Search tips
Search criteria 


Logo of nihpaAbout Author manuscriptsSubmit a manuscriptHHS Public Access; Author Manuscript; Accepted for publication in peer reviewed journal;
Protein Expr Purif. Author manuscript; available in PMC 2012 November 1.
Published in final edited form as:
PMCID: PMC3183367

A Bacillus anthracis strain deleted for six proteases serves as an effective host for production of recombinant proteins


Bacillus anthracis produces a number of extracellular proteases that impact the integrity and yield of other proteins in the B. anthracis secretome. In this study we show that anthrolysin O (ALO) and the three anthrax toxin proteins, protective antigen (PA), lethal factor (LF), and edema factor (EF), produced from the B. anthracis Ames 35 strain (pXO1+, pXO2), are completely degraded at the onset of stationary phase due to the action of proteases. An improved Cre-loxP gene knockout system was used to sequentially delete the genes encoding six proteases (InhA1, InhA2, camelysin, TasA, NprB, and MmpZ). The role of each protease in degradation of the B. anthracis toxin components and ALO was demonstrated. Levels of the anthrax toxin components and ALO in the supernatant of the sporulation defective, pXO1+ A35HMS mutant strain deleted for the six proteases were significantly increased and remained stable over 24 h. A pXO1-free variant of this six-protease mutant strain, designated BH460, provides an improved host strain for the preparation of recombinant proteins. As an example, BH460 was used to produce recombinant EF, which previously has been difficult to obtain from B. anthracis. The EF protein produced from BH460 had the highest in vivo potency of any EF previously purified from B. anthracis or E. coli hosts. BH460 is recommended as an effective host strain for recombinant protein production, typically yielding greater than 10 mg pure protein per liter of culture.

Keywords: Bacillus anthracis, protease deficient, protein overproduction, edema factor


The Gram-positive bacterial pathogen Bacillus anthracis secretes high levels of the three proteins that are collectively termed anthrax toxin: protective antigen (PA), edema factor (EF), and lethal factor (LF), when grown under conditions thought to mimic those in an infected animal host. PA is a receptor-binding component which acts to deliver LF and EF to the cytosol of eukaryotic cells; EF is a calmodulin-dependent adenylate cyclase and LF is a zinc metalloprotease that cleaves most members of the mitogen-activated protein kinase kinase family for reviews see [14]. Because anthrax pathogenesis is highly dependent on the actions of the anthrax toxin proteins, vaccine and therapeutic development efforts have focused on countering toxin action, typically by generating antibodies to PA. The anthrax vaccine currently licensed in the USA, and developed almost 50 years ago [5], consists of a partially purified culture supernatant of a protease-deficient strain (V770-NP1-R). PA is the most abundant protein and the key immunogen in this vaccine. Efforts to produce a recombinant PA vaccine from B. anthracis by scale-up of an established process [6] appear to have been hampered by instability of the final product, possibly due to protease contamination.

While the toxin components can be purified as recombinant proteins from B. anthracis culture supernatants [69], the integrity and yields are limited by the B. anthracis proteolytic enzymes that are co-secreted. Two extracellular proteases are reported to be abundant in the B. anthracis secretome: NprB (GBAA_0599) – neutral protease B, a thermolysin-like enzyme highly homologous to bacillolysins from other Bacillus species, and InhA1 (GBAA_1295) – immune inhibitor A1, a homolog of the immune inhibitors A from other members of the Bacillus cereus group [1012]. These two proteases contain zinc-binding motifs typical for the zincin tribe of metallopeptidases, His-Glu-Xxx-Xxx-His, and belong, respectively, to the M4 and M6 families of metalloproteases according to the MEROPS database ( A third metalloproptease, camelysin (GBAA_1290), belonging to the M73 family is found in the secretome of several B. anthracis strains. This protease is similar to the camelysin of B. cereus, a novel surface metalloprotease [13].

B. anthracis also contains a gene encoding the InhA2 metalloprotease (GBAA_0672, M6 family), although it is not known whether this protease is expressed and secreted. This gene is an ortholog of the InhA1 described above (68% amino acid identity). Similarly, the genome of B. anthracis also contains genes encoding TasA (GBAA_1288, M73 superfamily), which is an ortholog of camelysin (60% amino acid identity), and MmpZ (GBAA_3159, ZnMc superfamily), which is a putative extracellular zinc-dependent matrix metalloprotease, a member of the metzincin clan of metallopeptidases. This clan is characterized by an extended zinc-binding motif, His-Glue-Xxx-Xxx-His-Xxx-Xxx-Gly/Asn-Xxx-Xxx-His/Asp [14].

We hypothesized that secretion of these and other extracellular proteases might significantly decrease the levels of intact anthrax toxin components in the B. anthracis secretome, so that disruption of these protease-encoding genes might result in higher protein yields. We previously inactivated NprB and InhA1 individually and found that NprB deficiency reduced proteolysis of casein [15], while coagulation of human blood by B. anthracis required InhA1 for proteolytic activation of prothrombin and factor X [16].

In this report, we describe the adaptation of an improved Cre-loxP system for sequentially deleting additional protease-encoding genes of B. anthracis. Also, we describe a role of each protease in degradation of B. anthracis toxin components and another potential virulence factor, anthrolysin O (ALO) [17]. Our work parallels earlier work to knock out up to eight proteases from Bacillus subtilis [18,19] so as to produce an improved expression host. Finally, we suggest that the final B. anthracis strain generated, designated BH460, lacking six proteases and being sporulation deficient and free of the virulence plasmids, provides an improved host for production of recombinant proteins. As an example, we show that EF produced from BH460 is highly active, whereas previous B. anthracis host strains produced truncated proteins having low potency.



