|Home | About | Journals | Submit | Contact Us | Français|
Bacillus anthracis is a member of the Bacillus cereus group species (also known as the “group 1 bacilli”), a collection of Gram-positive spore-forming soil bacteria that are non-fastidious facultative anaerobes with very similar growth characteristics and natural genetic exchange systems. Despite their close physiology and genetics, the Bacillus cereus group species exhibit certain species-specific phenotypes, some of which are related to pathogenicity. B. anthracis is the etiologic agent of anthrax. Vegetative cells of B. anthracis produce anthrax toxin proteins and a poly-D-glutamic acid capsule during infection of mammalian hosts and when cultured in conditions considered to mimic the host environment. The genes associated with toxin and capsule synthesis are located on the B. anthracis plasmids, pXO1 and pXO2, respectively. Although plasmid content is considered a defining feature of the species, pXO1- and pXO2-like plasmids have been identified in strains that more closely resemble other members of the B. cereus group. The developmental nature of B. anthracis and its pathogenic (mammalian host) and environmental (soil) lifestyles of make it an interesting model for study of niche-specific bacterial gene expression and physiology.
The etiologic agent of anthrax, Bacillus anthracis, is perhaps the most renowned pathogen of the Bacillus genus. Like all Bacillus species, B. anthracis is a Gram-positive spore-former that is commonly found in the soil. Unlike most other Bacillus species, respiratory, gastrointestinal, or cutaneous entry of B. anthracis spores into mammals can result in systemic infection and lethal disease. Spores are considered to be the predominant form of B. anthracis outside of a host, but upon infection, spores germinate to become vegetative cells that can replicate to high numbers in virtually all body tissues. Death of the host and contact of infected tissues with air results in a return to the spore form of the bacterium. As is true for all Bacillus species, B. anthracis spores are highly resistant to adverse environmental conditions, but it has been proposed that B. anthracis is more dependent than other Bacillus species on sporulation for species survival (Turnbull, 2002). Vegetative cells appear to survive poorly in simple environments including water and bulk soil (Lindeque and Turnbull, 1994; Saile and Koehler, 2006; Turnbull, 2002). Thus, the spore - vegetative cell – spore cycle is essential for the pathogenic lifestyle of this developmental bacterium.
The evolution of B. anthracis and its relatedness to other Bacillus species is intriguing. B. anthracis is a member of the “group 1 bacilli” also known as the “B. cereus group”, which also includes B. cereus, B. thuringiensis, B. mycoides, B. pseudomycoides, and B. weihenstephanensis (Ash and Collins, 1992; Ash et al., 1991; Helgason et al., 2000; Helgason et al., 2004; Lechner et al., 1998; Rasko et al., 2005; Tourasse et al., 2006; Turnbull, 1999). The B. cereus group species exhibit remarkably similar cell structure, physiology, and natural genetic exchange systems but they are distinct with regard to pathogenicity. In large part, species-specific pathogenicity is associated with plasmid content. B. anthracis is distinguished by its virulence plasmids pXO1 (182 kb), which contains the structural genes for the anthrax toxin proteins, pagA (PA), lef (LF) and cya (EF), (Koehler, 2002) and pXO2 (96 kb), which carries the biosynthetic operon for capsule, capBCADE (Candela et al., 2005; Makino et al., 1989; Okinaka et al., 1999a). Although pXO1 and pXO2 are considered to be specific to B. anthracis, there are some reports of rare B. cereus strains harboring plasmids with similarity to these plasmids (Avashia et al., 2007; Hoffmaster et al., 2004; Klee et al., 2006; Rasko et al., 2007). Some of these unusual strains have been isolated from humans and animals that succumbed to an anthrax-like disease. Phenotypes conferred by plasmid genes have been the focus of many investigations comparing B. anthracis to related species. Nevertheless, not all unique attributes of B. anthracis are plasmid-associated. Plasmid-cured strains of the species exhibit some species-specific phenotypes despite the striking degree of sequence similarity and gene synteny associated with the chromosomes (Rasko et al., 2005).
