Recent advances in the rapid detection of Bacillus anthracis
Steven M. Levine, Yi-Wei Tang, Zhiheng Pei
Bacillus anthracisis a Gram-positive, spore-forming rod that causes anthrax. Culturebased methods are the gold standard for the identification of virulent B. anthracis strains but these require days for completion. The experience from the anthrax attacks in September and October of 2001 revealed the urgent need for methods that can rapidly detect this pathogen with high reliability. Because of the extensive homology among non-anthrax Bacillus sp. at the chromosomal level, rapid detection of virulent B. anthracis strains depends on markers associated with the two plasmids, pXO1 and pXO2, responsible for its virulence. Genes encoding toxins and capsules have been used as markers for pXO1 and pXO2, respectively, in methods that are designed for rapid and sensitive detection of B. anthracis DNA, such as real-time polymerase chain reaction,direct liquid phase hybridization, and DNA microarrays. A variety of platforms can be modified to suit the needs for rapid detection of B. anthracis antigens, but little is known about plasmid-encoded antigens expressed in spores. Future studies should be aimed at detecting markers for pXO1 and pXO2in viable spores.
Keywords: Bacillus anthracis, DNA, antigen, polymerase chain reaction, rapid diagnosis, bioterrorism
Though anthrax is a disease largely confined to the herbivore population, all mammals, including humans are susceptible. The disease is initiated by the entry of spores into the host body which can be transmitted via contact with sick animals or their products. Human anthrax has three major clinical forms: cutaneous, gastrointestinal, and inhalational (pulmonary). Each form can progress to fatal systemic anthrax. Cutaneous anthrax is recognizable as a small pimple that develops, within a few days, into a painless black eschar. In both the inhalational and gastrointestinal forms, early diagnosis is difficult. Initial presentation for the former will mimic bronchopneumonia, and for the latter, will mimic gastroenteritis, and eventually progress to hemorrhagic enteritis. Both diseases quickly progress to a systemic form that can become treatment-resistant and rapidly fatal secondary to sepsis and respiratory failure.
The USA experienced its first inhalational anthrax attack on September 27, 2001, immediately following the September 11, 2001 terrorist attacks. The effective delivery of anthrax spores through the mail system was unexpected and cause for great concern. There were numerous exposures, 19 infections, and five fatalities. 10 000 people in the USA took a 2-month course of antibiotics following possible exposure to this agent. Indeed, a 1970 report of the World Health Organization (WHO) estimated that the release of 50 kg of anthrax spores along a 2-km line downwind of a city of 500 000 would result in 95 000 deaths, and a 1993 report of the US Congressional Office of Technology Assessment predicted that the release of 100 kg of anthrax spores downwind of Washington, D.C., would result in 130 000 to 3 million deaths.
If the target of a biological attack is civilian, hospital-based clinical microbiology laboratories will be on the frontline for recognition of possible bioterrorism agents including Bacillus anthracis. The Laboratory Response Network (LRN), containing hospital-based clinical microbiology laboratories, local, state and federal health agencies, is expected to play a key role in the detection and identification of the agent. For this reason, it is important that the clinical microbiologist be familiar with the characteristics of these agents and the technologies available for their detection and identification. This article includes a review of molecular techniques associated with the detection and identification of B. anthracis. Bacillus anthracis is a Gram-positive, spore-forming rod. At the Centers for Disease Control and Prevention (CDC) and its Learning Resource Network, presumptive identification to the genus level ( Bacillus species) requires Gram stain and colony identification. Presumptive identification to the species level ( B. anthracis ) requires tests for motility, lysis by gamma phage, capsule production, hemolysis, wet mount, and Malachite green staining for spores. Confirmatory identification of B. anthracis carried out by the CDC may include phage lysis, capsular staining, and direct fluorescent antibody (DFA) testing on capsule antigens and cell wall polysaccharides.
