Bacillus anthracis secretome time course under host-simulated conditions and identification of immunogenic proteins

Background The secretion time course of Bacillus anthracis strain RA3R (pXO1+/pXO2-) during early, mid, and late log phase were investigated under conditions that simulate those encountered in the host. All of the identified proteins were analyzed by different software algorithms to characterize their predicted mode of secretion and cellular localization. In addition, immunogenic proteins were identified using sera from humans with cutaneous anthrax. Results A total of 275 extracellular proteins were identified by a combination of LC MS/MS and MALDI-TOF MS. All of the identified proteins were analyzed by SignalP, SecretomeP, PSORT, LipoP, TMHMM, and PROSITE to characterize their predicted mode of secretion, cellular localization, and protein domains. Fifty-three proteins were predicted by SignalP to harbor the cleavable N-terminal signal peptides and were therefore secreted via the classical Sec pathway. Twenty-three proteins were predicted by SecretomeP for secretion by the alternative Sec pathway characterized by the lack of typical export signal. In contrast to SignalP and SecretomeP predictions, PSORT predicted 171 extracellular proteins, 7 cell wall-associated proteins, and 6 cytoplasmic proteins. Moreover, 51 proteins were predicted by LipoP to contain putative Sec signal peptides (38 have SpI sites), lipoprotein signal peptides (13 have SpII sites), and N-terminal membrane helices (9 have transmembrane helices). The TMHMM algorithm predicted 25 membrane-associated proteins with one to ten transmembrane helices. Immunogenic proteins were also identified using sera from patients who have recovered from anthrax. The charge variants (83 and 63 kDa) of protective antigen (PA) were the most immunodominant secreted antigens, followed by charge variants of enolase and transketolase. Conclusion This is the first description of the time course of protein secretion for the pathogen Bacillus anthracis. Time course studies of protein secretion and accumulation may be relevant in elucidation of the progression of pathogenicity, identification of therapeutics and diagnostic markers, and vaccine development. This study also adds to the continuously growing list of identified Bacillus anthracis secretome proteins.