B. anthracis is best distinguished from related species by its ability to synthesize the anthrax toxin proteins and the poly-D-glutamic acid capsule. The bacterium produces these virulence factors in a number of complex and defined media. Optimal synthesis of the toxin proteins occurs during culture at 37°C in defined media containing glucose, as the carbon source, and bicarbonate (Leppla, 1988; Ristroph and Ivins, 1983). Capsule is produced at high levels in defined and complex media, but synthesis is dependent upon the presence of dissolved bicarbonate in the media (Green et al., 1985; Meynell and Meynell, 1964; Thorne et al., 1952). Toxin and capsule synthesis are highest as the culture transitions from exponential to stationary phase (Drysdale et al., 2004; Drysdale et al., 2005; Koehler et al., 1994; Leppla, 1988; Sirard et al., 1994). The “CO2/bicarbonate effect” on toxin and capsule synthesis is long known (Gladstone, 1946; Puziss and Wright, 1954; Thorne, 1993). Steady state levels of toxin gene and cap operon transcripts increase up to 60-fold in response to this signal (Bartkus and Leppla, 1989; Drysdale et al., 2004; Green et al., 1985; Koehler et al., 1994; Sirard et al., 1994). Deletion of the genes encoding a bicarbonate transporter results in decreased uptake of bicarbonate and the absence of toxin gene induction, indicating that induction of the virulence genes requires transport of bicarbonate into cells (Wilson et al., 2008). Nevertheless, the molecular basis for the transcriptional response to the signal remains elusive. Temperature also affects toxin gene expression. Cells growing at 37°C exhibit 4- to 6-fold higher levels of toxin gene transcripts than cells growing at 28°C (Sirard et al., 1994), however, no temperature effect has been reported for capsule synthesis. Given the mammalian body temperature of 37°C and bicarbonate/CO2 concentrations of 15 to 40 mM in the bloodstream (Lentner, 1981), these signals are likely significant during infection.
In addition to production of the anthrax toxin proteins and a poly-D-glutamic acid capsule, classic B. anthracis phenotypes include absence of β-hemolysis on sheep blood agar, absence of phospholipase C activity, lack of motility, susceptibility to penicillin, and sensitivity to the γ bacteriophage (Marston et al., 2006). Nevertheless, for some of these traits there are exceptions. Motile strains and strains that appear to produce flagella have been reported (Klee et al., 2006; Liang and Yu, 1999) although the genomes of sequenced strains, which are nonmotile and presumed nonflagelated, are devoid of some genes commonly associated with flagella (Read et al., 2003). Rare penicillin-resistant isolates have also been reported (Cavallo et al., 2002; Klee et al., 2006; Mohammed et al., 2002; Odendaal et al., 1991). Sequenced B. anthracis strains, which are penicillin-susceptible, harbor two structural genes for β-lactamase proteins (Chen et al., 2003; Read et al., 2003). These bla genes are transcriptionally silent in prototypical strains, but constitutively expressed in a rare highly penicillin-resistant isolate (Chen et al., 2004; Ross et al., in press). Although susceptibility to γ phage is part of the standard protocol used by many diagnostic laboratories for identification of B. anthracis, non-susceptible B. anthracis strains and susceptible B. cereus strains have been described (Abshire et al., 2005; Davison et al., 2005; Klee et al., 2006).
In most growth conditions, B. anthracis cells have distinctive morphology (see figure 1). The vegetative cells are square-ended rods, approximately 1 by 5-8 μm, with the propensity to form chains. The B. anthracis capsule is readily visualized microscopically using India Ink exclusion (Drysdale et al., 2004). B. anthracis generally grows as planktonic cells in liquid media, although formation of pellicles during static incubation and adherence to solid surfaces can occur (Charlton et al., 2007; Lee et al., 2007). B. anthracis forms long serpentine chains during the exponential phase of batch culture. In infected tissues, generally single cells or short chains of 2-3 cells are observed (Lincoln et al., 1965).
