|Home | About | Journals | Submit | Contact Us | Français|
Yersinia pestis is the etiological agent of the plague. Because of the disease's inherent communicability, rapid clinical course, and high mortality, it is critical that an outbreak, whether it is natural or deliberate, be detected and diagnosed quickly. The objective of this research was to generate a recombinant luxAB (“light”)-tagged reporter phage that can detect Y. pestis by rapidly and specifically conferring a bioluminescent signal response to these cells. The bacterial luxAB reporter genes were integrated into a noncoding region of the CDC plague-diagnostic phage A1122 by homologous recombination. The identity and fitness of the recombinant phage were assessed through PCR analysis and lysis assays and functionally verified by the ability to transduce a bioluminescent signal to recipient cells. The reporter phage conferred a bioluminescent phenotype to Y. pestis within 12 min of infection at 28°C. The signal response time and signal strength were dependent on the number of cells present. A positive signal was obtained from 102 cells within 60 min. A signal response was not detectable with Escherichia coli, although a weak signal (100-fold lower than that with Y. pestis) was obtained with 1 (of 10) Yersinia enterocolitica strains and 2 (of 10) Yersinia pseudotuberculosis strains at the restrictive temperature. Importantly, serum did not prevent the ability of the reporter phage to infect Y. pestis, nor did it significantly quench the resulting bioluminescent signal. Collectively, the results indicate that the reporter phage displays promise for the rapid and specific diagnostic detection of cultivated Y. pestis isolates or infected clinical specimens.
Yersinia pestis is a gram-negative member of the family Enterobacteriaceae, and the etiological agent of the plague, which is a zoonotic disease affecting rats, prairie dogs, and other rodents. Y. pestis is transmitted to humans typically through bites by infected fleas. Bubonic plague, which develops 2 to 8 days later, is characterized by fever, chills, weakness, and the development of swollen lymph nodes, or buboes. The disease can also develop into septicemia without a bubo or occasionally into pneumonic plague. Although the disease is relatively rare in the United States, there are 1,000 to 5,000 cases and 100 to 200 deaths each year worldwide (35). Y. pestis is considered to be a high-priority agent for use as a bioterror weapon because of its ability to be transmitted from person to person, rapid clinical course, and high mortality rate (17). The World Health Organization (WHO) estimates that an aerosolized release of 50 kg of weaponized material over a populated city could cause 150,000 cases of pneumonic plague and 36,000 fatalities (34). Pneumonic plague is nearly always fatal if not treated within the first 12 to 24 h after symptom onset (36). Therefore, early detection and diagnosis would be vital in order to quickly implement public health measures.
Recombinant reporter phages may provide a “natural” and specific approach for the detection of Y. pestis. Reporter phage-mediated detection systems have been developed for Listeria monocytogenes, salmonellae, mycobacteria, Bacillus anthracis, and Escherichia coli O157:H7 (4, 13, 19, 22, 23, 30, 31). Due to the availability of a species-specific phage that can infect a wide range of Y. pestis isolates, a similar approach may be feasible for the detection of Y. pestis. For example, the CDC plague-diagnostic phage A1122 is currently used as a tool (in lysis assays) for the identification of Y. pestis. A1122 belongs to serovar 1 and is a member of the Podoviridae family. Sequence analysis has shown that A1122 is very closely related to the E. coli phages T3 and T7 (15). A1122 is ideally suited as a “diagnostic phage” since it has an unusual ability to infect most Y. pestis isolates. According to the CDC, A1122 can grow and lyse on all but two of thousands of natural isolates of Y. pestis within the CDC collection (M. C. Chu, CDC, unpublished observations noted in reference 15). Advier (2) also demonstrated that A1122 lysed all 47 Y. pestis strains tested (2). Furthermore, A1122 is “specific” to Y. pestis, with the exception of some strains from the closely related species Yersinia pseudotuberculosis (16); however, temperature may be used to differentiate the two species, since the phage does not grow on Y. pseudotuberculosis at temperatures below 26°C. Consequently, due to its specific and broad strain infectivity, A1122 is used by the CDC, the WHO, and the U.S. Army Medical Research Institute of Infectious Diseases as a diagnostic standard (in lysis assay) for the confirmed identification of Y. pestis (9, 10).