Background
Bacillus anthracis is a Gram-positive spore-forming bacterium that is the etiologic agent of anthrax [1]. During the course of infection, the virulent spores germinate to become vegetative cells. The bacterium secretes two major virulence factors, the bipartite edema and lethal toxins that are encoded by plasmid pXO1 [1][2][3]. The common component of both toxins is the protective antigen (PA) which is not toxic by itself. PA binds a specific receptor on the host cells and translocates the edema factor (EF) and lethal factor (LF) inside the cells where they exert their damaging action. Transcription of the genes coding for these virulence factors has been shown to be coordinately induced by bicarbonate-CO 2 . High CO 2 tension is believed to simulate conditions encountered within the host [4]. In addition, the effect of temperature has been shown to be important for toxin production but not for production of antiphagocytic capsule, whose synthesis is encoded by genes on plasmid pXO2 [5]. Thus, when B. anthracis is cultured at 37°C in a bicarbonate-containing minimal medium, toxin production is enhanced.
The composition and identification of B. anthracis extracellular proteins (secretome) has been the subject of recent proteomic studies [6][7][8][9]. In B. cereus, a close relative of B. anthracis, most of these secreted proteins include collagenases, phospholipases, hemolysins, proteases, and enterotoxins that are positively regulated by plcR [10]. However, a non-sense mutation inactivates a homolog of this gene in B. anthracis resulting in a significant reduction of these secreted proteins [1,11,12]. Using 2-DE and MALDI-TOF MS, Chitlaru et al. [6] have identified 64 extracellular proteins in a virulent B. anthracis strain, of which 50 exhibit export signal peptides. Thirty-one of these secreted proteins harbor features that are characteristic of virulence determinants suggesting that in addition to the "classic" lethal and edema toxins, a large number of proteins may be essential for B. anthracis virulence [6]. Additionally, in minimal medium under high CO 2 tension, the presence of the plasmids led to the enhanced secretion of 12 chromosome-encoded and 5 pXO1 encoded proteins. Ten of the chromosome-encoded proteins could not be detected in the absence of the plasmids. These results suggest distinct plasmid and chromosome CO 2 -dependent crosstalk mechanisms that modulate extracellular proteolytic activities. Likewise, Antelmann et al. [7] have identified 64 extracellular proteins in the nonvirulent B. anthracis strain UM23C1-2 (pXO1 -/pXO2 -), 29 of which were predicted to be secreted. The remainder of the extracellular proteins were predicted to be associated with the cell wall or the cytosol. Based on the nature of B. anthracis secretome, it was suggested that this organism is adapted to life in a protein-rich environment due to the presence of a variety of proteases, peptidases, peptidebinding proteins, as well as enzymes required for the metabolism of amino acids. It was also proposed that these secreted proteases and peptidases could be useful targets for the development of improved vaccines.
In another study, Gohar et al. [8] compared the fully cured B. anthracis extracellular proteomes to two other members of the B. cereus group, B. cereus and B. thuringiensis, that were also cured of their plasmids. The secretomes of the three species included both cell wall and cytosolic proteins that are similar in all three species. However, whereas the extracellular proteins of B. cereus and B. thuringiensis contained a large number of secreted proteins such as degradative enzymes and toxins, those of B. anthracis contained only one secreted protein, a metalloprotease InhA1. These results are in contrast to those reported by Chitlaru et al. [6] and Antelmann et al. [7] where 31 and 29 secreted proteins were identified in the culture filtrate of fully cured B. anthracis cells, respectively. Using minimal medium under high CO 2 tension, a comparative analysis of the extracellular proteomes of three isogenic strains of B. anthracis that differed solely in their plasmid content was conducted by Lamonica et al. [9]. In this study, the use of SignalP for the prediction of cellular function and location of each protein indicated that most of the identified proteins were either cell wall-associated or cytoplasmic in the fully cured strain, RA3:00.
The present study characterized the time course of protein secretion and accumulation in the culture of B. anthracis strain RA3R (pXO1 + /pXO2 -). Secretion patterns at the different time points are due not only to differentially expressed proteins at various phases of exponential growth, but also to the accumulation of proteins in the culture and their stability therein. Particular focus was placed on the analysis of protein secretion during the exponential growth phase to preclude the identification of proteins that may have been caused by bacterial lysis. In addition, computer assisted comparative analysis using 2D Phoretix of the cytosol with the secretome at16 hr under induced conditions was performed. LC MS/MS was used as a complementary tool in identifying proteins that were not amenable to MALDI-TOF MS, significantly increasing the number of identified secreted proteins. The data from this analysis add to the growing list of identified proteins in the secretome database of B. anthracis and help in further identifying key pathways associated with virulence. Furthermore, immunogenic proteins, or the immunome, of the secretome were identified using sera from human patients infected with cutaneous anthrax. Identification of these immunodominant antigens is essential for development of more efficacious and safer vaccines against anthrax.

Time course of protein secretion
The extracellular proteomes of B. anthracis strain RA3R were analyzed at three time points during the exponential growth phase, i.e., early (6 hr), mid (10 hr) and late log (16 hr) under induced, host-simulated conditions (Fig.  1). 2D Phoretix computer-assisted analysis of 2-DE average gels revealed that the number of protein spots increased from 54 spots at 6 hr to 308 spots at 10 hr and to 536 spots at 16 hr. (Figs. 2A-C). In contrast, the number of protein spots at 16 hr under uninduced condition ( Fig. 3) is only 178 or 33% of the total number under induced condition (Fig. 4). The S-layer proteins Sap (surface array protein) and EA1 (extractable antigen 1), PA, and enolase were the predominant proteins that increased in secretion (Fig. 5). Other highly expressed proteins include alkyl hydroperoxide reductase, chaperonin 60kDa, phosphoglycerate isomerase, sulfatase, manganese superoxide dismutase, and a zinc-binding lipoprotein.

