Cloning, Purification and Characterization of the Collagenase ColA Expressed by Bacillus cereus ATCC 14579

Bacterial collagenases differ considerably in their structure and functions. The collagenases ColH and ColG from Clostridium histolyticum and ColA expressed by Clostridium perfringens are well-characterized collagenases that cleave triple-helical collagen, which were therefore termed as ´true´ collagenases. ColA from Bacillus cereus (B. cereus) has been added to the collection of true collagenases. However, the molecular characteristics of B. cereus ColA are less understood. In this study, we identified ColA as a secreted true collagenase from B. cereus ATCC 14579, which is transcriptionally controlled by the regulon phospholipase C regulator (PlcR). B. cereus ATCC 14579 ColA was cloned to express recombinant wildtype ColA (ColAwt) and mutated to a proteolytically inactive (ColAE501A) version. Recombinant ColAwt was tested for gelatinolytic and collagenolytic activities and ColAE501A was used for the production of a polyclonal anti-ColA antibody. Comparison of ColAwt activity with homologous proteases in additional strains of B. cereus sensu lato (B. cereus s.l.) and related clostridial collagenases revealed that B. cereus ATCC 14579 ColA is a highly active peptidolytic and collagenolytic protease. These findings could lead to a deeper insight into the function and mechanism of bacterial collagenases which are used in medical and biotechnological applications.


Introduction
Collagen is the most abundant component of the extracellular matrix (ECM) in vertebrates, which provides not only a flexible scaffold for embedded cells, but also regulates crucially important cellular processes including differentiation, cellular growth, survival, migration, and many more [1]. Dynamic remodeling of the ECM constantly requires redistributions, Studies on B. cereus collagenase activity were exclusively performed using secreted collagenase(s) enriched from supernatants of bacterial liquid cultures [17,18]. Comparable to clostridial collagenases, it was demonstrated that B. cereus collagenase activity is zinc-dependent, could be induced by calcium ions [17], and targets native collagen [17,18]. However, detailed biochemical analyses have not been performed so far.
In a previous study, we have detected gelatinolytic activities of B. cereus ATCC 14579 in zymography analyses and which are strongly active in the logarithmic growth phase [35]. To obtain deeper insights into the molecular functions of the collagenolytic B. cereus proteases, we identified the endogenously expressed collagenase and analyzed recombinant ColA acting as a highly active and secreted true collagenase from B. cereus ATCC 14579.

Bacteria
Bacillus cereus (ATCC 14579) and Bacillus thuringiensis (strain 407) were obtained from Nalini Ramarao (INRA, La Minière, Guyancourt, France). The B. cereus ATCC 14579 deletion mutant lacking the plcR regulon (BcΔplcR) was a kind gift from Michel Gohar (INRA, Génétique Microbienne et Environnement Jouy-en-Josas, France). Bacillus subtilis (DSM 402), Bacillus weihenstephanensis and Bacillus megaterium were included as further controls. All bacteria were grown in brain heart infusion (BHI) medium (Sigma Aldrich) at 37°C overnight shaking at 200 rpm. Overnight cultures were diluted 1:20 in fresh BHI medium and bacteria and their corresponding supernatants from the liquid cultures were harvested after indicated time periods by centrifugation at 3000 x g for 10 min at 4°C. Supernatants were sterile-filtered (0.22 nm filter, Greiner). Bacterial pellets were resuspended in lysis buffer (20 mM Tris pH 7.5, 100 mM NaCl, 1% Triton X-100, 0.5% DOC, 0.1% SDS, 0.5% NP-40) and sonicated 3 x for 30 sec with 50% power [35]. Bacterial lysates were cleared from debris by centrifugation at 16000 x g for 10 min at 4°C and protein amounts were determined by Bradford protein assay (Carl Roth).