Although phage lysis assays are robust, they suffer from two shortcomings: (i) they are laboratory based, and (ii) they require 24 to 36 h to complete. Because the disease progresses very rapidly and because a positive prognosis requires the early administration of antibiotics, the ability to reduce the time required for confirmed identification of Y. pestis would have significant value. Therefore, the goals of our research are to provide the foundation for the generation of a recombinant “light-tagged” reporter phage that can quickly (within minutes) confer a bioluminescent signal to Y. pestis and to demonstrate the feasibility of the detection system using cultivated Y. pestis and Y. pestis in the presence of a clinical matrix.
The attenuated Y. pestis strain A1122 (NR-15), Yersinia enterocolitica strains (NR-204, NR-205, NR-206, NR-207, NR-209, NR-210, NR-211, NR-212, NR-213, and NR-214), and Y. pseudotuberculosis strains (NR-4371, NR-4372, NR-4373, NR-4374, and NR-4375) were obtained from the Biodefense and Emerging Infections Research Resources Repository (BEI Resources). Y. pseudotuberculosis NCTC 1105, NCTC 9507, NCTC 21, NCTC 10275, and NCTC 9509 were obtained from the American Type Culture Collection (ATCC 6903, ATCC 13980, ATCC 23207, ATCC 28933, and ATCC 27802, respectively). Escherichia coli ER2738 was obtained from New England Biolabs, while E. coli B was from the laboratory stock of I. J. Molineux. The CDC diagnostic phage A1122 was obtained from the Division of Vector-Borne Infectious Diseases at the CDC.
Bacteria were grown in 2 ml of Luria-Bertani (LB) medium at 26°C, 28°C, or 37°C as indicated for 18 to 24 h with shaking (225 rpm) to generate saturated cultures. Cultures were then diluted 1:15 (Y. pestis) or 1:50 (E. coli, Y. enterocolitica, and Y. pseudotuberculosis) into fresh media and incubated at the same initial temperature until the desired A600 was reached. Clonings were performed in E. coli ER2738 as described by Sambrook et al. (29). Phage A1122 was propagated using standard techniques (6).
The Vibrio harveyi luxA and luxB genes were used as the reporter genes and were targeted for integration into the A1122 phage genome downstream of the A1, A2, and A3 promoters and upstream of gene 0.3. pBluescript II-SK− containing the luxAB genes (pluxAB-SK) was used as the parental vector for the subsequent cloning steps (31). A 178-bp fragment of A1122 DNA (encompassing positions 720 to 897 of the sequence under GenBank accession number AY247822) was PCR amplified using phage DNA as template. The 5′ 0.3 primers (Table (Table1)1) contained SacI and XbaI restriction sites, respectively, for cloning into the corresponding sites of pluxAB-SK to create the 5′-flanking phage DNA for homologous recombination (p5′0.3-luxAB-SK). A 176-bp fragment of phage DNA (encompassing positions 904 to 1079) was PCR amplified with primers containing XhoI and SphI (Table (Table1,1, 3′ 0.3 primers). The PCR product was cloned 3′ of luxAB into the corresponding sites of p5′0.3-luxAB-SK, thereby flanking the luxAB cassette with phage DNA (p5′0.3-luxAB-3′0.3-SK). The sequences of the cloned fragments were verified by deoxy dye terminator sequencing using the Applied BioSystems (ABI, Foster City, CA) standard instructions for the BigDye Terminator v1.1 cycle sequencing kit and the ABI Prism 377 sequencer.
The luxAB genes were integrated into the phage genome through homologous recombination (double-crossover event) between A1122 phage and integration cassette DNA. Y. pestis A1122 was transformed using the calcium chloride method (29) with p5′0.3-luxAB-3′0.3-SK, and transformants were selected after growth for 48 h at 28°C on LB agar supplemented with 100 μg/ml ampicillin. Transformants were positive for bioluminescence as expected (data not shown). Cultures were grown in LB agar supplemented with 100 μg/ml ampicillin at 28°C to an A600 of 0.3, infected with A1122 (1 ×105 PFU/ml [final concentration]), and incubated until lysis was observed. The culture supernatants (containing wild-type and a small number of recombinant A1122::luxAB phage) were clarified by centrifugation (8,000 × g, 5 min) and passed through sterile nylon 0.2-μm syringe filters (Pall Corporation, Ann Arbor, MI).