Extracellular proteins
A total of 275 proteins were identified in the extracellular proteome by a combination of MALDI-TOF MS and LC MS/MS ( Fig. 3/4, additional file 1). MALDI-TOF MS analysis has identified 52 proteins of which 47 were also identified by LC MS/MS analysis. In contrast, 270 proteins were identified by LC MS/MS. This indicates that the combined use of MALDI-TOF MS and LC MS/MS has significantly increased the total number of identified proteins in the extracellular proteome, which has also been observed in the analysis of the proteome of other bacteria such as Mycobacterium tuberculosis [13] and Brucella abortus [14].
2-DE gel images of SYPRO Ruby-stained secretome proteins of toxigenic, non-encapsulated B. anthracis strain RA3R (pXO1 + /pXO2 -) from pH 4 to 7 at 6 hr (A), 10 hr (B) and 16 hr (C) after inoculation under conditions that simulate those found inside the host (induced), and (D) Western blot analy-sis of immunogenic extracellular proteins at 16 hr using sera from patients infected with cutaneous anthrax Figure 2 2-DE gel images of SYPRO Ruby-stained secretome proteins of toxigenic, non-encapsulated B. anthracis strain RA3R (pXO1 + /pXO2 -) from pH 4 to 7 at 6 hr (A), 10 hr (B) and 16 hr (C) after inoculation under conditions that simulate those found inside the host (induced), and (D) Western blot analysis of immunogenic extracellular proteins at 16 hr using sera from patients infected with cutaneous anthrax. Spot numbers refer to additional file 1.
Growth curve for B. anthracis strain RA3R (pXO1 + /pXO2 -) in R medium under host simulated conditions (induced) Figure 1 Growth curve for B. anthracis strain RA3R (pXO1 + /pXO2 -) in R medium under host simulated conditions (induced). Secretome proteins were harvested at time points 6, 10, and 16 hrs as indicated by the arrows.
The B. anthracis secretome proteins were also grouped into major cellular functions (Fig. 6). Half of all identified secretome proteins are involved in energy metabolism (17.8%), protein synthesis (10.9%), cellular structure (8.7%), or are hypothetical (13%). This distribution is significantly different from role category predictions of all genes theoretically encoded by B. anthracis. Some groups were overrepresented or underrepresented relative to the theoretical total B. anthracis proteome, where energy metabolism accounts for 5.2%, protein synthesis is 2.4%, cellular structure is 7.1%, and hypothetical proteins are 40.6% of all cellular proteins. The role category predictions of B. anthracis were obtained from The Institute for Genomic Research website [15].
Some of the secretome proteins have metabolic functions that would typically place them in the cytoplasm. However, 2-DE secretome and cytosol subproteomes show distinct, significantly different patterns from another (Fig. 7). Using MALDI-TOF MS the protein spots in the cytosol were identified, the relative protein spot intensities analyzed, and their ratio to each other determined. These results were used for computer assisted comparison of cytosol and secretome 2-DE gels. Out of the 489 unique Relative protein expression of differentially expressed extra-cellular proteins of B. anthracis strain RA3R (pXO1 + /pXO2 -) during exponential growth at 6 hr (green bars), 10 hr (red bars) and 16 hr (yellow bars), based on spot size and pixel intensity Figure 5 Relative protein expression of differentially expressed extracellular proteins of B. anthracis strain RA3R (pXO1 + /pXO2 -) during exponential growth at 6 hr (green bars), 10 hr (red bars) and 16 hr (yellow bars), based on spot size and pixel intensity. The relative amount of each protein was determined using the 2D Phoretix software. The protein numbers refer to additional file 1.   protein spots in the cytosol, 56 are more than two-fold up regulated, 84 are more than two-fold down regulated, and 177 proteins spots have no corresponding spot in the secretome (Fig. 7B). If protein spots in the secretome were simply the result of cell lysis, not only should their relative position on the gel be the same between the cytosol and the secretome, but the ratio of the proteins to each other in each subproteome should be the same in both. The results show that two thirds of all protein spot positions as well as relative abundances are distinctively different between the cytosol and secretome. Therefore, cell lysis can be excluded as the major contributor to the protein accumulation in the secretome. The extracellular proteins were further analyzed using various bioinformatics software programs, such as SignalP, SecretomeP, PSORT, LipoP, TMHMM, and PROSITE for predicting protein secretion and localization.