DNA markers for B. anthracis
Bacillus anthracis is a member of the Bacillus cereus group, which also includes B. cereus, B. mycoides, B. thuringiensis, B. weihenstephanensis and B. pseudomycoides. Strains with <3% difference between their 16S rRNA genes, are in general, considered the same species. However, differences between 16S rRNA genes for some Bacillus species, such as B. anthracis, B. cereus, and B. thuringiensis, are <1%. Currently, there is no known chromosomal marker that can reliably distinguish between them. Variable region 1 of the rpoB gene (Ba813), a 152-bp chromosomal fragment was originally thought to be specific to B. anthracis but when strains from closely related Bacillus species were screened, this marker was found in strains of non-anthrax B. cereus. VrrA and intergenic spacer regions for the 16S–23S rDNA are most suitable for genotyping B. anthracis strains when using sequencing-based techniques, but have not yet been adapted for rapid pathogen detection. Because of the marked similarity between the chromosomes of B. anthracis and non-anthrax Bacillus sp, it is the extrachromosomal elements that must be targeted to distinguish B. anthracis from other closely related species.
Fully virulent B. anthracis strains contain two plasmids, pXO1 and pXO2, encoding toxins and capsule, respectively. pXO1 is 181 654 bp in length and composed of 143 open reading frames (ORF). It encodes two toxins composed of three proteins, the protective antigen (PA), lethal factor (LF), and edema factor (EF). The toxins follow the general model of A-B toxins. The ‘A’subunit is the active subunit which acts intracellularly and is responsible for the toxic effect. The ‘B’ subunit is for binding the toxin molecule to the cell surface. In B. anthracis, PA is the B moiety and LF/EF, the alternative A moieties. PA plus LF forms the lethal toxin and PA plus EF forms the edema toxin [17,18]. A large majority of the 143 ORF have significant homologous sequences in at least one non-anthracis species, likely associated with large extra-chromosomal elements in B. thuringiensis (327 kb) and B. cereus (341.8 kb). Only 21 of the ORF tested were potentially specific for B. anthracis, including the three toxin genes, six additional ORF in the pathogenic island, and 12 ORF from the remainder of the plasmid. These genes or gene products might be unique to B. anthracis and are candidate targets for DNA and/or antigen-based diagnostic tests. pXO2 has been completely sequenced at Los Alamos National Laboratory (GenBank accession number AF188935). It is 96 231 bp in length and composed of 85 ORF. Genes required for the synthesis of the poly-y-D-glutamic acid capsule, including capA, capB, capC, and dep are encoded by pXO2. The main role of the capsule is to protect B. anthracis from phagocytosis by macrophages. Although the uniqueness of pXO2 sequences have not been thoroughly examined, published primers or probes targeting the capB or capC genes amplified homologous sequences in at least one unknown soil organism. The extensive homology found between pXO1 and plasmids of closely related B. cereus species dictates the need for a thorough examination of pXO2-like sequences in non-anthrax Bacillus spp before pXO2 anthrax- specific genes or gene products are chosen as targets for diagnostic assays.
Antigen markers for B. anthracis
Detection of B. anthracis antigens forms the basis for rapid detection using immunological methods. Because of the close relationship between B. anthracis and non-anthrax bacilli, the first task in the development of a rapid antibody-based test is to identify an antigen that is specific for B. anthracis. Because anthrax is most commonly spread in spore form, the selection of a target antigen should give preference to antigens that are located on the spore surface. If a non-spore antigen is chosen, the suspected materials must undergo a lengthy germination process in order to express the antigen. An internally located spore antigen needs to be released mechanically or chemically
before the assay can be performed.
The outermost surface of B. anthracis spores is covered by a protein-rich layer called the exosporium. The exosporium contains five proteins, including alanine racemase, iron/manganese superoxide dismutase, BclA, BxpA, and BxpB. All five proteins are nearly identical to their counterparts in B. cereus. Serological studies have indicated extensive cross-reactivity between spores of B. anthracis and non-anthracis bacilli, but there is at least one B. anthracis-specific spore antigen/epitope. PA is a major soluble antigen produced by vegetative cells and can also be found in the spore coat by immunogold electron microscopy. Antibodies to recombinant PA promote uptake of B. anthracis spores by phagocytes and inhibit spore germination, suggesting that PA might be located on the surface of spores. LF and EF genes are unique for B. anthracis at the gene level, however, unlike PA, their gene products have not been thoroughly examined for antigenic cross-reactivity with other non-anthrax bacilli. LF and EF are produced by vegetative B. anthracis, but their association with anthrax spores is undetermined.