The titers of the mixed phage lysates were determined on Y. pestis using the agar overlay technique (6) on multiple plates (>30) to give near-confluent lysis. The resulting phage (approximately 5,000 plaques/plate) were eluted in SM buffer (50 mM Tris-HCl [pH 7.5], 0.1 M NaCl, 8 mM MgSO4·7H2O, 0.01% gelatin), and an aliquot from each plate was screened by PCR for the presence of luxA. The titers of phage lysates from plates that were positive for luxA were redetermined and these lysates were rescreened using successively higher dilutions, with each dilution predicted to contain a higher proportion of recombinant phage. This process of determining the titers of successively higher dilutions and identifying luxA-positive plate lysates was repeated eight times until individual plaques could be readily picked and screened for luxA; at this stage, 4 of 20 individual plaques were positive for luxA. Clonal A1122::luxAB phage were prepared from single plaques and amplified on exponentially growing Y. pestis at 28°C, and the phage supernatants were clarified by centrifugation and filtration as described in the previous paragraph. Clonal stocks of A1122::luxAB contained 1010 to 1011 PFU/ml and were stored at 4°C in the dark until needed.
To identify the presence of the A1122::luxAB recombinant phage and to confirm that luxAB integration had occurred at the correct site in the phage genome, cell-free phage supernatants were analyzed by PCR. Primers were designed to detect the presence of luxA (reporter), or to span the 5′ and 3′ integration junction sites (Table (Table1,1, 5′-INT and 3′-INT). Each integration primer set was designed to ensure that primer binding occurred both inside and outside the original integration cassette. PCR analysis was performed using standard techniques as recommended by the manufacturer of the Taq DNA polymerase (New England Biolabs, Ipswich, MA).
Unless otherwise stated, A1122::luxAB phage (~108 PFU/ml final) and Y. pestis cells were mixed and incubated at the designated temperatures. “Flash” bioluminescence was measured using a Biotek Synergy II multiplate detection reader. Cultures (200 μl per reading) were injected with n-decanal (0.5%) and read for 10 s. Controls consisted of cells alone or phage alone. Bioluminescence is depicted as relative light units (RLU), and the results presented are the averages from three experiments ± standard deviations (SD). Statistical significance was determined using Student's t test (P < 0.05).
PCR analysis using the luxA, 5′-INT and 3′-INT primers generated PCR products of the correct predicted size (Fig. (Fig.1A).1A). This indicated the presence of the reporter and that the luxAB cassette had integrated at the correct genome site as expected.
To investigate whether the addition of the heterologous reporter compromised the “fitness” of the recombinant phage, the ability of A1122::luxAB to lyse Y. pestis was analyzed and compared to that of the wild-type phage A1122. Lysis was assessed by monitoring bacterial growth (culture optical density) in the absence or presence of the wild-type or recombinant phage. Both the wild-type A1122 and recombinant A1122::luxAB were able to cause a significant drop in the optical density of the cultures due to phage-mediated cell lysis (Fig. (Fig.1B).1B). The lysis time of the recombinant was delayed compared to that of the wild type. Nevertheless, recombinant stock titers were in the range of 1010 to 1011 PFU/ml and were comparable to the parental phage.
The ability of A1122::luxAB to transduce a bioluminescent phenotype to Y. pestis A1122 was assessed. Y. pestis A1122 is an exempt select-agent strain (26), which lacks critical virulence factors. The time required to generate a bioluminescent signal response was assessed at 28°C and 37°C (Fig. (Fig.2A).2A). A growth temperature of 28°C is optimal for Y. pestis (8). However, incubation at 37°C could potentially result in a shorter signal response time, since the closely related T7 phage infects and replicates faster at elevated temperatures (5, 37). Therefore, Y. pestis was grown at 28°C or 37°C until the early exponential stage of growth and then mixed with phage, and the ability of the reporter phage to transduce bioluminescence (RLU) was monitored over time at the designated temperatures. A steady increase in bioluminescence was detected from Y. pestis phage-infected cells at 28°C and 37°C (Fig. (Fig.2A).2A). Moreover, a detectable light signal above background (phage alone or cells alone) was evident within 12 min of phage addition. Although the magnitude of the signal was initially higher (approximately eightfold) at the 12-min time point at 37°C as expected, the signal response time and the strength of signal were generally similar at 28°C and 37°C. Collectively, the results indicated that (i) the A1122::luxAB phage were able to infect and rapidly transduce a bioluminescent phenotype to Y. pestis, (ii) the luxAB genes were functional in Y. pestis, and (iii) phage-mediated detection of Y. pestis was comparable at both 28°C and 37°C. It should be noted that incubation of the reporter phage with the cells for time periods longer than shown (>90 min) resulted in a gradual decline in signal strength, irrespective of the temperature, presumably reflecting phage-mediated cell lysis (data not shown).