Mass-Spectrometry
Supernatants of B. cereus ATCC 14579 were loaded on a preparative gelatin zymogram as described before [35]. Briefly, transparent bands were excised from the zymogram and proteins were electro-eluted using D-Tube™ Dialyzer Midi (Merck Millipore). After concentrating using Centricon 1 (Merck Millipore), the protein samples were separated via SDS-PAGE. To visualize the proteins, gels were stained using 1% Coomassie Brilliant Blue G250 (BioRad). Protein bands were excised and digested with the ProteoExtract All-in-One Trypsin Digestion Kit (Merck Millipore). Resulting peptides were separated by reverse-phase nano-HPLC (Dionex Ultimate 3000, Thermo Fisher Scientific). Peptides were loaded onto the trap column (PepSwift Monolithic Trap Column, Dionex) and desalted with 0.1% (v/v) heptafluorobutyric acid at a flow rate of 10 μl/min. After 5 minutes, trap and separation column (PepSwift Monolithic Nano Column, 100 μm x 25 cm, Dionex) were coupled with a switching valve and the peptides were eluted with an acetonitrile gradient (Solvent A: 0.1% (v/v) FA/0.01% (v/v) TFA/5% (v/v) ACN; solvent B: 0.1% (v/v) FA/0.01% (v/v) TFA/90% (v/v) ACN; 5-45% B in 60 min) at flow rate of 1 μl/min at 55°C. The HPLC was directly coupled via nano electrospray to a Q-Exactive Orbitrap mass spectrometer (Thermo Fisher Scientific). Capillary voltage was 2 kV. For peptide identification, a top 12 method was used with the normalized fragmentation energy at 27%. For protein identification, Proteome Discoverer version 1.4 (Thermo Fisher Scientific) with SequestHD and UniProtKB was used.
To analyze the cleavage site of the GST-ColA ΔSP, the GST fragment was precipitated using glutathione sepharose. After elution, the samples were diluted in 0.1% (v/v) aqueous formic acid to a final concentration of 0.1 mg/ml. Intra-molecular disulfide bonds were reduced with 5 mM TCEP (Sigma-Aldrich Chemie GmbH) at 60°C for 30 minutes. The analysis was carried out on a high-performance liquid chromatography (HPLC) system (Ultimate 3000, Thermo Fisher Scientific) at a flow rate of 200 μl/min. A Supelco Discovery C18 column (150 × 2.1 mm i.d., 3 μm particle size, 300 Å pore size, Sigma-Aldrich) was operated at a column temperature of 50°C. 10 μl of sample were injected in in-line split-loop mode. Separation was carried out with a gradient of 5.0-50.0% (v/v) acetonitrile (ACN, VWR) in 0.1% (v/v) formic acid (FA, Sigma-Aldrich Chemie GmbH) in 20 min, followed by column regeneration at 99.99% ACN in 0.1% FA for 10 min and re-equilibration at 5.0% in 0.1% FA for 15 min. UV-detection was carried out with a 2.5 μl flow-cell at 214 nm. Mass spectrometry was conducted on a quadrupole-Orbitrap instrument (QExactive) equipped with an Ion Max source with a heated electrospray ionization (HESI) probe from Thermo Fisher Scientific. Mass calibration of the instrument was conducted with Pierce™ LTQ Velos ESI Positive Ion Calibration Solution from Life Technologies. The instrument settings were as follows: source heater temperature of 200°C, spray voltage of 4.0 kV, sheath gas flow of 15 arbitrary units, auxiliary gas flow of 5 arbitrary units, capillary temperature of 300°C, S-lens RF level of 60.0, in-source CID of 20.0 eV, AGC target of 1e6 and a maximum injection time of 200 ms. The intact protein measurements were carried out in full scan mode with a range of m/z 1,000-2,500 at a resolution of 140,000 at m/z 200. Deconvolution of the intact ion spectra was carried out with the Xtract algorithm integrated into the software Xcalibur 3.0.63 (Thermo Fisher Scientific).