The proteases, oligonucleotide primers, plasmids, and bacterial strains used and analyzed in this study are listed in Tables 14.

B. anthracis Ames ancestor strain genes inactivated or analyzed in this study.
Bacterial strains used in this study.

Bacterial growth conditions and phenotypic characterization

E. coli strains were grown in Luria-Bertani (LB) broth and used as hosts for cloning. LB agar was used for selection of transformants [20]. B. anthracis strains were also grown in LB or FA medium [21]. Antibiotics (Sigma-Aldrich, St. Louis, MO) were added to the medium when appropriate to give the following final concentrations: ampicillin (Ap), 100 μg/ml (only for E. coli); erythromycin (Em), 400 μg/ml for E. coli and 10 μg/ml for B. anthracis; spectinomycin (Sp), 150 μg/ml for both E. coli and B. anthracis; kanamycin (Km), 20 μg/ml (only for B. anthracis). SOC medium (Quality Biologicals, Inc., Gaithersburg, MD) was used for outgrowth of transformation mixtures prior to plating on selective medium. B. anthracis spores were prepared as previously described [22] after growth on NBY minimal agar (nutrient broth, 8 g/liter; yeast extract, 3 g/liter; MnSO4·H2O, 25 mg/liter; agar, 15 g/liter) at 30 °C for 5 days. Spores and vegetative cells were visualized with a Nikon Eclipse E600W light microscope (Nikon Instrument Inc., New York).

DNA isolation and manipulation

Preparation of plasmid DNA from E. coli, transformation of E. coli, and recombinant DNA techniques were carried out by standard procedures [20]. E. coli SCS110 competent cells were purchased from Stratagene (La Jolla, CA) and E. coli TOP10 competent cells from Invitrogen (Carlsbad, CA). Recombinant plasmid construction was carried out in E. coli TOP10. Plasmid DNA from B. anthracis was isolated according to the protocol for the purification of plasmid DNA from B. subtilis (Qiagen, Valencia, CA). Chromosomal DNA from B. anthracis was isolated with the Wizard genomic purification kit (Promega, Madison, WI). B. anthracis was electroporated with unmethylated plasmid DNA isolated from E. coli SCS110 (dam dcm). Electroporation-competent B. anthracis cells were prepared and transformed as previously described [23]. Restriction enzymes, T4 ligase, and Antarctic phosphatase were purchased from New England Biolabs (Ipswich, MA). Taq polymerase, Platinum PCR SuperMix High Fidelity kit and the TOPO TA cloning kit were from Invitrogen. The pGEM-T Easy Vector system was from Promega. Ready-To-Go PCR Beads were from GE Healthcare Biosciences Corp. (Piscataway, NJ). For routine PCR analysis, a single colony was suspended in 200 μl of TE buffer [20] (pH 8.0), heated to 95 °C for 45 s, and then cooled to room temperature. Cellular debris was removed by centrifugation at 15,000 × g for 10 min. Two microliters of the lysate contained sufficient template to support PCR. The GeneRuler DNA Ladder Mix from MBI Fermentas (Glen Burnie, MD) was used to assess DNA fragment length. All constructs were verified by DNA sequencing and/or restriction enzyme digestion.

Construction of vectors for protease gene inactivation

B. anthracis Ames 35 (pXO1+ pXO2) (A35) was used for genetic manipulations. The GenBank database (GenBank Accession No. for the Ames strain is NC_003997) was analyzed for the identification of target genes and for the corresponding primer design. The Cre/Lox genetic modification method was adapted to introduce precise genetic knockouts into B. anthracis genes encoding putative proteases. The general schemes for producing B. anthracis mutants using Cre-loxP system were described previously [15,23]. The system employs vectors we designate generically as pDC, for double-crossover plasmid or pSC, for single-crossover plasmid. These plasmids are derived from the highly temperature-sensitive plasmid pHY304 [24], which has permissive and restrictive temperatures of 30 °C and 37 °C, respectively. The pDC vector was used to inactivate the spo0A (GBAA_4394), nprB (GBAA_0599) [15], inhA1 (GBAA_1295) [16], inhA2 (GBAA_0672) and calY (GBAA_1290) genes. The pSC vector was used to inactivate the tasA (GBAA_1288) and mmpZ (GBAA_3159) genes. Both plasmids were used to produce a genomic deletion of the region from tasA (GBAA_1288) to inhA1 (GBAA_1295). To inactivate the inhA2 gene we amplified left and right fragments with primer pairs 0672LL/0672LR and 0672RL/0672RR (Table 2) and inserted them into pDC to produce the pInhA2I plasmid (I at the end of the proteases gene number means inactivation). To inactivate tasA we amplified left and right fragments with primer pairs 1288LL/1288LR and 1288RL/1288RR and inserted them into pSC to produce the pTasALI and pTasARI plasmids. To inactivate the camelysin gene calY we amplified a DNA fragment overlapping the protease gene with primer pairs 1290L/1290R and inserted the fragment into the EcoRI-site of pHY304. The internal BglII-fragment of the calY gene was replaced with a loxP-Ω-sp-loxP cassette flanked by two BglII sites [15] to create the pCamI plasmid for the gene inactivation. To inactivate the mmpZ gene, we amplified left and right fragments with primer pairs 3159LL/3159LR and 3159RL/3159RR and inserted them into pSC to produce pMmpZLI and pMmpZRI plasmids. To delete the tasA-inhA1 gene cluster we transformed the double protease mutant A35DM strain (having a LoxP site in the inhA1 gene) with pTasALI and integrated the plasmid into the genome as described previously [23]. Subsequent transformation of the recombinant strain with the pCrePAS plasmid eliminated the complete tasA-inhA1 gene cluster and produced the tetra-protease mutant strain A35TM.