To investigate assay sensitivity and dose-dependent characteristics, 10-fold serially diluted cells (4.1 × 106 to 4.1 × 103 CFU/ml) were mixed with A1122::luxAB phage and analyzed for bioluminescence over time (Fig. (Fig.2B).2B). The highest CFU/ml produced the strongest signal, at over 10,000 RLU within 40 min. As the cell number decreased from 106 to 103 CFU/ml, the signal response (RLU) decreased and the signal response time increased, indicating dose-response characteristics (Fig. (Fig.2B).2B). Nevertheless, 4.1 × 103 CFU/ml were detectable (P < 0.05) at 60 min after phage addition; since 200 μl of the sample was measured per reading, this equates to the detection of 820 cells.
To investigate the relationship between reporter phage-mediated detection and cell fitness, Y. pestis was collected at various stages during the growth cycle (Fig. (Fig.3A)3A) and assessed for the ability to emit a bioluminescent signal. The signal response times for cells harvested at different phases of growth were similar, as indicated by a positive signal at the earliest time point analyzed (Fig. (Fig.3A)3A) (within 15 min). As expected, the magnitude of the signal response was reduced for cells collected during the early stages of stationary growth (set 4) compared to the other samples. Nevertheless, the results suggest that cell “fitness” does not eliminate the ability of the reporter phage to transduce a bioluminescent phenotype to Y. pestis.
The ability (or inability) of the reporter phage to transduce a signal response to closely related Yersinia species other than Y. pestis was assessed. Exponentially growing cells of 10 Y. enterocolitica strains, 5 Y. pseudotuberculosis strains, and 2 E. coli strains (B and K-12) were mixed with reporter phage and analyzed for a bioluminescent signal (Table (Table2).2). As expected, a signal response was not obtained with the two E. coli strains. At the permissive temperature of 37°C, 5 (out of 10) of the Y. pseudotuberculosis strains elicited an attenuated response. At the discriminatory temperature (26°C), only two of the Y. pseudotuberculosis strains elicited a signal, indicating a temperature-differential response, as expected. However, the two “positive” strains displayed a signal response that was approximately 100-fold lower than that for Y. pestis. Unexpectedly, 1 (out of 10) Y. enterocolitica strains produced a bioluminescent signal (strain CDC 497-70, with 106 RLU). The onset of the signal response was delayed in comparison to that for Y. pestis, and the magnitude of the response did not increase significantly at the time points analyzed (up to 110 min) (data not shown). Moreover, the signal strength was approximately 100-fold less than that observed for Y. pestis under comparable conditions (Table (Table22).
To assess whether the reporter phage can detect Y. pestis directly in a clinical matrix, the ability of the reporter phage to detect Y. pestis in deliberately spiked human serum was assessed. Serum was chosen as a suitable matrix since bacteremic infection is often the result of both pneumonic and bubonic plague.
The feasibility of whether detection was possible directly in serum was investigated initially by assessing whether serum quenches the bioluminescent signal. Human serum was spiked with 10-fold serial dilutions of Y. pestis harboring plasmid-borne copies of the luxAB genes, and the ability to detect the bioluminescent signal was measured and compared to that of analogous dilutions in LB medium (Fig. (Fig.4A).4A). The sensitivity limits of detection (100 CFU) and the pattern of response were the same. To investigate the ability of the reporter phage to bind, infect, and transduce a bioluminescent signal to Y. pestis in serum, reporter phage were mixed with Y. pestis in the presence of 0, 24, 49, 74, and 99% serum (Fig. (Fig.4B).4B). The reporter phage was able to confer a bioluminescent signal to Y. pestis in samples containing serum; however, serum did increase the signal response time and decrease the signal strength, most notably within the 99% serum samples. Although the time to initial detection and signal peak were delayed compared to those for samples containing LB only, the signal responses of samples containing 24 to 74% serum and those in LB were similar.