B. anthracis generally produces off-white colonies that have irregular edges and a rough “ground glass” appearance when incubated in conditions that are not conducive for capsule formation. The colonies adhere firmly to the agar surface and transfer using a loop or toothpick can result in strings of cells that can be drawn up and remain perpendicular to the agar surface without support. Specific differences in the degree of tenacity have been associated with genotypic groups (Smith et al., 2000). Colonies on solid media that facilitate capsule synthesis appear extremely mucoid, reflecting the large capsule which can be more than 3 μm in thickness (Drysdale et al., 2004).
Upon nutrient starvation, vegetative cells can develop into the dormant spore form of the bacterium. Prolonged culture of B. anthracis in a variety of media can facilitate spore formation (Leighton and Doi, 1971; Schaeffer et al., 1965; Tarr, 1933; Thorne, 1962; Turnbull et al., 2007). Phase microscopy of sporulating cells reveals forespores appearing as oval refractile bodies of 1- to 1.5-μm in diameter, that are central or subterminal within the mother cell and do not cause significant swelling of the mother cell (Battisti et al., 1985). A distinguishing characteristic of the spores of B. anthracis and related species is the exosporium, an outermost layer that is not present on spores of many other Bacillus species. The exosporium has been visualized using electron crystallography and atomic-force microscopy (Ball et al., 2008; Chada et al., 2003).
B. anthracis is a facultative anaerobe that can multiply readily in a variety of common laboratory media using multiple sugars and amino acids as carbon sources (Charlton et al., 2007; Puziss and Wright, 1959). Most studies of carbohydrate metabolism have been performed to determine relationships between sugar utilization and toxin synthesis. The precise nutritional requirements of B. anthracis are not clear and appear to vary between strains. Methionine and thiamine are required for growth, but growth in glucose-salts medium containing methionine and thiamine requires supplementation with multiple amino acids, which most likely serve as sources of nitrogen (Thorne, 1993). Hydrolyzed casein medium containing glucose, adenine, uracil, thiamine, tryptophan, cysteine, glycine, and various salts is used frequently for investigations of B. anthracis physiology and gene expression (Thorne and Belton, 1957). Defined medium comprised of glucose, thiamine, uracil, adenine, guanine, calcium and multiple amino acids are also used (Brewer et al., 1946; Ristroph and Ivins, 1983). The most minimal defined medium, XO, contains glucose, ferric chloride, thiamine, glutamic acid, glycine, methionine, proline, serine, threonine, ammonium sulfate, magnesium sulfate, manganese sulfate, potassium phosphate, and sodium citrate (Hoffmaster and Koehler, 1997). The optimal growth temperature for B. anthracis is 37°C and the bacterium is unable to grow at temperatures above 43°C. Cell doubling times during growth in complex media at 37°C range from 30- to 60-min (Dai and Koehler, 1997; Passalacqua et al., 2007; Yang and Miller, 2008).
Iron acquisition by B. anthracis has been the subject of recent investigations. Like most mammalian pathogens, B. anthracis has multiple means of obtaining iron that is tightly bound to carrier and storage proteins in host tissues. In response to low iron availability, B. anthracis secretes the siderophores bacillibactin and petrobactin (Cendrowski et al., 2004; Koppisch et al., 2005; Koppisch et al., 2008). These ferric iron chelators bind iron with very high affinity and transport it into the bacterium via substrate-binding proteins (SBPs) and other ABC transporter components. B. anthracis can also scavenge iron from heme and heme-containing proteins, including hemoglobin, via iron-regulated surface determinant (Isd) proteins (Gat et al., 2008; Maresso et al., 2006; Maresso et al., 2008; Skaar et al., 2006). The three B. anthracis proteins that bind heme, IsdX1 (also known as IsdJ), IsdX2 (also known as IsdK), and IsdC (Maresso et al., 2006; Maresso et al., 2008) contain NEAT (near iron transporter) domains that are associated with binding to iron-containing ligands (Andrade et al., 2002). IsdC is bound covalently to the cell envelope by the transpeptidase SrtB while IsdX1 and IsdX2 are found in cell-free culture supernates (Skaar et al., 2006).