Primers used in this study

The Cre recombinase-expressing plasmids pCrePAS and pCrePA both have permissive and restrictive temperatures of 30 °C and 37 °C, respectively, and differ only in the selectable marker. The pCrePA was used for elimination of DNA regions containing a spectinomycin resistance cassette located between two similarly oriented loxP sites [15], while pCrePAS was used in a similar way when the recipient strain did not contain a spectinomycin marker. In that case, the region to be deleted generally contained an erythromycin resistance gene along with backbone vector pSC [23]. In both cases, a single loxP site replaced the DNA region targeted for deletion.

To complement the mutation in the mmpZ gene (A35ΔMmpZ strain) and to restore the deleted tasA-inhA1 region in the A35TM strain, the 3159CF/3159CR and deltaCF/deltaCR PCR fragments were inserted into the pSC plasmid. The resulting pMmpZC and pΔTasA-InhA1C plasmids containing, respectively the intact mmpZ gene and the whole tasA-inhA1 region (C at the end of the proteases gene means complementation), were used for complementation. Each plasmid was inserted separately into the corresponding mutant by a single crossover event. During the crossover both the mmPZ gene and whole tasA-inhA1 region were inserted into the genomes of the corresponding mutants. Subsequent elimination of the vector sequences by Cre recombinase left an intact functional copy of the originally mutated gene along with an inactive duplicate copy of a fragment of the gene.

Preparation of mutants

The strains used and their relevant characteristics are listed in Table 4. The B. anthracis NprB, InhA2, TasA, camelysin, InhA1, and MmpZ mutants were constructed in the A35 strain by the replacement of coding sequences with the loxP element as described in the previous section. The double NprB, InhA1 mutant (A35DM) was created starting from the A35ΔNprB strain (Table 4) as described before [16]. The tetra-protease mutant, A35TM was created by deletion of the tasA-inhA1 gene region in the A35DM strain. The A35TM was then used for inactivation of the inhA2 and mmpZ genes with plasmids pInhA2I and pMmpZLI/pMmpZRI. The resulting penta- and hexa-protease mutants were designated A35PM and A35HM, respectively. The mutant strains were checked at each step to ensure they had retained the ability to sporulate [25]. To intentionally produce a sporulation-deficient hexa-protease mutant of A35HM, we inactivated the spo0A (GBAA_4394) gene in A35HM (Table 4) using the plasmid pSΩL304 [15], obtaining A35HMS. The final protease-deficient spo0A-negative mutant lacking pXO1 was obtained by repeated passage of the A35HMS mutant at elevated temperatures to cure pXO1 as described previously [26]. The final strain was designated B. anthracis BH460 (Table 4).

PCR and sequence analysis of chromosomal modifications

PCR fragments containing loxP sites within mutated genes were amplified and sequenced using primers listed in Table 2 (0672seqF/0672seqR, 1288seqF/1288seqR, 1290seqF/1290seqR, 3159seqF/3159seqR). All primers for PCR and sequencing were synthesized by Operon Biotechnologies, Inc. (Huntsville, AL) or the FDA core facility (Bethesda, MD). Sequences were determined from both sides of the PCR fragments (Macrogen, Rockville, MD). For verification of the genomic deletion in the tasA-inhA1 gene area, the region encompassing the start of tasA gene and the end of inhA1 gene was amplified in the mutant strains by PCR using the primer pair deltaF/deltaR. The location of the loxP site inside the PCR fragments was determined by sequencing the fragments.

Western blot analyses

The A35 strain and the mutants were grown at 37 °C in LB in air to analyze B. anthracis toxin production by Western blot. Overnight cultures were diluted into fresh LB to give A600 = 0.002 and growth was measured at 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 17 and 24 h. Supernatant samples (6 ml) from each time point were filtered (0.22 μm Millex syringe-driven filter units, Millipore, Cork, Ireland) and concentrated 10-fold using Amicon Ultra-4 membranes (Millipore). Samples of 5 μl were mixed with 5 μl of 2X Tris-glycine SDS sample loading buffer (126 mM Tris-HCl, pH 6.8, 4% SDS, 20% glycerol, 0.005% bromphenol blue) (Invitrogen), heated (96 °C, 10 min) and separated on 4–12% Bis-Tris NuPAGE gels) using NuPAGE MOPS SDS running buffer (Invitrogen). Precision Plus Protein Standard (All Blue, Bio-Rad, Hercules, CA) was used as molecular weight marker. Proteins were transferred to MagnaCharge 0.45 μm nylon membranes (Osmonics Inc., Minnetonka, MN) using transfer buffer (25 mM Tris, 190 mM glycine, and 20% methanol), blocked in PBS + 3% skim milk (Difco, Lawrence, KS) for 1 h at room temperature, followed by three 10 min washes in PBST (PBS + 0.05% Tween20). Membranes were then incubated with primary antibody diluted in PBS + 1% skim milk overnight at 4 °C. Mouse monoclonal antibody PA-05-A-G1 Lot#071100-02, 5.9 mg/ml (Naval Medical Research Center, Biological Defense Research Directorate), for detection of PA, and mouse monoclonal antibody LF-03-A-G1 Lot# 150900-01, 7.4 mg/ml (NMRC, BDRD), for detection of LF, were both used at 1:2000. Rabbit antisera for development of blots included anti-EF serum #5900 (used at 1:8000), anti-camelysin (used at 1:2000), and anti-recombinant ALO (used at 1:2000). The latter was a kind gift of Richard Rest, Drexel University College of Medicine, Philadelphia, PA. Appropriate HRP-conjugated secondary IgGs (KPL, Gaithersburg, MD) were used at 1:10000 followed by development with TMB (3,3′,5,5′-tetramethylbenzidine) (KPL).