Capsule, surface layer, and cell wall antigens are not ideal targets for anthrax detection. The capsule poly-y-D- glutamic acid antigen is identical to that of B. licheniformis. The two surface array proteins, SAP (surface array protein) and EA1 (extractable antigen 1), contain epitopes shared with those of non-anthrax bacilli, as well as epitopes unique to B. anthracis. However, the unique epitopes are completely covered by the capsule, and therefore not accessible by antibodies in encapsulated vegetative cells. B. anthracis galactose-N-acetyl-D-glucosamine polysaccharide is another cell wall antigen that is antigenically related to that of B. cereus.
The suitability of DNA and antigens for use as markers in detection of B. anthracis is summarized in Table 2. In general, DNA markers need to be released from spores and amplified before they can be detected due to their intracellular location and low copy numbers. In contrast, each cell usually contains many copies of a particular antigen, and surface antigens can be directly accessed without the need for spore lysis. Because of the extensive homology between B. anthracis and B. cereus, specific detection of anthrax spores heavily depends on markers present in or encoded by the plasmids. While there are plenty of DNA markers for the plasmids, little is known about the expression of plasmid-encoded antigen markers in spores.
Methods for rapid detection of B. anthracis
Real-time assays that can rule out the presence of B. anthracis spores in environmental samples can eliminate the need for further investigation of suspicious contaminations, personnel exposures, prophylactic antibiotic use, laboratory investigation, and shut-down of routine operations. Conceptually, sensors would best suit this aim. The ideal sensor systems should detect and identify anthrax spores with high sensitivity in nearly real time, as compared to classic microbiological methods that take one or more days to perform. Sensors can be grouped into two types: chemical and biological sensors. One prototype of chemical sensors is based on mass spectrometry. Mass spectrometry involves the ionization and fragmentation of molecules into characteristic pieces. Positively charged fragments are then accelerated through an electric field that sorts the fragments by mass and charge. By analysing characteristic information on the structure and molecular weight of the material, mixtures of closely related bacterial species can be identified. Because some strains of B. anthracis are nearly identical to B. cereus and B. thuringiensis except for the two plasmids, success of a chemical sensor will depend on its ability to recognize patterns generated by the plasmids of B. anthracis. Currently, chemical sensors for the detection of B. anthracis are still in early stages of development. This review will focus on biological sensors.
Biological sensors can be roughly classified into two categories: amplification-based and tissue-based. The most widely used amplification-based sensors detect B. anthracis
by amplification of its nucleotide sequence using polymerase chain reaction (PCR). The prototype of biological tissue-based sensors uses antibodies to detect antigens.
Detection of B. anthracis DNA
Amplification of DNA Nucleic acid-based assays that are capable of detecting the presence of B. anthracis include: the ligase chain reaction, the QB replicase-based system, the nucleic acid sequence-based amplification (NASBA) system, strand displacement amplification (SDA), and PCR. While each of these techniques has benefits and drawbacks, PCR is most commonly used for detecting DNA.
Conventional PCR is composed of three phases: DNA template preparation, nucleic acid amplification, and post-amplification analysis. The whole process may require 1–2 days to complete. Although development of rapid PCR tests are beneficial to patient care, the main driving force for shortening the time required comes from the demand of rapid detection and identification of biocrime agents on scene. With the advent of real-time PCR, significant progress has been made in shortening all three phases of traditional PCR.