Bacteriophage of B. anthracis comprise a genetically diverse group, including members of the Siphoviridae, Podoviridae, Myoviridae, and Tectiviridae families (Ackermann et al., 1978; Inal and Karunakaran, 1996; Minakhin et al., 2005; Nagy et al., 1976; Schuch and Fischetti, 2006; Sozhamannan et al., 2008; Walter and Baker, 2003). The γ phage is the most well-studied B. anthracis phage due to its diagnostic use and more recently, the proposed use of the lysin enzyme PlyG for decontamination (Fischetti, 2006; Kikkawa et al., 2008; Schuch et al., 2002). γ is morphologically similar to the Siphoviridae family, possessing a long contractile tail and an icosohedral head. The double-stranded DNA phage has a 37-kb genome, encoding 53 ORFs (Schuch and Fischetti, 2006). γ phage is actually a lytic variant of the temperate phage, W, a phage first identified in an atypical “B. cereus” isolate (Brown and Cherry, 1955; Fouts et al., 2006; McCloy, 1958). The γ phage is lytic on capsulated and noncapsulated B. anthracis cells, unlike phage W which is relatively inefficient for lysis of capsulated cells. Although γ is generally considered to be B. anthracis-specific, some nonsusceptible B. anthracis strains and susceptible non-B. anthracis isolates have been reported (Abshire et al., 2005; Davison et al., 2005; Klee et al., 2006; Schuch et al., 2002). Susceptible strains produce a cell-wall anchored protein, GamR (Gamma Receptor), which is associated with phage binding (Davison et al., 2005). Interestingly, there is some evidence of lysogenic conversion by γ or related phage. The γ phage DNA sequence contains a gene encoding the fosphomycin-resistance protein Gp41 and it has been suggested that resistance to this soil antibiotic may play a role in the survival of vegetative B. anthracis cells outside of the host (Schuch and Fischetti, 2006).
Bacteriophage with generalized transducing capability have served as useful tools for creation of recombinant B. anthracis strains. The transducing phage CP-51 and CP-54 were originally isolated from the soil as phage that could infect B. cereus (Thorne, 1968; Thorne, 1978). Transducing phage TP-21 was originally identified in B. thuringiensis subsp. kurstaki as a plasmid and shown subsequently to exist as a plasmid prophage (Thorne, 1993). The host ranges of CP-51, CP-54, and TP-21 extend to multiple strains of B. anthracis, B. cereus, and B. thuringiensis (Ruhfel et al., 1984; Thorne, 1968; Thorne, 1978; Thorne, 1993; Yelton and Thorne, 1970). CP-51 and CP-54 have a similar, but not identical structure and are serologically related (Thorne, 1978; Thorne and Holt, 1974; Yelton and Thorne, 1970). Despite the value of these transducing phage for genetic studies, they remain relatively uncharacterized.
Finally, genome sequence analyses have revealed the presence of apparent intact and defective prophage (Inal and Karunakaran, 1996; Read et al., 2003; Schuch and Fischetti, 2006; Sozhamannan et al., 2006). Potential phenotypes conferred by the prophage have not been discerned, but the prophage sequences can facilitate distinction of B. anthracis from other B. cereus group species.
Since the discovery of pXO1 and pXO2 in the 1980s (Green et al., 1985; Mikesell et al., 1983; Uchida et al., 1985), the plasmids have served as the hallmark genetic feature of B. anthracis. Yet given the safety and regulatory concerns related to use of fully virulent B. anthracis strains in the laboratory, many investigators employ attenuated strains harboring only one of the two virulence plasmids. “Sterne” type strains, so named because of a strain first described by M. Sterne in 1939 (Sterne, 1939), carry pXO1 but are missing pXO2. These strains include Sterne, 34F2, STI, Weybridge, and others. “Pasteur” type strains (Mikesell et al., 1983), such as 6602 and 4229, harbor pXO2 but are pXO1-negative. pXO1 and pXO2 are stably maintained by B. anthracis growing in vivo during infection. During culture in the absence of any apparent selective pressure, spontaneous loss of pXO2 occurs more frequently than loss of pXO1 (Green et al., 1985; Thorne, 1985). The stability of pXO1 may be related to a plasmid-encoded type 1 topoisomerase (Fouet et al., 1994). B. anthracis can be cured of either plasmid during culture at 43°C or in the presence of sub-lethal concentrations of novobiocin (Chen et al., 2004; Green et al., 1985; Mikesell et al., 1983; Thorne, 1985).