Isolation and purification of EF protein

EF protein with an N-terminal six-histidine tag was expressed from plasmid pProEx-H6-EF (provided by Wei-Jen Tang) in E. coli BL21(DE3) (Promega) as previously described [27]. EF was expressed in B. anthracis host strains from plasmid pSJ136EFOS, which contains the EF structural gene in plasmid pYS5 under the control of the PA promoter and signal sequence [21]. Host strains BH450 and BH460 containing pSJ136EFOS were grown in FA medium containing 15 μg/ml of kanamycin at 37 °C for 14 h, largely following procedures previously used for production of LF [9]. The cultures were cooled, supplemented with 2 μg/ml of AEBSF [4-(2-Aminoethyl)-benzenesulfonylfluoride·HCl] (US Biological, Swampscott, MA) and centrifuged at 4550 × g for 30 min. All subsequent steps were performed at 4 °C. The supernatants were filter sterilized and supplemented with 5 mM EDTA. Solid ammonium sulfate was added to the supernatants to obtain 40% saturation. Phenyl-Sepharose Fast Flow (low substitution, GE Healthcare Biosciences Corp.) was added and supernatants gently mixed in the cold for 1.5 h. The resins were collected on porous plastic funnels (BelArt Plastics, Pequannock, NJ) and washed with buffer containing 1.5 M ammonium sulfate, 10 mM Tris-HCl, and 1 mM EDTA (pH 8.0). The EF proteins were eluted with 0.3 M ammonium sulfate, 10 mM Tris-HCl, and 1 mM EDTA (pH 8.0), precipitated by adding an additional 30 g ammonium sulfate per 100 ml eluate, and centrifuged at 18,370 × g for 20 min. The proteins were dissolved and dialyzed against 5 mM HEPES, 0.5 mM EDTA (pH 7.5). The dialyzed samples were applied to a Q-Sepharose Fast Flow column (GE Healthcare Bio-sciences Corp.) and eluted with a 0–0.5 M NaCl gradient in 20 mM Tris-HCl, 0.5 mM EDTA (pH 8.0). The EF-containing fractions identified by SDS-PhastGel analysis were purified on a column of ceramic hydroxyapatite (Bio-Rad Laboratories, Hercules, CA) with a gradient of 0.02–1.0 M potassium phosphate containing 0.1 M NaCl (pH 7.0). The fractions containing EF were dialyzed overnight against 5 mM HEPES and 0.5 mM EDTA, pH 7.5, concentrated as necessary, frozen, and stored at −80 °C. The molecular mass of purified EF was estimated by liquid chromatogram-electrospray mass spectrometry using an HP/Agilent 1100 MSD instrument (Hewlett Packard, Palo Alto, CA) at the NIDDK core facility, Bethesda, MD).

Analysis of EF activity in vitro and in vivo

EF activity was measured by analysis of cAMP production in the RAW264.7 macrophage cell line. Cells were grown in Dulbecco’s Modified Eagle Medium (DMEM) with 10% fetal bovine serum, 2 mM Glutamax, 2 mM HEPES and 50 μg/ml gentamicin (all from Invitrogen) at 37 °C in 5% CO2. Cells were seeded in 96-well plates 24 h prior to assays. EF preparations were serially diluted in a constant PA concentration (250 ng/ml) prior to addition to cells and incubation for 1 h at 37 °C. Total cAMP levels were assessed using the BioTRAK cAMP enzyme immunoassay kit (GE Healthcare Biosciences Corp.) according to the manufacturer’s protocol. For analysis of EF potency, groups of five 8-week old female Balb/cJ mice (Jackson Laboratories, Bar Harbor, ME) were injected via tail vein with EF preparations combined with an equal dose of PA. Toxin was prepared in sterile PBS. Mice were monitored for survival for 168 h. All experiments involving animals were performed under protocols approved by the Animal Care and Use Committee of the National Institute of Allergy and Infectious Diseases, National Institutes of Health.


Insertion of loxP into protease genes results in truncation of the corresponding proteins

A number of protease genes on the B. anthracis chromosome as well as the spo0A gene were targeted for inactivation in this study. They are identified here with the gene numbers assigned by The Institute for Genomic Research (Rockville, Maryland; for the “Ames ancestor” strain chromosome (GenBank Accession No. NC_007530, gene designation GBAA_gene number) [28]. The gene numbers are coincident with the previous “Ames” strain chromosome (GenBank Accession No. NC_003997, gene designation BA_gene number) [29]. All genes and proteins inactivated or analyzed in this study are listed in Table 1 together with corresponding locus tags.

The inactivation of the nprB and inhA1genes has been previously described [15,16]. In this study we inactivated the inhA2, tasA, calY, and mmpZ genes. The B. anthracis InhA2 protein is 96% identical in sequence to the InhA2 protease of Bacillus thuringiensis, which is an essential virulence factor in that insect pathogen [30]. The tasA gene (GBAA_1288) located downstream of the putative signal protease gene sipW (GBAA_1287) is only 5 genes upstream of the InhA1 gene (GBAA_1295). The intervening genes include the calY protease gene, and the two regulatory genes sinI and sinR (Table 4). The SinI and SinR proteins play important roles in B. subtilis biofilm formation, acting through protease-dependent processes [31], while in B. anthracis these proteins regulate secreted proteases [32]. It is interesting that the genes corresponding to calY and inhA1 are not found in the sinI, sinR region of the B. subtilis genome. Only the gene for the TasA protease, tasA, is located downstream of sipW and upstream of sinR and sinI [32]. Both TasA and camelysin are similar to the B. subtilis TasA protease (36% and 34% sequence identities, respectively). The final gene selected for inactivation, mmpZ, is reported to form an operon with the downstream gene (GBAA_3160) [33]. The latter gene encodes a hypothetical secreted protein that is overproduced in B. anthracis [10]. The absence of the MmpZ protease in the B. anthracis secretome indicates that this protease could be a target for proteolytic degradation by other proteases during the stationary phase of growth [10].