The sample preparation phase may be completely eliminated because there is a trace amount of free bacterial DNA attached on the spore surface that can directly serve as templates. Larger amounts of DNA may be released from spores by heating at 96°C for 15 s or by microsonication for 30 s . PCR amplification can detect as few as 100 spores, or a similar number of vegetative cells prepared by germination, the classic method for spore lysis. However, spore lysis by germination requires heat activation and subsequent germination, which takes about 60 min. Depending on the nature of the material, PCR inhibitors may be present and need to be removed by time-consuming purification, usually by commercial kits.
With improvements in engineering designs, new generations of real-time PCR reduce amplification time, from hours required for conventional PCR, to minutes. Traditional microfuge tubes are replaced with glass capillaries that have a high surface area to volume ratio and heating/cooling blocks are replaced by air-phase thermoconduction. These changes allow rapid tempera- ture transition at a rate of 20° C/s compared with the traditional 2°C/s. Amplification time can be further reduced by designing primers that can anneal with templates at 72°C, the commonly used PCR elongation temperature, thereby eliminating one of the classic three steps of PCR.
The real-time PCR machine has a built-in fluorimeter designed to detect amplification products at the end of each cycle. Because the sample is analysed in situ in a sealed PCR tube, contamination by PCR products is virtually eliminated. There are two classes of fluorescent dyes. Sequence-independent dyes are DNA intercalators that bind non-specifically to double-stranded DNA, including primer dimers and PCR products. Confirmation of specific products requires melting curve analysis at the end of the amplification phase rather than in real time. Sequence-specific probes can reduce non-specific signals to near zero and allow real-time monitoring and detection of specific amplification products at the end of each cycle. A number of fluorescent resonant energy transfer probes are available, including TaqMan, molecular beacons, and dual hybridization probes. With these modifications, the whole process, including all three phases, can produce a positive signal from 500 bacterial cells in 7 min or 1–5 cells in approximately 1 h.
Real-time PCR is also suitable for the detection of multiple target sequences. In the Mayo-Roche Rapid Anthrax Test, two genes, the pagA and capB are targeted as markers for the presence of pXO1 and pXO2, respectively. Both genes are amplified in multiplex using Roche’s LightCycler machine. Modified protocols using commercially available kits have been shown to detect B. anthracis from human tissue, even subsequent to patient antibiotic prophylaxis. In the real-time PCR assay developed by CDC for B. anthracis,a chromosomal marker is included in addition to the two plasmid markers [http://www.bt.cdc.gov/agent/anthrax/faq/labtesting.asp]
Direct liquid-phase hybridization
DNA molecules can also be detected by hybridization, without the need for amplification. A prototype of the amplification-independent methods is the well-known Southern blot, which is insensitive and takes days to complete. Newer versions of these types of assays are performed in liquid so that the blotting, blocking, and washing steps are eliminated. A method using liquid- phase hybridization coupled with a laser-based fluorescence detector was developed at Los Alamos National Laboratory . The method consists of using two nucleic acid probes complementary to different sites on a target DNA. The two probes are each labeled with different fluorescent dyes. When mixed with a sample containing the target DNA, the two probes hybridize to their respective binding sites on the same target DNA molecule. The sample is then analysed by a laser-based fluorescent system capable of detecting a single fluorescent molecule on two different wavelength channels simultaneously. Since the probes are bound to the same molecule, their signals appear simultaneously. Thus, coincident detection of both dyes provides the necessary specificity to detect non-amplified target DNA. If the target is not present, only uncorrelated events originating from both probes will be observed at either channel. This assay detected about 3000 copies of B. anthracis pXO2 plasmid. The test requires DNA extraction, 30 min hybridization, and 200 s of post-amplification analysis. Direct detection of target DNA molecules is useful in situations where they cannot be amplified due to the presence of PCR inhibitors.