Cloning, Mutagenesis and Purification of ColA
The collagenase ColA (BC3161, Gene ID: 1205508) was amplified from bacterial genomic DNA from B. cereus strain ATCC 14579. PCR primers (Table 1) were designed to amplify ColA lacking the predicted signal peptide (ColA ΔSP, aa 31-960) or lacking the predicted propeptide (ColA ΔPP, aa 93-960). The amplified BamHI/XhoI flanked PCR products were cloned into the pGEX-6P-1 plasmid (GE Healthcare Life Sciences) and transformed in E. coli BL21 to create a GST-ColA fusion protein. To generate a protease-inactive ColA protein, glutamic acid 501 was substituted by an alanine (ColA E501A ) (primers in Table 1) using the QuikChange Lightning Site-Directed Mutagenesis Kit (Agilent) according to the manufacturer's instructions. For heterologous expression and purification of ColA proteins, transformed E. coli was grown in 300 ml LB medium to an OD 600 of 0.5-0.7 at 37°C and 200 rpm and the expression was induced by the addition of 0.1 mM isopropylthiogalactosid (IPTG) at 30°C for 3 h. The bacterial culture was pelleted at 4500 x g for 30 minutes and bacteria were lysed in 10 ml ice- cold PBS by sonication. The lysate was cleared by centrifugation and the supernatant was incubated with glutathione sepharose (GE Healthcare Life Sciences) at 4°C overnight as described earlier [36]. The fusion protein was either eluted with 10 mM reduced glutathione for 10 minutes at room temperature or cleaved with 180 U Prescission Protease for 16 h at 4°C (GE Healthcare Life Sciences).

SDS-PAGE and Western Blot
Proteins were separated by SDS-PAGE under reducing conditions and stained using 1% Coomassie Brilliant Blue G250 (BioRad). For Western blot analyses, equal amounts of proteins were separated by SDS-PAGE and blotted on PVDF membranes. B. cereus ATCC 14579 ColA was detected using a polyclonal anti-ColA antiserum produced in rabbits immunized with recombinant ColA ΔPP E501A (Davids Biotechnology, Germany). To detect the GST-tag, an anti-GST antibody (Rockland) was applied. Visualizing was performed using Odyssey 1 Fc Imaging System (LiCor).

In Vitro Cleavage Assays
Enzymatic assays with N-[3-(2-Furyl)acryloyl]-Leu-Gly-Pro-Ala (FALGPA) as substrate were performed as described by van Wart and Steinbrink [37] and detailed in the manufacturer's protocol (AppliChem) with some minor modifications. Briefly, ColA ΔPP wt or ColA ΔPP E501A were incubated with FALGPA in 250 mM Hepes, 10 mM CaCl 2 , 10 mM ZnCl 2 . Collagenase G from Clostridium histolyticum (ColG) (Q9X721; Y 119 -K 1118 ) and the peptidase domain of C. tetani ColT (ColT PD ) (Q899Y1, D 340 -K 731 ) were used as controls [9]. The enzymes were analyzed at 0.67 μM using a substrate concentration of 2 mM FALPGA in 60 μL reaction volume. The concentration of FALGPA was verified in solution via UV absorbance at 305 nm (ε 305 = 24.70 mM -1 cm -1 ). All measurements were performed at 25°C, and the decrease in absorbance upon substrate cleavage was monitored at 345 nm in 1 min intervals for 25 min with an Infinite M200 plate reader (Tecan). Complete substrate turnover (100%) was accomplished by ColT PD and ColA ΔPP wt after 20 min. All experiments were performed in triplicates and repeated three times. Ratio of substrate turnover (+/-SD) was calculated normalized on the lowest substrate absorbance measured. Significance was calculated using Student´s t-test (paired, one tailed) with a significance threshold of p < 0.05 (GraphPad Prism 5). For the collagenolytic assays, pepsin-extracted soluble type I collagen from bovine skin (Cell Guidance Systems Ltd) was used at a final concentration of 1 mg/ml (i.e. 3.33 μM assuming a molecular mass of 300 kDa, corresponding to non-polymerized tropocollagen) in 250 mM Hepes pH 7.5, 150 mM sodium chloride, 5 mM calcium chloride and 5 μM zinc chloride. The collagen was incubated with 0.83 μM alpha-chymotrypsin (Sigma-Aldrich) and 0.055 μM ColG, ColT PD , ColA ΔPP wt and ColA ΔPP E501A at 25°C. Samples were taken at the indicated time points and the reactions were stopped with 10 mM copper chloride for alpha-chymotrypsin, and with 50 mM EDTA for the other proteins. To test whether ColA targets additional proteins from the ECM, 1 μg ColA ΔPP was incubated with 1 μg fibrinogen (Bovine Plasma, Calbiochem), 1 μg human vitronectin (R&D), 1 μg laminin (mouse, BD Bioscience), 1 μg collagen (Sigma) or 5 μg casein (Carl Roth) in 50 mM HEPES pH 7.4 or collagenase cleavage buffer containing 50 mM Tris, 500 mM NaCl, 20 mM CaCl 2 , pH 7.6 for 16 hours at 37°C.