The protease genes were inactivated as described in the Materials and Methods section, followed by sequencing to locate the loxP insertions and infer the corresponding amino acid changes in the mutated proteins. These are shown in Fig. 1. Typically, the 34-bp loxP sequence will generate a frameshift and early downstream occurrence of a stop codon in an alternative reading frame either within the loxP sequence or soon thereafter. Thus, all four inactivated protease genes encode greatly shortened proteins.

Fig. 1
Sequences of truncated and inactivated B. anthracis proteases

Degradation of proteins in B. anthracis mutants compared with the A35 strain

The levels of the three B. anthracis toxin components, ALO, and camelysin produced by the B. anthracis mutants were compared to those of the parental A35 strain by Western blot (Fig. 2). The ten strains were grown in LB at 37 °C over a 24-h period. All strains grew similarly except the six-protease mutant BH460, which appeared to have a slight lag before reaching exponential growth phase.

Fig. 2
Growth curves and PA, EF, LF, ALO, and camelysin production analyses

Expression of PA by Ames 35 was detectable starting at 5 h of growth. However, full length PA (83 kDa) disappeared by the 9th hour of growth due to proteolytic degradation. Inactivation of the InhA2 protease did not influence production of PA while inactivation of Spo0A or camelysin actually reduced the half-life of PA by 1–2 h. Inactivation of NprB, TasA, or MmpZ resulted in increased PA stability up to 10 h. However, PA produced by all these strains was completely degraded after 17 h. The most stable production of PA among the single knockout mutants was found in the InhA1 strain. PA was present even at 24 h of growth with this strain, although some degradation occurred. Surprisingly, the A35DM double mutant did not demonstrate enhanced PA production. PA degradation generally began at 6–7 h of growth and continued through 24 h. Strikingly, the A35HMS strain with six inactivated proteases produced PA with minimal to no degradation during the full 24 h of growth. Similar analyses were performed to follow production of the other B. anthracis toxin components, EF and LF. Full length EF (89 kDa) was found to be more vulnerable to degradation than PA, while LF (90 kDa) was quite stable when produced from most mutant strains. The levels of full length EF and LF, and the timing of their production, paralleled what was seen for PA for each mutant strain.

The enhanced breakdown of PA and EF found in the A35DM strain may be explained by increased production of camelysin in A35DM compared with A35 or the two corresponding single protease knockouts. Although both the NprB and InhA1 knockout strains had increased levels of camelysin, the double mutation produced what seems to be a greater than additive effect. The camelysin levels produced by the A35DM strain were similar to those by the Spo0A mutant strain, A35ΔSpo0A. Both strains produced camelysin (19 kDa) that remained stable over the 24-h period. These observations on post-translational regulation of camelysin production by several proteases support and expand recently published data demonstrating that the concentration of InhA1 in culture supernatants is inversely proportional to the concentration of camelysin [32]. Another interesting finding of our work was that the global transcriptional regulator Spo0A inhibited camelysin production. Pflughoeft et al. [32] recently demonstrated that B. anthracis sinR (which was deleted in those mutants containing the tasA-ihhA1 deletion) also negatively regulates transcription of camelysin and InhA1, both of which have been suggested to be associated with virulence [12,34].

Knocking out TasA also increased production of camelysin but to a lesser extent than elimination of NprB and InhA1. The InhA2 knockout did not result in any change in camelysin degradation or production when compared to A35, while MmpZ elimination was actually detrimental to camelysin production. We conclude that inactivation of six proteases in B. anthracis leads to increased production of full-length toxin proteins in culture supernatants relative to A35, single, or double protease mutants.

Very low levels of ALO were produced from the A35 strain compared to all the protease knockouts, with the exception of the InhA2 mutant. The ALO produced by the A35 strain completely disappeared after 9 hours of growth. Every protease knockout strain showed increased levels of production or greater stability of ALO (53-kDa band). It is interesting that ALO breakdown started in A35DM (less in A35ΔInhA1) only after 17 h of growth. A similar effect was not seen in the A35HMS strain, which allowed accumulation of full length ALO throughout the 24 h of growth. The ALO gene is under control of a PlcR-dependent promoter, so that the truncation of the PlcR protein in B. anthracis is expected to greatly limit ALO synthesis, along with all the other PlcR-dependent proteins [35]. The fact that ALO can be observed in several of the protease-deficient mutants implies that previous reports of low ALO production may be attributed, at least in part, to its degradation rather than to low expression. Camelysin overexpression did not influence ALO production, indicating that this toxin is not a target for camelysin. Taken together, the results demonstrate the involvement of multiple proteases in controlling the accumulation of extracellular proteins in B. anthracis.

Complementation of mutations restores proteolytic activity

To verify that the changes observed above were due to the intended gene knockouts rather than to unrecognized second-site mutations, we complemented several mutants. To complement the mmpZ mutation, the A35ΔMmpZ strain was transformed with pMmpZC. This plasmid undergoes a single crossover to insert the full length, wild type mmpZ gene next to the mutated one. The pSC vector was eliminated by Cre-recombinase treatment as described previously [23]. The presence of the intact mmpZ gene in A35MmpZC was confirmed by PCR and sequencing. Western blot analysis of LF production from A35ΔMmpZC (Fig. 3A) over 24 h indicated that proteolytic activity was restored to levels similar to that of A35 (Fig. 2). To restore the large tasA-inhA1deletion in the genome of the A35TM strain, we transformed with the plasmid pΔTasA-InhA1C (containing the entire tasA-inhA1 region) and followed steps similar to what was described above. Analysis of ALO production from the complemented strain verified restoration of proteolytic activity (Fig. 3B). Complementation of the A35ΔInhA2 strain, and restoration of the InhA2 protease in the same manner, however, did not result in any difference in secreted proteins (data not shown).