Biological agents of mass destruction include many pathogens or toxins other than B. anthracis, therefore the development of technology that detects multiple pathogens is necessary. Researchers at Lawrence Livermore National laboratory have developed the Multi-Pathogen Identification (MPID) microarray to detect a number of pathogenic organisms including B. anthracis. The first two steps of this system are similar to PCR, including sample DNA preparation and amplification. The post-amplification analysis was performed by hybridization with 53 660 overlapping oligonucleotide probes which were synthesized as a high-density microarray on a solid surface. The probes are complimentary to DNA sequences of 142 unique regions of 11 bacteria, five viruses, and two eukaryotes. After hybridization, the target is stained with a streptavidin-conjugated fluorophore, and the interaction of the target with specific probes is measured with epifluorescence confocal microscopy. This system is capable of detecting two copies of purified target DNA sequence. The total time required for the assay was 6 h, but it is expected to be significantly reduced with the recently developed microfibricated electromechanical sample preparation systems, rapid thermocycling used in real-time PCR, and fast, active electric field-enhanced hybridization.
Detection of B. anthracis antigens
Principles for the detection of B. anthracis by antibody–antigen reactions have been established for decades. Methods that have been developed vary based upon reporter systems, including enzyme-linked immunosor-bent assay, immunoflourescent microscopy, and flow cytometry. These methods can detect B. anthracis with high specificity and sensitivity, but involve many intermediate steps of incubation and washing, and take several hours plus technical expertise to perform. The ideal assay would produce results within minutes. Recent developments in immunological methods for rapid testing focus on the improvement of reporter systems that would provide test results in real-time.
Hand-held immunochromatographic devices, that are portable and require little training to use, work with liquid suspension of suspicious material. The sample liquid, when passing through an absorbent paper, would solubilize a lyophilized antibody. The antibody is specific to B. anthracis and has been labeled with colloidal gold particles. Anthrax antigen, if present in the sample, will form an antibody–antigen complex with the labeled antibody. This complex is carried by capillary forces through a nitrocellulose membrane, thereby passing a line of immobilized antibodies specific for anthrax. If an immune complex of anthrax antigen-gold labeled antibody are formed, the complex will be captured by the antibodies to anthrax on the test line, creating a visible line on the membrane within 1–15 min. None of the currently available hand-held devices have been evaluated with specimens from sporadic cases of anthrax for their actual specificityand sensitivity. They have a high detection threshold, requiring more than 104 spores for a positive reading, which is above the number the US Army Medical Research Institute of Infectious Diseases (USAMRIID) cites as necessary to cause infection. So, the tests need to be followed by a more sensitive one. They are also not very specific since the antibodies can cross-react with environmental strains of B. cereus. Because of these inadequacies, a federal memorandum released on July 21, 2002 advised First Responders not to use hand-held anthrax tests for field evaluation of suspected biological threats.
In flow cytometry, individual bacterial cells confined within a rapidly flowing jet of water pass a measuring window, in which multiple parameters pertinent to each single cell can be simultaneously measured. Bacterial cells can be labelled with either nucleic acid probes or antibodies and thousands of cells can be measured per second. Because B. anthracis spores are not permeable to nucleic acid probes, the spores can only be labeled using antibodies to spore surface proteins. Early attempts of using flow cytometry to detect B. anthracis took several hours to complete. By reducing the reaction time of the antibody with spores and eliminating washing steps, Stopa developed an assay that took 5 min and detected as low as 1000 spores/ml. Because flow cytometry only recognizes markers associated with particular matter, it is less susceptible to hoax attacks that use free DNA or antigen molecules. The current limitation to this method is that non-B. anthracis-specific spore surface antigens have been identified.
The Naval Research Laboratory developed a portable fibro-optical biosensor to perform fluoroimmunoassays. Similar to the flow immunochromatography test, this sensor is a kind of sandwich immunoassay. The target antigen was PA. A capture antibody attached on the fiber surface of an immunoprobe captured a cyanine dye labeled antibody–antigen complex. The fluorophores were excited by a laser beam and then converted to electronic signals. Initial studies yielded a limit of detection of 10 ng/ml for this assay, which is relatively low compared with most other immunoassays. As the next generation of lightweight, multichannel fiber optic instrument becomes available, this methodology of antigen detection could be applicable for multiple antigen/agent detection.