Homology Model
The full length amino acid sequence of B. cereus ATCC 14579 ColA (Q81BJ6) was used as a target sequence to search for templates using BLASTp against the protein data bank (PDB) structures [38] as well as the SWISS-MODEL server [39]. Among the templates obtained from the SWISS-MODEL, the 2.5 Å crystal structure of apo collagenase G (ColG) from C. histolyticum (PDB 2Y3U) was considered as the most suitable candidate to model the residues Y93-K850 of B. cereus ATCC 14579 ColA. A sequence identity of 49.80% was observed between the corresponding regions of B. cereus ATCC 14579 ColA and C. histolyticum ColG proteins. The quality of the obtained model (Y93-K850) was evaluated by RAMPAGE [40]. The electrostatic potential surface of both ColA and ColG was generated using the program PyMOL (The PyMOL Molecular Graphics System, Version 1.7.4 Schrodinger, LLC).

Bacterial Species of Bacillus cereus s.l. Express Gelatinolytic Proteases
In a previous study, we detected a gelatinolytic protease in B. cereus ATCC 14579 [35]. To analyze whether also other species from the B. cereus s.l. group and additional Bacillus species express similar active proteases, gelatinolytic activities in lysates of B. subtilis, B. megaterium, B. thuringiensis and B. weihenstephanensis were compared with B. cereus ATCC 14579 in gelatin zymography. Equal protein amounts of B. subtilis (Bs), B. megaterium (Bm), B. thuringiensis (Bt), B. weihenstephanensis (Bw) and B. cereus ATCC 14579 (Bc) lysates were separated in a gelatin zymogram ( Fig 1A). Transparent bands in the coomassie-stained gel refer to an active protease cleaving gelatin. As described previously [35], we detected active proteases in a B. cereus ATCC 14579 lysate with molecular weights between 80 kDa and 110 kDa (Fig 1A, lane 5). A similar protease pattern was also identified in a lysate of B. thuringiensis (Fig 1A, lane 3). In the lysate of B. weihenstephanensis, we observed a predominant proteolytic activity at a molecular weight of 120 kDa (Fig 1A, lane 4). In contrast, the lysate of B. subtilis showed a different activity pattern with two distinct bands at 55 kDa and 35 kDa (Fig 1A, lane 1). In B. megaterium, no activity was detectable ( Fig 1A, lane 2). In order to control equal protein yield of the bacterial cultures, all lysates were analyzed by SDS PAGE and stained with coomassie brilliant blue ( Fig 1B). To identify the gelatinolytic protease of B. cereus ATCC 14579, we performed a preparative zymogram, electro-eluted the proteins from the excised gel and analyzed the proteins by coomassie-stained SDS PAGEs. Excised proteins were finally examined by mass-spectrometry. Among several protein hits, two peptides of the collagenase ColT (bcere0001_4490) from B. cereus m1293 were identified that correspond to the orthologous collagenase ColA (BC_3161) from B. cereus strain ATCC 14579 exhibiting 76.8% identity (  According to the predicted domain architecture, B. cereus ATCC 14579 ColA variants lacking the putative signal peptide (ColA ΔSP) or also the predicted propeptide (ColA ΔPP) were cloned and overexpressed in E. coli to purify ColA ΔSP and ColA ΔPP as N-terminally tagged GST-ColA fusion proteins (Fig 2A). Lysates of E. coli transformed with expression constructs for GST-ColA