The toxin and capsule genes are the most thoroughly investigated plasmid genes, but functions of some other plasmid genes and structural features of the plasmids have been reported. DNA sequence annotation of pXO1 indicates 203 open reading frames (ORFs) (Okinaka et al., 1999b). An 8.7-kb element resembling a class II cointegrative transposon lies within a 44.8-kb pathogenicity island on the plasmid (Van der Auwera and Mahillon, 2005). The element, TnXO1, carries genes for a transposase and a site-specific recombinase in addition to a germination operon, gerX, that is associated with germination in phagocytes (Guidi-Rontani et al., 1999). The pathogenicity island also includes pagA, lef, and cya, the structural genes for the toxin proteins; atxA, a major virulence gene regulator (Koehler et al., 1994; Uchida et al., 1993); bslA, encoding an adhesion protein located in the S-layer of vegetative cells (Kern and Schneewind, 2008); and other ORFs. Comparison of the pXO1 DNA sequences of multiple B. anthracis strains reveals one strain in which the pathogenicity island is inverted (Thorne, 1993), but overall, the pXO1 DNA sequence appears to be stable.
Investigations aimed at identifying the pXO1 origin of replication (ori) and genes required for plasmid replication have resulted in dissimilar models (Pomerantsev et al., 2009; Robertson et al., 1990; Tinsley and Khan, 2006). A 5-kb region of pXO1 cloned in an E. coli vector was reported to facilitate replication of the plasmid in B. anthracis (Tinsley and Khan, 2006). The recombinant plasmid contains a 158-bp region, considered to be an ori, and four ORFs, pXO1-43 to pXO1-46. RepX, encoded by pXO1-45, belongs to the tubulin family of cytoskeletal proteins that are associated with cell division and DNA segregation. Purified RepX can polymerize in a GTP-dependent manner and has been shown to form polymers in vivo in B. anthracis, properties consistent with a role in plasmid segregation and partitioning (Akhtar et al., 2009). More recently, another group described an 8.8-kb mini-pXO1 plasmid that replicates in B. anthracis and B. cereus (Pomerantsev et al., 2009). The mini-replicon does not contain the previously implicated pXO1 sequences. Rather, it carries two ORFs, pXO1-14 and pXO1-16, which appear to be essential for replication. ORF14 is predicted to encode a protein with a helix-turn-helix motif. ORF16 is the putative replication initiator Rep protein. The region upstream of pXO1-16 bears the attributes of the ori from theta-replicating plasmids, which include an exclusively A+T-containing segment, five 9-bp direct repeats, an inverted repeat, and a σA-dependent promoter.
Sequence annotation of the smaller B. anthracis virulence plasmid, pXO2, reveals 110 ORFs, including the capBCADE operon that encodes proteins required for capsule synthesis, assembly, and transport (Candela et al., 2005; Green et al., 1985; Makino et al., 1989; Okinaka et al., 1999a), the cap regulatory genes acpA and acpB (Drysdale et al., 2004; Vietri et al., 1995), and an L-alanine amidase gene amiA (Mesnage and Fouet, 2002; Rollins et al., 2005). A 2.4-kb region of pXO2 cloned into an E. coli vector is sufficient for replication in B. anthracis (Tinsley et al., 2004). The cloned DNA carries a putative 60-bp ori located immediately downstream of a gene encoding the replication initiation protein, RepS. Recombinant RepS protein binds to a DNA fragment corresponding to the ori. Like pXO1, pXO2 appears to replicate via a theta-type mechanism with replication proceeding unidirectionally from the origin of replication (Tinsley et al., 2004).