Classical Sec pathway
A total of 53 proteins were predicted by SignalP to be secreted in the classical Sec pathway, which is characterized by the presence of a signal peptide [16,17] (additional file 1). Of the 53 proteins containing the signal peptides, 38 proteins have the cleavage site for signal peptidase I (SpI). These proteins are predicted to be secreted into the external environment, because they lack additional retention signals. However, 10 proteins have the cleavage site for signal peptidase II (SpII) and the retention signals for lipid anchors while three proteins were predicted to have transmembrane helices (TMH). These proteins are predicted to have an extracytoplasmic but cell-associated location. LipoP predictions discriminate between lipoprotein signal peptides, other signal peptides, and N-terminal membrane helices with 93% accuracy in Gram positive bacteria [18]. In contrast to LipoP, TMHMM predicted 25 integral membrane proteins with one to ten transmembrane helices [19] (additional file 1). TMHMM predicts transmembrane protein topology with a hidden Markov model with a 98% accuracy. Additional predictions using PROSITE identified several secreted proteins that are also know to be associated with the bacterial cell wall by S-layer homology domains or lipoprotein lipid attachment sites [7,20]. No cell wall associated proteins with LPXTG-motifs were found in the secretome. Despite its name, the LPXTG-motif cell wall anchor domain protein does not contain such a domain. In fact, it contains a NEAT domain that might be involved in the transport of iron.

Alternative Sec pathway
Twenty-three proteins were predicted by SecretomeP to be secreted by the non-classical Sec Pathway characterized by the lack of typical export signals [21]. In contrast to the predictions of SignalP and SecretomeP, PSORT [22] predicted 171 extracellular proteins, 7 cell wall-associated proteins, and 6 cytoplasmic proteins. Using the whole genome of B. anthracis, Binnewies et al. have calculated that the Bacillus anthracis Ames strain has 6% secreted proteins predicted using SecretomeP, 3% using LipoP, and 6% using SignalP [23].

Immunogenic extracellular proteins
The main immunogenic proteins detected by 2-DE Western blot analysis using sera from humans infected with cutaneous anthrax include the 83 and 63 kDa charge variants of protective antigen (PA), followed by charge variants of enolase and transketolase (Fig. 2D). The protective antigen produced in vivo has a molecular mass of 83 kDa Proteins encircled in red are more than two fold up-regulated, proteins encircled in yellow are more than two fold downregulated, proteins encircled in green are between these limits, and proteins encircled in blue are unmatched between the two sub proteomes.
and is subsequently cleaved by cell-associated protease activity resulting in a 63 kDa protein that binds the lethal factor to form lethal toxin [24]. Notably, 17 charge and mass variants of PA were detected. These include five charge variants of the 83 kDa PA isoforms, seven charge variants of the 63 kDa PA isoforms and five charge variants of the 37 kDa PA isoforms. Charge variants of PA have been reported previously. Their generation was due to the spontaneous deamidation of asparagine residues which is dependent on pH, temperature, primary sequence ("nearest neighbor" effect) and protein conformation [25]. Charge variants of enolase and transketolase were also noted. Enolase has also been reported previously as component of the B. anthracis spore [26] and as one of the immunodominant spore antigens [27]. Other minor immunogenic proteins were also detected by 2-DE Western blot analysis but were not identified.

Discussion
The secretome of B. anthracis has been the subject of recent proteomic studies [6][7][8][9]. Interest in its study comes from the fact that several secreted proteins of B. anthracis are known virulence factors (e.g., PA, EF, LF). Other secreted proteins may potentially be involved in the adherence of the bacteria to host cells while some may be required for the suppression of the host's defense mechanisms. While several of the virulence proteins have been identified by recent proteomic studies [6,7], the time course of protein secretion by a toxigenic but non-encapsulated strain RA3R (pXO1 + /pXO2 -) under host simulated conditions has not been reported. In this study, the dynamics of secretion of several proteins were investigated at different time points during the exponential growth phase of B. anthracis. The same collection points have been used previously by Lamonica et al. [9] for the isolation of secretome proteins. Chitlaru et al. [6] and Antelmann et al. [7] isolated secretome proteins at analogous growth stages as well. As shown in Fig. 5, the rate of protein secretion varies for each protein. It should be noted that at 6 hours, PA has the highest rate of secretion of all secreted proteins. Since PA is a major component of the anthrax toxin, its high secretion rate at this early time point confirms the critical role this protein plays during the onset of anthrax pathogenesis. Two other proteins that were detected in high abundance in the culture supernatant are the S-layer proteins, EA1 and Sap, both of which are components of the cell wall. Both proteins contain a signal peptide followed by three SLH (S-layer homology) anchoring domains and are considered major surface antigens. The amount of these two proteins was highest at the 16-hour time point compared to the other identified extracellular proteins. It has been previously described that Sap is sequentially replaced by EA1 [28]. This rapid Slayer turnover results in the release and spillover of these proteins into the secretome.
Several of the secretome proteins have metabolic functions that would typically place them in the cytoplasm. This is similar to the results by Antelmann et al. [7] in which more than half of the identified secretome proteins were associated with the cell wall or cytosol. We also identified 25 proteins that were predicted to have transmembrane helices which would typically associate them to the cell wall. All but one of these proteins were predicted to be secreted by SignalP or SecretomeP and half of these proteins have also been identified before as a natural component of the secretome (see additional file 1).