ΔSP wt were analyzed by coomassie-stained SDS PAGE (S3A Fig, lane 1). GST-ColA ΔSP wt , which had a predicted molecular weight of approximately 132.8 kDa was not observed, but appeared to be processed into a~100 kDa protein and a 26 kDa GST protein as detected by SDS PAGE (S3A Fig). The identity of GST was also confirmed by Western blotting (S3A Fig, lanes 9-12). To analyze whether an autoproteolytic processing of GST-ColA ΔSP wt   occurred, a protease-inactive variant was created by the exchange of the glutamic acid 501 to an alanine in the active center (ColA ΔSP E501A ). In contrast to GST-ColA ΔSP wt in bacterial lysates, proteolytic-inactive GST-ColA ΔSP E501A migrated with the predicted molecular weight of approximately 132.8 kDa (S3A Fig, lane 2). In GST pull-down (PD) experiments with GST-ColA ΔSP wt , only the 26 kDa GST tag could be enriched, while the GST-ColA ΔSP E501A samples yielded the 132.8 kDa full length protein (S3A Fig, lanes 3-4). Furthermore, the gelatinolytic activities of these proteins were analyzed by zymography. GST-ColA ΔSP is gelatinolytically active, but exhibited a lower molecular weight in the range of 100 kDa than the expected 132.8 kDa (S3A Fig, lane 5, asterisk). The GST-ColA ΔSP E501A mutant was proteolytically inactive (S3A Fig, lanes 6 and 8). Together, these data indicate that the N-terminal GST tag of ColA ΔSP was autocatalytically removed. Subsequently, we generated isogenic GST-ColA ΔPP wt and its inactive GST-ColA ΔPP E501A mutant for the further characterization of recombinant B. cereus ATCC 14579 ColA (Fig 2A  and 2B). GST-tagged proteins lacking the propeptide exhibited the predicted molecular weight of 125.6 kDa and were detected in the lysates of IPTG-induced E. coli (Fig 2B, lanes 1-2) and in GST-PD experiments (Fig 2B, lanes 3-4). Comparing the molecular weight of autoprocessed GST-ColA ΔSP wt in E. coli lysates (S3B Fig, lane 7) with the different recombinant protein variants GST-ColA ΔSP E501A (132.8 kDa), ColA ΔSP E501A (106.8 kDa), GST-ColA ΔPP wt (125.6 kDa), ColA ΔPP wt (99.6 kDa), GST-ColA ΔPP E501A (125.6 kDa) and ColA ΔPP E501A (99.6 kDa) (S3B Fig) showed that the autoprocessed GST-ColA ΔSP wt migrated at the size of the 106.8 kDa ColA ΔSP E501A protein. This indicated that ColA wt cleaves off the GST tag, but not the propeptide and could further indicate that ColA cleaves-off its signal peptide in an autocatalytic manner. Therefore, we aimed to identify the cleavage site by mass-spectrometry analyses of the processed GST tag. These analyses revealed that ColA cleaves in the amino acid stretch FQ # GPL in the linker between the GST tag and ColA ΔSP wt protein (S4 Fig).
For further analyses of the proteolytic activity of ColA, we finally expressed GST-ColA ΔPP wt protein in E. coli. In contrast to the inactive GST-ColA ΔPP E501A mutant (Fig 2B, lanes 6  and 8), GST-ColA ΔPP wt proteins were gelatinolytic active in zymography analyses (Fig 2B,  lanes 5 and 7). Both, induction and enrichment of GST-ColA ΔPP wt protein were analyzed by coomassie-stained SDS PAGEs (Fig 2B, lanes 9-11). To remove the GST tag from the fusion protein, GST-ColA ΔPP wt coupled to GST sepharose was incubated with PreScission protease resulting in the release of ColA ΔPP wt (Fig 2B, lane 12).