Plasmids with high DNA sequence similarity and synteny to pXO1 and pXO2 have been identified in strains of other B. cereus group species (Hoffmaster et al., 2004; Hu et al., 2006; Hu et al., 2009; Rasko et al., 2007; Van der Auwera and Mahillon, 2008). In most reported cases, these pXO1-like and pXO2-like plasmids have not been shown to carry the pXO1 pathogenicity island or the cap region of pXO2. Notable exceptions include one clinical B. cereus isolate associated with a human illness resembling anthrax, that carries a pXO1-like plasmid with 99.6% similarity to pXO1 (Hoffmaster et al., 2004). Also, clinical isolates from wild great Apes that died of an anthrax-like disease do not bear classic phenotypes associated with B. anthracis, but harbor pXO1- and pXO2-like plasmids and produce protective antigen and capsule. Multilocus sequence typing of these strains indicates that they are closely-related to B. anthracis and two uncommonly virulent B. cereus and B. thuringiensis isolates (Klee et al., 2006).
Inter- and intra-species transduction and conjugation methods were crucial for early investigations of B. anthracis physiology and genetics. Before the advent of molecular biology methods and DNA sequencing, similarities in the B. cereus group species chromosomes were demonstrated by co-transduction of linked markers. Mating experiments in which conjugative plasmids transferred between species, and in some cases mobilized pXO1, pXO2, and other non-conjugative plasmids, revealed plasmid-specific phenotypes.
Plasmids pXO1 and pXO2 can be transferred in intra- and inter-species matings by conjugative plasmids originally found in B. thuringiensis. B. cereus group species harboring plasmids pXO12 and pXO14 can mobilize small selectable plasmids, such as the tetracycline-resistance plasmid pBC16, in inter-species matings with frequencies up to 10-2 transconjugants per donor (Battisti et al., 1985; Reddy et al., 1987). Co-transfer of the conjugative plasmid with the selected plasmid approaches frequencies of up to 80%. B. anthracis transconjugants carrying pXO12 or pXO14 can serve as donors of pXO1 and pXO2 (Green et al., 1985; Green et al., 1989; Reddy et al., 1987). Transfer of the virulence plasmids appears to occur via conduction. Co-integrates of pXO12 and pXO1 involving the insertion element IS4430 form in the donor strain and are resolved, sometimes improperly, in ex-conjugants (Green et al., 1989). pXO12 can also transfer pBC16 between B. anthracis strains co-cultured in a plant/soil system, suggesting that conjugative transfer of the plasmids may occur in nature (Saile and Koehler, 2006).
Although pXO1 and pXO2 are not self-transmissible, it is notable that they carry homologues of genes encoding type IV secretion system (T4SS) machinery, a system that mediates delivery of protein and DNA substrates from donors to recipients during conjugative transfer. However, the T4SS gene homologues on pXO1 display low levels of homology to known T4SS genes and are physically scattered throughout the plasmid (Grynberg et al., 2007). The “transfer region” of pXO2, is very similar to that of pXO2-like conjugative plasmids pAW63 and pBT9727 originally found in B. thuringiensis strains (Van der Auwera et al., 2005). Yet some of the transfer genes on pXO2 carry nonsense mutations or frameshifts (Hu et al., 2009).
The most widely employed transducing bacteriophage of the B. cereus group is CP-51. The phage was instrumental in demonstrating that pXO2 is required for capsule synthesis (Green et al., 1985). Propagation of CP-51 on pXO1- pXO2+ Pasteur-type strains results in phage lysates that can transduce pXO2 to pXO2- B. anthracis and B. cereus at a frequency of about 10-8 transductants per plaque-forming unit. Cap+ isolates can be selected readily from transduction mixtures using phage CP-54. CP-54, which can infect B. anthracis and B. cereus, lyses non-capsulated cells, but does not adsorb to capsulated cells (Green et al., 1985). CP-51-mediated transduction of smaller plasmids and chromosomal genes occurs at higher frequencies of up to 10-5 transductants per PFU respectively (Ruhfel et al., 1984). The 60-MDal size estimate for CP51 DNA reported by Yelton and Thorne (Yelton and Thorne, 1971) is in agreement with the lower frequency of pXO2 transduction and likely indicates that the size of the plasmid (96 kb) approaches the limit for DNA packaging by the phage.