An automated electrochemiluminescence (ECL) detection system has been developed with the capability of detecting several organisms, including B. anthracis spores. The ECL system as described in detail by Higgins et al. is another kind of sandwich assay. It is performed by adding magnetic beads conjugated to a capture antibody and a ruthenium-conjugated detector antibody to an unknown sample. The chemiluminescent signal is generated by a voltage-dependent, cyclic oxidation–reduction reaction of the ruthenium heavy metal chelate. Automated applications for this technology are being developed with the ultimate goal of developing portable, hand-held devices capable of rapidly detecting all biological threats. The major advantage of ECL over fluorescence assays is a relatively low intrinsic background signal from biological materials and media. The resultant high signal-to-noise ratio allows ECL to detect antigens at a level comparable to radioisotopic techniques. The minimum detection sensitivity for a variety of biotoxoids is in the femtogram range, equivalent to several thousand to several hundred thousand molecules. This detection threshold is, in general, several orders of magnitude lower than most fluorescence methods. It also takes less than 20 min to detect a soluble antigen. When applied to B. anthracis spores, the method can detect as few as 10–100 bacterial cells within approximately 1 h of reaction time.
Rapid detection of viable
B. anthracis cells None of the rapid tests discussed above are able to differentiate live from dead B. anthracis cells. Conceivably, anthrax-positive results may also be due to a hoax since they may not necessarily mean live infectious anthrax spores are present. Schuch and colleagues have designed a new detection system for B. anthracis based on PlyG, a type of lysin, isolated from bacterial phage y, a phage that specifically attacks B. anthracis. In this method, spores were first heat-activated for 5 min and then incubated with PlyG which lyses B. anthracis cells. ATP released from the lysis of cells is measured as light emitted in the presence of luciferin/luciferase. The lysin-based method can detect a sample of 2500 spores in 10 min. Because the signals released are generated during metabolism, dead bacterial cells or surreptitious molecules will not trigger this system. However, this system still cannot replace the gold standard method of culturing, because it can be triggered by any B. anthracis strain regardless of their plasmid profiles.
Pros and cons of the two classes of detection methods are summarized in Table 4. Methods detecting DNA have a wide selection of markers specific for B. anthracis but do not truly provide immediate results because DNA extraction and/or amplification are necessary. Methods that detect antigens can be developed into a simple to use device, a real-time sensor, or tests detecting whole bacteria; however, they suffer from a lack of specific markers for virulent B. anthracis. Rapid detection of viable B. anthracis is still in its early stage of development. The main issue yet to be solved is how to discriminate between attenuated and fully virulent strains.
Limitations and future trends
Rapid molecular tests, nucleic acid-based or antigen-based, are operated based on the principle that the presence of a portion (target molecule) of an organism can be used as a marker to indicate the presence of the whole viable organism. In naturally occurring microbial diseases, the sole source of the marker is the whole organism, and discrepancy rarelyoccurs. In a bioterrorism event, material used may not be the true virulent organism, but attenuated organisms, DNA fragments, antigens, or even non-biological additives. If the fully virulent organisms were used, as in the anthrax attack that occurred in 2001, molecular tests would work just as well as in the diagnosis of a naturally occurring infectious disease and would warn the presence of virulent anthrax bacterium. However, the principle may not apply in hoax attacks.
Rapid detection of anthrax requires a multi-dimensional approach but current methods are designed to detect only a single dimension such as nucleic acids, antigens, other molecules, or bacterial cells. These protocols are not comprehensive and therefore suboptimal for current needs in this time of heightened fears of bioterrorism. Therefore, the ultimate goal must be to design a highly sensitive and specific method that detects all three parameters in a given sample: viability, virulence plasmids, and specific marker antigens. Without testing all three parameters, the presence of dead, non-virulent, or merely the DNA or antigen component of B. anthracis in a suspicious sample, may lead to a false-positive diagnosis. Although, none of the current assays can accomplish this goal alone, a combination of them can significantly, though not completely, reduce false alarms. At present, culture-based identification of B. anthracis remains an irreplaceable gold standard.