Recombinant B. cereus ATCC 14579 ColA Acts as a True Collagenase
To examine the peptidase activity of recombinant ColA ΔPP wt from B. cereus ATCC 14579 in more detail, N-[3-(2-Furyl)acryloyl]-Leu-Gly-Pro-Ala (FALGPA) was used as a synthetic peptide-substrate mimicking the consensus collagenase cleavage site. We compared the activity of the B. cereus ATCC 14579 ColA ΔPP wt with the recombinant protease domain of ColT from C. tetani (ColT PD ) [12]. Proteolytic inactive ColA ΔPP E501A and ColG from C. histolyticum exhibiting low peptidolytic activity [12] were included as negative controls. Incubation of FALGPA with ColA ΔPP wt led to a rapid substrate turnover of~85% after 12 min, which was comparable with the activity of ColT PD (Fig 3A).
To further analyze the collagenolytic activity, we tested collagenases on their ability to cleave tropocollagen I composed of the β, α 1 and α 2 chains (Fig 3B, lanes 1 and 6). Incubation of tropocollagen with ColG induced a slight fragmentation of tropocollagen after 6 h, which drastically increased after 24 h (Fig 3B, lanes 2-5). A strong fragmentation of tropocollagen was observed after incubation of tropocollagen with ColA ΔPP wt as no collagen was detectable after 24 h (Fig 3B, lanes 7-10). This underlines that ColA from B. cereus ATCC 14579 targets not only denatured collagen, but efficiently cleaves tropocollagen and thus can be considered as a true collagenase. Most likely, ColA is the main collagenolytic entity described in B. cereus supernatants [17]. In contrast, the peptidase domain of ColT PD which lacked the activator domain, and the inactive ColA ΔPP E501A did not cleave tropocollagen (Fig 3C), similar to chymotrypsin which only induced a limited cleavage of collagen (Fig 3D). These data underline the finding that ColA expressed by B. cereus ATCC 14579 can target native collagen. ColA activity appeared highly specific as we did not identify further substrates from the ECM, such as laminin and fibrinogen (S5A Fig) or vitronectin (S5B Fig). Casein was included as an additional protease substrate (S5B Fig). In comparison with ColG, B. cereus ATCC 14579 ColA is exceptionally active. However, a crystal structure of B. cereus ATCC 14579 ColA is not available yet, which could explain the increased activity. Therefore, comparative protein modelling was used to model the residues  (ColT PD ) from C. tetani were tested for their peptidase activity using FALGPA as a substrate. * p = 0.0161 indicates statistical significance (Student´s t-test, paired, one-tailed). (B) In in vitro cleavage assays, the positive control ColG and ColA ΔPP wt were incubated with tropocollagen type I for the indicated time periods and analyzed by coomassie stained SDS PAGE to analyze their collagenolytic activities. (C) As negative controls, ColT PD and ColAΔPP E501A were investigated. (D) As indicated, α-chymotrypsin was incubated with tropocollagen type I as an additional negative control and compared to untreated tropocollagen type I for the indicated time periods and analyzed by coomassie stained SDS PAGE.
doi:10.1371/journal.pone.0162433.g003 Y 93 -K 850 of B. cereus ATCC 14579 ColA, which comprises the activator domain, a catalytic subdomain and a catalytic helper subdomain (Fig 4A). The quality of the obtained model (Y 93 -K 850 ) was evaluated by RAMPAGE [40] and 95.8% of the modeled residues are present in the favored region, 3.1% of them are in the allowed region and 1.0% residues are outliers. Based on this homology model, the electrostatic surface potential of B. cereus ATCC 14579 ColA was calculated and compared to ColG (Fig 4B and 4C). Interestingly, this analysis revealed that the surface properties of ColA and ColG at the activator domain (Fig 4B and 4C, dotted lines) differed significantly. ColA exhibited a highly basic surface, whereas the corresponding region in ColG was acidic, which might potentially explain the difference in their activities.