There are no reported methods for generation of naturally competent cells of B. anthracis, but bifunctional plasmid vectors containing replication origins for Gram-negative and Gram-positive species can be introduced into B. anthracis using electroporation and an E. coli - B. anthracis conjugal mating system. Selectable markers commonly used in B. anthracis include genes encoding erythromycin, kanamycin-, and spectinomycin-resistance (Koehler, 2002). Highest electroporation frequencies are achieved when the plasmid DNA is isolated from DNA methyltransferase-deficient (dam) E. coli strains or from B. subtilis 168, a Bacillus strain that lacks adenine methylation activity (Marrero and Welkos, 1995). In the conjugal system, shuttle vectors containing the transfer origin of the IncP plasmid RK2 are mobilized by self-transmissible IncP plasmids that are co-resident in the donor strain (Cataldi et al., 1990; Pezard et al., 1991).
Numerous methods for generating mutations in the B. anthracis genome have been published. Integrative vectors, carrying B. anthracis DNA sequences but lacking a replication origin functional in B. anthracis, can be transferred to B. anthracis using the RK2 system with selection for plasmid-encoded antibiotic resistance (Cataldi et al., 1990; Pezard et al., 1991). Electroporation of unstable and temperature-sensitive plasmids into B. anthracis facilitates insertional mutagenesis and allelic replacement (Bongiorni et al., 2007; Chen et al., 2004; Fisher and Hanna, 2005; Janes and Stibitz, 2006; Saile and Koehler, 2002). The Cre-loxP system, first employed in E. coli to excise antibiotic resistance genes after their introduction into specific loci (Palmeros et al., 2000), has been developed for use in B. anthracis (Pomerantsev et al., 2006) and can be used to generate very large deletions in the virulence plasmids (Pomerantsev et al., 2009). Random mutagenesis employing mini-Tn10 and mariner-based transposon systems have proved useful in genetic screens for germination, toxin gene expression, and other phenotypes (Barua et al., 2009; Day et al., 2007; Tam et al., 2006; Wilson et al., 2007).
A pXO1-encoded protein, AtxA (for anthrax toxin activator) is a global regulator of transcription in B. anthracis (see figure 2). AtxA has a strong positive effect on the toxin genes, pagA, cya, and lef, and the capsule biosynthetic operon, capBCADE, and varying control of many other genes on the B. anthracis plasmids and chromosome (Bourgogne et al., 2003; Drysdale et al., 2004; Guignot et al., 1997; Hoffmaster and Koehler, 1997; Sirard et al., 2000; Uchida et al., 1997). In agreement with strong regulation of the toxin and capsule genes, an atxA-null mutant is highly attenuated for virulence in a murine model for anthrax (Dai et al., 1995). The molecular mechanism for AtxA protein function is unknown; there are no reports of direct interaction of AtxA with any of its target genes. AtxA exerts its affect on capBCADE expression via control of acpA (for anthrax capsule activator), a pXO2 gene located 5′ of the cap operon (Drysdale et al., 2004; Guignot et al., 1997; Uchida et al., 1997; Vietri et al., 1995). Transcriptional read-through of capBCADE results in increased expression of the 3′ proximal gene acpB, which affects cap operon expression in a positive feedback loop (Drysdale et al., 2005). Interestingly, AtxA, AcpA, and AcpB have sequence and functional similarity, yet while AtxA has far-reaching effects on B. anthracis gene expression, AcpA and AcpB effects are much more limited. PagR is another regulator that functions downstream of AtxA. The pagR gene is cotranscribed with the protective antigen gene in the AtxA-regulated pagAR operon (Hoffmaster and Koehler, 1999). PagR is a weak repressor of pagAR and thus serves in a negative feedback loop. In addition to autogenous control of pagAR, PagR regulates expression of the chromosome genes sap and eag which encode surface proteins of vegetative cells. PagR is a member of the ARSR family of transcriptional regulators and binds DNA specifically in the promoter regions of its target genes to repress transcription, in the cases of pagAR and sap, and activate transcription, in the case of eag (Mignot et al., 2003).