Extracellular accumulation of B. anthracis proteins
Combined analysis using SignalP and SecretomeP of the B. anthracis secreted proteins indicated that 28% of the detected proteins are extracellular whereas 62% are predicted to be extracellular using PSORT. The remaining proteins were predicted to be cytoplasmic since they lacked known export signals or are cell-associated because they have membrane-anchoring or cell wall retention signals. Using LipoP, some of the proteins are predicted to be bound at the trans surface of the cytoplasmic membrane. Contrary to their predicted location, a number of proteins with retention signals for covalent or non-covalent attachment to the cell walls were also found in the extracellular environment.
While bioinformatics tools are useful for predicting cellular function and localization, considerable variation exists in the number of proteins that were predicted to be extracellular, cytoplasmic, or membrane bound. This is expected because each of these bioinformatics tools uses different algorithms and assumptions in their predictions. Further empirical studies are therefore required to verify the precise location of proteins for which conflicting predictions are noted.

Immunogenic extracellular proteins
The use of sera from human patients infected with cutaneous anthrax confirmed the high immunogenicity of PA in the secreted proteins of the toxigenic but non-encapsulated B. anthracis strain RA3R (pXO1 + /pXO2 -). Although several mass and charge variants of PA were detected, the most immunogenic are the 63 and 83 kDa charge variants. In addition to PA, the major component of AVA (anthrax vaccine adsorbed), enolase and transketolase were found to be highly immunogenic. AVA is an alum precipitate prepared from B. anthracis culture filtrates and is a licensed vaccine currently used to protect humans against anthrax. Enolase was also described previously as part of the AVA culture filtrate [29]. AVA preparations contain contaminating proteins whose benefits are not known, but the presence of the highly immunogenic enolase in this formulation might contribute to its protective immunity over a vaccine containing just PA. The membrane pro-teins Sap and EA1 were also found in the secretome by Chitlaru et al. [6] and Antelmann et al. [7] as the most abundant extracellular cell wall proteins. Both proteins were reported as major surface antigens and potential vaccine carriers in vivo [30,31]. EA1 and Sap are also major components of AVA [29] and are associated with a variety of functions ranging from evasion of host recognition, cell adhesion and resistance, and phagocytosis [1,32,33]. Interestingly no reactivity to these proteins was observed using sera from patients with subcutaneous anthrax infection. Overall fewer immunogenic proteins were detected using human sera from patients infected with cutaneous anthrax than in a recent study by Chitlaru et al. [30] who used experimentally challenged rabbit and guinea pig sera. There is a quantitative and qualitative difference between sera from recovering humans and sera from multiple challenged animals. Since with human sera enolase and transketolase were major immunogens beside PA, they might be promising candidates for next generation anthrax subunit vaccines besides the newly identified proteins using animal sera [30].