B. cereus ATCC 14579 ColA Expression Is Under Control of the plcR Regulon
Based on the finding that both endogenous and recombinant ColA are highly active and true collagenases, we analyzed the activities of collagenases in bacterial lysates and supernatants of B. cereus ATCC 14579 wildtype (wt) and an isogenic plcR deletion mutant (ΔplcR). B. cereus ATCC 14579 strains were cultured in BHI medium and after indicated time periods, equal amounts of bacterial lysates and supernatants were analyzed for collagenase activity in gelatin zymography. As a control, recombinant ColA ΔPP wt (rColA) was included (Fig 5, lane 9). In fact, we observed several gelatinolytic activities in the lysates of B. cereus ATCC 14579 wt ( Fig  5, lanes 1-4, upper panel). To correlate the observed activity with ColA expression, we generated a polyclonal anti-ColA antibody raised against rColA (ColA ΔPP E501A ). We detected ColA protein in bacterial lysates in Western blot analysis (S6 Fig), which corresponds to gelatinolytic activities (Fig 5, upper panel). ColA activity was increased at 2 h and 4 h and decreased again at 8 h post-inoculation. Only a weak activity was observed in the plcR deletion mutant (Fig 5,  lanes 5-8, upper panel) indicating that ColA is regulated by the plcR regulon as described previously [31,33]. Collagenase activities appeared at least at four different molecular weights suggesting that multiple ColA variants were present (Fig 5, upper panel). The intensity of the individual proteolytic activities was slightly changed in the culture supernatants at different growth phases (Fig 5, middle panel). Here, we detected three main activities, which increased at 2 h and 4 h of culture. The activity with the highest molecular weight decreased at 8 h, while the activity of intermediate size increased (Fig 5, lanes 1-4, middle panel). The identity of secreted ColA was further validated by Western blot analysis (Fig 5, lower panel). Our data suggest that ColA processing occurs during secretion, which could be substantiated by the detection of different ColA versions in Western blot analyses (Fig 5, lower panel). In comparison with 99.6 kDa rColA, the largest secreted ColA version from B. cereus ATCC 14579 migrates at the same molecular weight (Fig 5, lanes 1-4). Putative cleavage products were produced, which can result from N-terminally, but also from C-terminal processing as described for ColG [44].

Discussion
Bacterial proteases are often directly or indirectly related to their virulence as they can either interfere directly with host cell functions or are involved in the processing of other bacterial virulence factors [45,46]. Collagenases attracted special attention since they can hydrolyze collagen, which represents an important structural component of the ECM. For instance, the collagenase of V. vulnificus has been described to promote bacterial invasion into human tissue [47,48]. Helicobacter pylori expresses the collagenase Hp0169, which was demonstrated to be an essential factor for bacterial colonization of the murine stomach [49]. Similar roles have been suggested for clostridial collagenases indicating the importance of bacterial collagenases in microbial pathogenesis [7,8]. In the last decades, clostridial collagenases also have been applied in safe and efficient therapeutic intervention strategies for Peyronie's or Dupuytren diseases [50,51]. Hence, intensive research on clostridial proteases increased the knowledge on structure, activity and function of collagenases. However, less is known about collagenases expressed by B. cereus s.l. In our previous work, we observed a highly active collagenase expressed by B. cereus ATCC 14579 [35], which has been reported as a collagenase in the supernatants of B. cereus s.s. in earlier studies [17,18], but was not further characterized. In this study we provide evidence that the described gelatinolytic activity is the true collagenase ColA secreted by B. cereus ATCC 14579, and which targets helical collagen.
Collagenases from Clostridia were described as zinc-dependent metalloproteases composed of an N-terminal signal peptide, a putative propeptide, an activator domain followed by the catalytic peptidase domain, and a varying number of PKD and CBD domains [8,9]. Clostridial domain architecture and structure have been intensively investigated revealing the molecular mechanisms of true collagenases [11][12][13]. Corresponding to previous reports [17,18], ColA also acts as a true collagenase, which targets gelatin, FALGPA and native tropocollagen type I, indicating that ColA prefers the typical collagen motifs Gly-Pro-X and Gly-X-Hyp [14] as cleavage sites. In comparison with the collagenase paradigm ColG from C. histolyticum, B. cereus ATCC 14579 ColA exhibited a fast enzymatic degradation of tropocollagen. Analyzing structural differences, a homology model of the collagenase activator and peptidase domains of B. cereus ColA points to alterations of the surface charge of ColA and ColG that could possibly explain the different enzymatic activities. This feature might play a significant role in collagen recognition and binding and could help to explain the difference in the activities of these two proteases.
