. anthracis. We show that in B. anthracisinfected macrophages, secreted BaSET localizes to the nucleus. Although the data is not shown, the anti-BaSET antibody also detects the SET protein from B. cereus, which stems from its high degree of homology with BaSET. Results from mice infected with either the BaΔSET mutant or its parental Sterne, do suggest an important role for BaSET in pathogenesis. Our inability to identify BaΔSET organisms in the organs of infected mice 3h post-administration suggests changes in both their pathogenicity, i.e., their ability to be cleared by phagocytes, as well as growth rate. We do in fact see changes in the in vitro growth rate, with the mutant having an extended period of adaptation before a normal looking but slightly suppressed exponential phase compared to Sterne. However, during this exponential phase, the BaΔSET mutant is still in an elongated state and has not yet separated into individual bacilli like the Sterne. This extended delay in adaptation and septation coupled with yet unknown further alterations in the mutant could explain their nearly complete eradication from the tissues of infected mice.
Though there was similarity in the controlled expression of BaSET in B. anthracis and infected macrophages, we did observe a difference in the timing at which BaSET first appears (2 h in macrophages versus 6 h in bacillus). This early expression in macrophages could be due to the need of repressing the host's gene transcription, thus countering the immune pressure with a unique mechanism. Based on mouse virulence studies, our results suggest that the BaSET protein plays a role in the pathogenesis of B. anthracis, since deletion of the gene eliminates the capacity for the organism to cause disease and death, as well as survival in the infected host. Whether this is a direct result of BaSET on septation or some other unidentified effect is currently unknown. Therefore, BaSET might play a critical role for both, B. anthracis growth and pathogenesis.
In earlier studies the highly conserved SET domain was considered to be mainly associated with regulation of developmental genes in metazoan. Subsequently, it became clear that the occurrence of SET domains is widespread including unicellular organisms. While the functions of SET domains in higher multicellular organisms are well characterized, it has been speculated that SET domain proteins have no biological function in lower organisms. Based on the presence or absence of a post-SET domain motif, in this study, we first highlight the significance of Zinc ion binding in reorganizing the phylogenetic tree of the SET domains. Next, we demonstrate that the SET domain from lower organisms, particularly belonging to pathogens like B. anthracis, can also methylate chromatin of macrophages, like the SET domains of metazoan organisms. The tightly regulated intracellular expr ession of BaSET in the bacillus raises the possibility that BaSET could also methylate intracellular bacillus proteins, which need to be identified. A clue as to SET's role in B. anthracis comes from complementation experiments where inducing SET at time 0 prevents the morphological elongation stage typically seen in Sterne, and results in the hyperseptation of the cBaΔSET mutant causing >60% shorter bacilli by 24h (4.1 µM vs. 10.4 µM). It is unknown at this time whether SET expression and methylation alone in a window between 6 and 12 h of cell growth results in the control of septation seen in Sterne. However, it has been shown in E. coli, that the S-adenosylmethionine deficiency results in septation defects, suggesting that methylation of one or more of the proteins in the formation of the septal ring (i.e., from FtsA to FtsQ) may be required for assembly of the septum (55).
Histones H3 and H4 are the most prominent targets of methylation by multiple SET domain-containing proteins that can result in gene activation or repression. In this study, we show that BaSET methylates Histone H1. Mass spec data suggest that BaSET methylates eight lysine residues, which could be the strategy of the bacillus to hypermethylate host chromatin to silence the host's inflammatory response. Numerous eukaryotic methyltransferases, like EZH2 and G9a methylate more than one lysine residue, for example, G9a methylates histone H3 at two sites K9 and K27, histone H1 at K187 and many cellular proteins (56)(57)(58)(59). However, the expression of eukaryotic methyltransferases is 15 tightly regulated depending upon the cell type and stage of development, which may not be the case with BaSET that mainly has an unregulated expression to achieve the repressor function within the mammalian host. Given the substrate selectivity of BaSET, our results indicate that histone H1 methylation by BaSET leads to transcriptional repression of inflammatory genes thus facilitating the establishment of B. anthracis in the host during early stages of infection. To the best of our knowledge, this is the first instance of the use of histone methylation by pathogenic bacteria for survival in its host. Our future goals are to determine the mechanism by which the BaSET protein leaves B. anthracis and enters the nucleus, and to identify the target lysine residues on H1. We will also utilize BaSET to dissect the role of H1 in gene regulation because most studies have focused on histones that are part of the nucleosomal octamer. While this may be the first instance of histone methylation by a pathogen as a potential virulence determinant, we believe that the presence of SET-like proteins in other bacterial pathogens may reveal similar modes of action.