Mechanistic predictions for AtxA function have been made using in silico protein modeling programs. AtxA is an approximately 56-kD protein with a calculated isoelectric point of 9.5. Winged-helix and helix-turn-helix motifs are found at the amino-terminus of the protein, suggesting that AtxA has DNA-binding activity, however specific binding of AtxA to the promoter regions of target genes has not been shown. Although apparent transcriptional start sites and cis-acting sequences have been established for the highly AtxA-regulated toxin genes, sequence similarities are not apparent in the control regions. In silico modeling and in vitro gel mobility experiments indicate significant curvature associated with DNA in AtxA-regulated promoters, suggesting that the topology of promoter DNA may play a role in AtxA-mediated control of gene expression (Hadjifrangiskou and Koehler, 2008).
Two motifs for phosphotransferase system regulation domains (PRDs) are located centrally in the AtxA amino acid sequence and have been implicated in AtxA function (Tsvetanova et al., 2007). The phosphotransferase system (PTS) is responsible for uptake of specific sugars by many bacterial species (Gorke and Stulke, 2008) and PRDs are common to some bacterial proteins that respond to carbohydrate availability. Specific histidine residues in the apparent PRDs of AtxA can be phosphorylated in vivo; phosphorylation of H199 in PRD1 has been reported to positively affect AtxA activity, while phosphorylation of H379 has a negative impact on AtxA function (Tsvetanova et al., 2007). It is intriguing that carboxy-terminal region of AtxA bears some similarity to part of the EIIB protein of the PTS. EIIB phosphorylates sugars passing through the EIIC channel (Deutscher et al., 2006; Zeng and Burne, 2009). Thus, the presence of PRDs and an EIIB-like region in AtxA suggests a relationship between AtxA activity and sugar uptake in B. anthracis.
Certain signals and trans-acting factors have been reported to affect atxA gene expression. AtxA protein and atxA transcript levels are 5- to 6-fold greater in cultures grown at 37°C compared to 28°C (Dai and Koehler, 1997). Control of atxA expression has also been tied to cytochrome c biogenesis. Mutants harboring insertions in genes encoding two heme-dependent small c-type cytochromes, CccA and CccB, and other components of the cytochrome c biogenesis pathway show decreased atxA expression (Wilson et al., 2009). The influence of c-type cytochromes on atxA expression indicates a potential connection between cellular redox state and virulence gene regulation in B. anthracis. The only trans-acting factor known to interact directly with the atxA promoter to affect transcription is AbrB. AbrB binds to DNA sequences overlapping the putative SigA recognition sequence for the major transcriptional start of the gene (Strauch et al., 2005), suggesting that the regulatory protein competes with RNA polymerase for promoter binding. During growth in batch culture, AbrB represses atxA transcription, exerting the greatest effect during exponential growth (Saile and Koehler, 2002). The well-characterized AbrB homologue from B. subtilis, is a “transition state regulator” that affects timing of expression for multiple genes in that species. Thus, it appears that AbrB is at least in part, responsible for growth phase control of atxA expression (Saile and Koehler, 2002).
Research efforts concerning B. anthracis have focused on attributes related to pathogenicity. Certainly, specific investigations of B. anthracis physiology and genetics as related to virulence will further progress toward development of new therapeutics and vaccines. However, independent of its “Select Agent” status, B. anthracis can be viewed as an outstanding model organism for studies of microbial ecology, evolution, cell development, and host-pathogen interactions. Further investigations, including functional and comparative genomic studies, will continue to increase knowledge of niche-specific lifestyles of this interesting bacterium.
The author thanks J. Dale, T. Hammerstrom, K. Pflughoeft, J. Rall, and J.-H. Roh for critical reading of the manuscript, and J. Hutt, J. Lovchik, and C. Lyons for help with figure design. Work in the author's laboratory is supported by Public Health Service grant AI33537 from the National Institutes of Health.