B. cereus collagenase has previously been described as a secreted protease [17]. Accordingly, we mainly observed two proteolytic activities in supernatants of B. cereus ATCC 14579. Although we cannot exclude the possibility that additional collagenases were expressed in B. cereus, our data suggest that ColA processing occurs during secretion, which could be substantiated by the detection of different ColA variants in Western blot analyses. Since the size of secreted ColA matched the size of recombinant ColA ΔPP, we assume that secreted ColA lacks not only the signal peptide, but also the propeptide. As the maturation of ColA was not investigated in this study, it remains unknown if ColA is a self-processing protease or if additional proteases are implicated in ColA secretion. In fact, several functions have been proposed for Nterminally located signal peptides and propeptides in Gram-positive and Gram-negative bacteria. It was suggested that propeptides are required for proper folding, for secretion of the mature protease, for maintaining the protease in an inactive state, or for anchoring the protease to the membrane [52]. N-terminal cleavage and conversion from the proform of the protease into an active form was already described for other well characterized metalloproteases like for MMP´s [53] and for trypsin-like proteases [54]. For the metalloprotease Npr of Streptomyces cacaoi, it has been proposed that the signal peptide directs Npr to the membrane. After removal of the signal peptide, Npr is released into the environment, where the propeptide is cleaved-off [55,56]. According to our data, the secreted B. cereus ColA did not contain the propeptide, therefore we assume that the removal of the signal peptide and propeptide is involved in ColA secretion via a yet unknown mechanism. Additional processing of ColA in bacterial supernatants was observed, which could reflect C-terminal cleavage events as already proposed for other collagenases [7,13,57]. However, the molecular consequences of ColA cleavage on the protease function needs to be investigated in future studies.
Many virulence factors of B. cereus are controlled by the transcriptional regulator PlcR [31] via binding to the PlcR box upstream of regulated target genes [32,33]. Described as a pleiotropic regulator, PlcR can control the expression of a wide range of virulence-associated genes including enterotoxins, cytotoxins, and hemolysins [33]. Here, we observed that ColA expression is down-regulated in a plcR-negative B. cereus mutant, which is in line with a previous study that mapped a PlcR box upstream of the bc3161 gene in the B. cereus ATCC 14579 strain [33]. Transcription of PlcR is increased in bacteria at the beginning of the stationary growth phase [58], which correlates with PlcR-controlled expression of secreted B. cereus factors [59]. This is in contrast to another report suggesting a decrease of the plcR transcript and PlcR-controlled nhe and hbl expression in the stationary growth phase of the anaerobic B. cereus F4430/ 73 [60]. Although we did not analyze the PlcR expression in B. cereus ATCC 14579 in our experiments under aerobic conditions, we conclude that PlcR-dependent ColA expression and secretion already start in the logarithmic growth phase [35].
In conclusion, we identified ColA as the major secreted true collagenase of B. cereus ATCC 14579. The existence of additional orthologs to the colA gene bc_3161 in other sequenced B. cereus sensu stricto strains as well as B. cereus group strains (Table 2) implies that expression of bacilli collagenase is a wide spread phenomenon and needs to be investigated in future. Cloning and purification of recombinant ColA revealed that it is a highly active protease that efficiently targets native tropocollagen. Furthermore, it might contribute to bacterial virulence of B. cereus in endophthalmitis or opportunistic infections via collagen degradation in the ECM. Clostridial collagenases already represent a widespread compound in the treatment of several diseases, such as Peyronie's or Dupuytren diseases. The high activity of B. cereus ColA makes it an additional attractive candidate for medical treatments.