Fig. 3
LF and ALO production by mutated and complemented strains

All the studies described above were done in derivatives of Ames 35, where the toxin proteins are encoded on the large virulence plasmid pXO1. Production of the toxin proteins and secreted proteases in these strains is highly dependent on the growth medium. In particular, the production of the three toxin proteins and certain proteases is greatly enhanced by the addition of bicarbonate [11,33,36]. Thus, it is likely that growth in certain media could produce higher concentrations of both proteases and substrates (e.g., PA, LF, EF, etc.) and this could lead to even greater degradation than observed here.

Bacillus host strains are widely used in biotechnological processes, and avirulent B. anthracis strains have been used in this laboratory for a number of years to produce PA and LF [7,9,37]. However, these strains have not been useful as generic hosts for recombinant protein production due to the secreted proteases demonstrated by the work presented above. Our successful elimination of many of the most abundant proteases suggested that the resulting strains could have value as protein expression hosts. To create an optimal host, we further modified the A35HMS strain by curing it of plasmid pXO1. The resulting strain, designated BH460, is non-toxigenic, and can be considered innocuous since it lacks the major virulence factors of B. anthracis. The permanent deletion of the spoOA gene assures that the strain dies rapidly at the end of exponential growth, eliminating concerns regarding laboratory contamination.

B. anthracis strain BH460 produces full-length EF protein

Production of EF from B. anthracis hosts has previously been difficult because this protein is more susceptible to proteolytic degradation than are PA and LF. The plasmid pSJ136EFOS encodes the mature EF protein with its native N-terminus (thus, the “OS” for original sequence) fused to the PA signal sequence and under the control of the PA promoter. This plasmid is otherwise similar to the plasmids pYS5 and pSJ115 that are routinely used in this laboratory to produce PA and LF, respectively. Protein purified from the transformant BH460(pSJ136EFOS) was compared to a preparation made in a similar way from the single protease mutant host BH450, and to a His6-tagged EF protein purified from E. coli [27], the latter being the type of material used in previous toxicity analyses reported by our lab [38]. SDS-PAGE profiles of the recombinant EF proteins are shown in Fig. 4A. The EF produced from BH460 appeared to be slightly less degraded than that isolated from BH450. This finding was confirmed by mass-spectrometry analyses (Fig. 4B). The molecular mass of the recombinant protein isolated from BH460 (88,820 Da) compared well with the theoretical molecular weight for EF (88,822 Da), differing by 2 Da, which is within the instrumental error. A second species was found that had a lower mass (88,687 Da) consistent with loss of the N-terminal methionine. These two protein species were present in about equal amounts (47% for the larger, 53% for the smaller). Mass spectra of the EF produced from BH450 showed degradation as indicated by losses of 1495.7 and 2455.7 Da, resulting in proteins of 87,326 and 86,366 Da, found with similar abundances of 54% and 46%, respectively (Fig. 4C). EF purified from E. coli was monomorphic, with a mass of 89,995 Da, differing by only 6 Da from the theoretical molecular weight (89,989 Da, Fig. 4C). These results clearly demonstrate production of full length EF from BH460 compared to the truncated proteins made by BH450.

Fig. 4
Comparison of EF proteins purified from either E. coli BL21(DE3) or B. anthracis BH450 and BH460 strains

Activity of EF produced from B. anthracis protease mutants

As noted above, recombinant EF has previously been difficult to produce from B. anthracis host strains transduced with plasmids such as pSJ136. Furthermore, the EF that was obtained either from B. anthracis Sterne strain culture supernatant [39] or E. coli [27] consistently had higher potency in inducing cAMP production in cultured cells or lethality to mice than EF produced from B. anthracis (data not shown). In fact, no previous recombinant EF preparations from B. anthracis have been lethal to mice even when injected in doses as high as 100 μg (combined with equimolar PA) (data not shown). The ES-MS analyses shown in Fig. 4 suggest that the low potency of previous B. anthracis-derived EF preparations could be due to degradation. Consistent with prior results, the BH450-derived EF displayed an extremely low level of specific activity and was not lethal for mice (Fig. 5A and 5B). However, the recombinant EF purified from the BH460 culture supernatant had a specific activity exceeding that of highly active E. coli BL21(DE3)-derived EF (Fig. 5A). Similarly, when injected with equal doses of PA, recombinant EF prepared from BH460 was lethal to animals at the 25 μg dose, whereas this dose of E. coli-derived EF had minimal effect (Fig. 5B). EF purified from BH450 was not lethal at 50 μg (Fig. 5B) and even at doses up to 100 μg (data not shown). Thus, the BH460 strain, which produces 5–7 mg of EF per liter of culture, allows for the first time the purification of substantial amounts of highly active EF from B. anthracis. Because other proteins purified in the same way from B. anthracis have consistently been free of endotoxin, use of these EF preparations also eliminates concerns regarding endotoxin-mediated cAMP co-signaling associated with E. coli-derived preparations. The BH460 strain has also proven very useful for expression of a variety of other proteins, typically yielding in excess of 10 mg of final pure protein per liter of culture (data not shown), as will be reported in separate reports on proteins produced in this system.