Conclusion
The work presented here is the first description of the time course of protein secretion for B. anthracis grown under host-simulated conditions. The combined use of two types of mass spectroscopy led to the identification of 275 proteins and their secretion patterns during the exponential growth phase of B. anthracis. While the secretome contained several predictable proteins (e.g. PA, LF, Sap, and EA1) many proteins were identified which would not be expected in the secretome such as proteins involved in energy metabolism and protein translation. The discovery of proteins in the secretome that are traditionally thought to be strictly cytosolic has often been assumed to result from contamination. However, this may not be true since cytosolic proteins, such as aldolase, enolase, elongation factor G, and various dehydrogenases, have also been detected in the secretome of group A streptococci [34], mycobacteria [35,36], and B. subtilis [37]. Some proteins may be cytoplasmic at one point of the cell cycle and secreted via pathways which are yet to be understood during other stages of the cell cycle. While prediction software are a good tool to characterize a large group of proteins, different algorithms give diverging results. It is far more difficult to accurately predict a precise location within a cell of non-classical secretory proteins than to recognize proteins which are secreted by a signal peptide. The ultimate proof is to empirically validate their results. This study helps to bridge the gap between pure in silico prediction and in vivo observation.
This study also identified the major immunoreactive proteins of the B. anthracis secretome using 2-DE Western blot analysis from humans infected with the pathogen.
These proteins included the expected PA as well as enolase and transkelolase. The immunoreactive secretome proteins contribute to the list of other B. anthracis immunogenic proteins that were idenfied in other subproteomes or life-stages of the infectious agent. This knowledge will ultimately lead to the development of a more-specific, safer, and highly efficacious vaccine against B. anthracis. In addition, the identification of early high abundance secretome proteins may aid in the development of detection and diagnostic kits for those cases, where the direct capture of the pathogen is not possible.   To identify immunogenic proteins, the PVDF membranes were washed three times with PBST for 5 min and then probed with a 1:1000 dilution of sera pooled from patients who recovered from cutaneous anthrax. The PVDF membrane was washed three times with PBST and incubated with 1:5000 dilution of appropriate secondary antibody. Chemiluminescent signals were visualized using the Western Lightning reagents (Perkin-Elmer, Wellesley, MA). Corresponding sera from uninfected humans were used as controls (Cambrex, Charles City, IA). Three replicate blots were used for computer analysis using the Phoretix 2D Expression software to identify the immunogenic proteins.

In-gel trypsin digestion and MALDI-TOF MS
Protein spots were excised, washed, and trypsin digested from 2-DE gels according to manufacturer's instructions using the Xcise robotic workstation (Shimadzu Biotech, Columbia, MD). Briefly, gel plugs were washed with 50 mM ammonium bicarbonate and 50% acetonitrile (ACN), dried, and treated with 1.6 µg/ml of trypsin in 50 mM ammonium bicarbonate at 37°C overnight. Tryptic peptides were applied to a MALDI-TOF MS plate in a solution of 10 mg/ml alpha-cyano-4-hydroxycinnamic acid in 0.1% trifluoroacetic acid and 50% ACN. MS spectra were obtained using an Axima-CFR plus (Shimadzu Biotech) in a positive ion reflectron mode and analyzed against the theoretical spectra of B. anthracis strain Ames, using the Mascot Daemon software package (Matrix Science, Boston, MA). The search parameters were: maximum of one missed cleavage by trypsin, fixed modification of oxidation, charged state of + 1, and mass tolerance of ± 0.5 Da.

Protein identification by LC MS/MS
All LC MS/MS analyses were performed using an Agilent 1100 nanopump system coupled with a Vydac C18 reverse phase column (75 µM) (Agilent, Santa Clara, CA) and a QTRAP 2000 LC MS/MS system (Applied Biosystems, Foster City, CA) outfitted with a nanospray source and controlled with Analyst 1.4.1 software. Proteins were prepared for digestion using heat denaturation and a modified organic-aqueous digestion method as described before [14]. Each digested and desalted sample was re-suspended in 10 µl of Buffer A (95% water, 5% ACN, 0.1% formic acid). Each sample (7 µl) was loaded onto the LC MS/MS system and analyzed using an independent data acquisition method with the following parameters: single enhanced mass spectra (EMS, 500-1500 m/z) from which the three most intense peaks were subjected to an enhanced resolution, from which ions with a charge state of + 2 to + 4 were subjected to an enhanced product ion [EPI (MS/MS)] scan. Once the three most intense peaks were subjected to downstream analysis, they were ignored for a period of 60 sec. LC gradient was 5-60% Buffer B (95% ACN, 5% water, 0.1% formic acid) over 35 min. General parameter settings are as follows (curtain gas: 15.00, collision gas: High, ion spray: 2200 v, interface heater: on, declustering potential: 30, entrance potential: 10, collision energy: 10, rolling collision energy). All MS/ MS data were searched against the theoretical spectra of B. anthracis Ames using the MASCOT software (Matrix Science). The search parameters used were: maximum of one missed cleavage by trypsin, fixed modification of oxidized methionine, charge state of + 2 and + 3, an MS tolerance of ± 1.2 Da and an MS/MS tolerance of ± 0.8 Da. Only