Fig. 5
EF activity analyses
Plasmids used in this study


We thank Richard Rest (Drexel University College of Medicine, Philadelphia, PA) for supplying rabbit antibody against recombinant ALO, Wei-Jen Tang (University of Chicago) for providing plasmid pPro-EX-H6-EF, and Aaron Firoved and Pradeep K. Gupta for constructing pSJ136EFOS. We thank Lisa Yeh for assistance with some experiments, D. Eric Anderson (NIDDK) for assistance with ESI-MS, Devorah Crown for help with animal studies, and Xavier Gomis Ruth for his support and encouragement. This work was supported by the Intramural Research Program of the National Institute of Allergy and Infectious Diseases (NIAID), Bethesda, MD, USA.


Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.


1. Leppla SH. Anthrax Toxin. In: Aktories K, Just I, editors. Bacterial Protein Toxins. Springer; Berlin: 2000. pp. 445–472.
2. Moayeri M, Leppla SH. Anthrax toxins. In: Bergman NH, editor. Bacillus anthracis and anthrax. John Wiley & Sons, Inc; Hoboken, NJ: 2011. pp. 121–156.
3. Moayeri M, Leppla SH. Cellular and systemic effects of anthrax lethal toxin and edema toxin. Mol Aspects Med. 2009;30:439–455. [PMC free article] [PubMed]
4. Young JA, Collier RJ. Anthrax toxin: receptor-binding, internalization, pore formation, and translocation. Annu Rev Biochem. 2007;76:243–265. [PubMed]
5. Puziss M, Wright GG. Studies on immunity in anthrax. X. Gel-adsorbed protective antigen for immunization of man. J Bacteriol. 1963;85:230–236. [PMC free article] [PubMed]
6. Farchaus JW, Ribot WJ, Jendrek S, Little SF. Fermentation, purification, and characterization of protective antigen from a recombinant, avirulent strain of Bacillus anthracis. Appl Environ Microbiol. 1998;64:982–991. [PMC free article] [PubMed]
7. Varughese M, Teixeira AV, Liu S, Leppla SH. Identification of a receptor-binding region within domain 4 of the protective antigen component of anthrax toxin. Infect Immun. 1999;67:1860–1865. [PMC free article] [PubMed]
8. Singh Y, Klimpel KR, Quinn CP, Chaudhary VK, Leppla SH. The carboxyl-terminal end of protective antigen is required for receptor binding and anthrax toxin activity. J Biol Chem. 1991;266:15493–15497. [PubMed]
9. Park S, Leppla SH. Optimized production and purification of Bacillus anthracis lethal factor. Protein Expr Purif. 2000;18:293–302. [PubMed]
10. Antelmann H, Williams RC, Miethke M, Wipat A, Albrecht D, Harwood CR, Hecker M. The extracellular and cytoplasmic proteomes of the non-virulent Bacillus anthracis strain UM23C1-2. Proteomics. 2005;5:3684–3695. [PubMed]
11. Chitlaru T, Gat O, Gozlan Y, Ariel N, Shafferman A. Differential proteomic analysis of the Bacillus anthracis secretome: distinct plasmid and chromosome CO2-dependent cross talk mechanisms modulate extracellular proteolytic activities. J Bacteriol. 2006;188:3551–3571. [PMC free article] [PubMed]
12. Chung MC, Popova TG, Millis BA, Mukherjee DV, Zhou W, Liotta LA, Petricoin EF, Chandhoke V, Bailey C, Popov SG. Secreted neutral metalloproteases of Bacillus anthracis as candidate pathogenic factors. J Biol Chem. 2006;281:31408–31418. [PubMed]
13. Grass G, Schierhorn A, Sorkau E, Muller H, Rucknagel P, Nies DH, Fricke B. Camelysin is a novel surface metalloproteinase from Bacillus cereus. Infect Immun. 2004;72:219–228. [PMC free article] [PubMed]
14. Gomis-Ruth FX. Catalytic domain architecture of metzincin metalloproteases. J Biol Chem. 2009;284:15353–15357. [PMC free article] [PubMed]
15. Pomerantsev AP, Sitaraman R, Galloway CR, Kivovich V, Leppla SH. Genome engineering in Bacillus anthracis using Cre recombinase. Infect Immun. 2006;74:682–693. [PMC free article] [PubMed]
16. Kastrup CJ, Boedicker JQ, Pomerantsev AP, Moayeri M, Bian Y, Pompano RR, Kline TR, Sylvestre P, Shen F, Leppla SH, Tang WJ, Ismagilov RF. Spatial localization of bacteria controls coagulation of human blood by ‘quorum acting’ Nat Chem Biol. 2008;4:742–750. [PMC free article] [PubMed]
17. Shannon JG, Ross CL, Koehler TM, Rest RF. Characterization of anthrolysin O, the Bacillus anthracis cholesterol-dependent cytolysin. Infect Immun. 2003;71:3183–3189. [PMC free article] [PubMed]
18. Wu XC, Lee W, Tran L, Wong SL. Engineering a Bacillus subtilis expression-secretion system with a strain deficient in six extracellular proteases. J Bacteriol. 1991;173:4952–4958. [PMC free article] [PubMed]
19. Murashima K, Chen CL, Kosugi A, Tamaru Y, Doi RH, Wong SL. Heterologous production of Clostridium cellulovorans engB, using protease-deficient Bacillus subtilis, and preparation of active recombinant cellulosomes. J Bacteriol. 2002;184:76–81. [PMC free article] [PubMed]
20. Sambrook J, Russell DW. Molecular Cloning; A Laboratory Manual; Cold Spring Harbor, New York: Cold Spring Harbor Laboratory Press; 2001.
21. Singh Y, Chaudhary VK, Leppla SH. A deleted variant of Bacillus anthracis protective antigen is non-toxic and blocks anthrax toxin action in vivo. J Biol Chem. 1989;264:19103–19107. [PubMed]
22. Thorne CB. Bacillus anthracis. In: Sonenshein AB, Hoch JA, Losick R, editors. Bacillus subtilis and other Gram-Positive Bacteria: biochemistry, physiology, and molecular genetics. American Society for Microbiology; Washington, D.C: 1993. pp. 113–124.
23. Pomerantsev AP, Camp A, Leppla SH. A new minimal replicon of Bacillus anthracis plasmid pXO1. J Bacteriol. 2009;191:5134–5146. [PMC free article] [PubMed]
24. Pritzlaff CA, Chang JC, Kuo SP, Tamura GS, Rubens CE, Nizet V. Genetic basis for the beta-haemolytic/cytolytic activity of group B Streptococcus. Mol Microbiol. 2001;39:236–247. [PubMed]
25. Sastalla I, Rosovitz MJ, Leppla SH. Accidental selection and intentional restoration of sporulation-deficient mutants of Bacillus anthracis. Appl Environ Microbiol. 2010;76:6318–6321. [PMC free article] [PubMed]
26. Pomerantsev AP, Kalnin KV, Osorio M, Leppla SH. Phosphatidylcholine-specific phospholipase C and sphingomyelinase activities in bacteria of the Bacillus cereus group. Infect Immun. 2003;71:6591–6606. [PMC free article] [PubMed]
27. Soelaiman S, Wei BQ, Bergson P, Lee YS, Shen Y, Mrksich M, Shoichet BK, Tang WJ. Structure-based inhibitor discovery against adenylyl cyclase toxins from pathogenic bacteria that cause anthrax and whooping cough. J Biol Chem. 2003;278:25990–25997. [PubMed]
28. Ravel J, Jiang L, Stanley ST, Wilson MR, Decker RS, Read TD, Worsham P, Keim PS, Salzberg SL, Fraser-Liggett CM, Rasko DA. The complete genome sequence of Bacillus anthracis Ames “Ancestor” J Bacteriol. 2009;191:445–446. [PMC free article] [PubMed]
29. Read TD, Peterson SN, Tourasse N, Baillie LW, Paulsen IT, Nelson KE, Tettelin H, Fouts DE, Eisen JA, Gill SR, Holtzapple EK, Okstad OA, Helgason E, Rilstone J, Wu M, Kolonay JF, Beanan MJ, Dodson RJ, Brinkac LM, Gwinn M, DeBoy RT, Madpu R, Daugherty SC, Durkin AS, Haft DH, Nelson WC, Peterson JD, Pop M, Khouri HM, Radune D, Benton JL, Mahamoud Y, Jiang L, Hance IR, Weidman JF, Berry KJ, Plaut RD, Wolf AM, Watkins KL, Nierman WC, Hazen A, Cline R, Redmond C, Thwaite JE, White O, Salzberg SL, Thomason B, Friedlander AM, Koehler TM, Hanna PC, Kolsto AB, Fraser CM. The genome sequence of Bacillus anthracis Ames and comparison to closely related bacteria. Nature. 2003;423:81–86. [PubMed]
30. Fedhila S, Gohar M, Slamti L, Nel P, Lereclus D. The Bacillus thuringiensis PlcR-regulated gene inhA2 is necessary, but not sufficient, for virulence. J Bacteriol. 2003;185:2820–2825. [PMC free article] [PubMed]
31. Chai Y, Kolter R, Losick R. Reversal of an epigenetic switch governing cell chaining in Bacillus subtilis by protein instability. Mol Microbiol. 2010;78:218–229. [PMC free article] [PubMed]
32. Pflughoeft KJ, Sumby P, Koehler TM. The Bacillus anthracis sin locus and regulation of secreted proteases. J Bacteriol. 2010;193:631–639. [PMC free article] [PubMed]
33. Passalacqua KD, Varadarajan A, Ondov BD, Okou DT, Zwick ME, Bergman NH. The structure and complexity of a bacterial transcriptome. J Bacteriol. 2009;191:3203–3211. [PMC free article] [PubMed]
34. Liu YT, Lin SB, Huang CP, Huang CM. A novel immunogenic spore coat-associated protein in Bacillus anthracis: Characterization via proteomics approaches and a vector-based vaccine system. Protein Expr Purif. 2008;57:72–80. [PMC free article] [PubMed]
35. Sastalla I, Maltese L, Pomerantseva OM, Pomerantsev AP, Keane-Myers A, Leppla SH. Activation of the latent PlcR regulon in Bacillus anthracis. Microbiology. 2010;156:2982–2993. [PMC free article] [PubMed]
36. Bartkus JM, Leppla SH. Transcriptional regulation of the protective antigen gene of Bacillus anthracis. Infect Immun. 1989;57:2295–2300. [PMC free article] [PubMed]
37. Gupta PK, Moayeri M, Crown D, Fattah RJ, Leppla SH. Role of N-terminal amino acids in the potency of anthrax lethal factor. PLoS ONE. 2008;3:e3130. [PMC free article] [PubMed]
38. Firoved AM, Miller GF, Moayeri M, Kakkar R, Shen Y, Wiggins JF, McNally EM, Tang WJ, Leppla SH. Bacillus anthracis edema toxin causes extensive tissue lesions and rapid lethality in mice. Am J Pathol. 2005;167:1309–1320. [PubMed]
39. Leppla SH. Purification and characterization of adenylyl cyclase from Bacillus anthracis. Methods Enzymol. 1991;195:153–168. [PubMed]