The use of phages for the detection of bacterial species is well documented. A1122 is used in a phage lysis assay by the CDC for the confirmed identification of Y. pestis isolates (9, 10). A similar lysis assay using γ phage is FDA approved as a standard for the identification of B. anthracis isolates (1). In an effort to significantly reduce the time to detection and to enable detection in complex matrices such as food or clinical specimens, reporter phages are being developed as biodetectors of pathogenic bacteria (27). For example, reporter phages have been generated for E. coli O157:H7 and L. monocytogenes and have proven effective for the detection of bacteria in a range of diverse food matrices such as milk, ground beef, spinach, and soft cheese (4, 22, 23, 25, 28). Similarly designed reporter mycobacteriophages have proven effective for the clinical identification of Mycobacterium tuberculosis from sputum samples (3). Therefore, the unique attributes of phages are being exploited in the detection field.
The bacterial luxAB genes were used as the reporter of choice because (i) the luxAB genes have been successfully used as a reporter for the phage-mediated detection of gram-positive and -negative bacterial pathogens (4, 19, 22, 23, 31); (ii) the bioluminescent signal may be visualized by a simple hand-held photon detection device (e.g., a camera); and (iii) no processing of the sample is required, as the only requirement is the addition of the substrate n-decanal. However, reporter phages typically require active cell metabolism in order to express luxAB and generate a bioluminescent response. Although this ensures that only viable, potentially infectious cells can be detected (31), it is likely that phage infection and luxAB expression are directly correlated with host fitness. Thus, cells that are actively dividing will presumably be more susceptible to reporter phage infection and have a greater capacity to generate the ensuing bioluminescent signal response.
The A1122 genome has recently been sequenced, making genetic manipulation more easily attainable. The genome consists of 37,555 bp, encodes 51 predicted gene products, and has a nucleotide identity of 89% to the E. coli phage T7 (15). The luxAB genes were targeted for integration into a noncoding region of the A1122 genome downstream of the A1, A2, and A3 promoters and upstream of gene 0.3. In doing so, 6 bp of phage DNA was replaced with 2,117 bp of reporter DNA and thereby luxAB expression was placed under A1 to A3 transcriptional control. This strategy was chosen for several reasons: (i) insertion of reporter DNA (2 kb) into the phage genome was not expected to compromise phage packaging, since the homologous T7 capsid is able to accommodate a similarly sized genome; (ii) the use of endogenous phage promoters (as opposed to the introduction of a heterologous promoter) was less likely to disrupt/interfere with phage function; and (iii) in E. coli, the homologous T7 A1, A2, and A3 promoters are among the strongest unregulated promoters known and are strongly expressed during the early stages of phage infection (11, 12). Consequently, the placement of luxAB downstream of A1 to A3 was expected to result in a strong and rapid bioluminescent signal in Y. pestis. Moreover, the fact that recombinant stock titers were comparable to those of the parental phage suggests that the burst size of the recombinant phage was not adversely compromised by the introduction of the heterologous reporter genes.
The results presented demonstrate the ability of the reporter phage to rapidly detect Y. pestis, since the reporter phage transduced a bioluminescent signal response to Y. pestis cultures within 10 to 15 min after phage addition. Compared to the standard phage lysis assays, which generally require 24 h for completion, this significantly decreases the time to detection. Rapid detection and diagnosis are essential for a positive prognosis, since the plague, especially pneumonic plague, is nearly always fatal if treatment is not administered within the first 12 to 24 h after symptom onset (36). The reporter phage also demonstrated the ability to directly detect Y. pestis in human serum, without the prerequisite of isolation and subsequent cultivation. With the availability of portable photon detection devices, the potential exists for the generation of a diagnostic kit for direct use with clinical samples in the field. A caveat of the phage detection system, however, is the potential of the phage to infect strains of the closely related species Y. pseudotuberculosis and hence cause false-positive results. Gunnison et al. (16) found that A1122 lysed 21 out of 45 Y. pseudotuberculosis strains analyzed (approximately 50% of the strains). In contrast to the case for Y. pestis, lysis occurred only at 37°C and not at 20°C, and thus the test is specific provided that it is performed at reduced temperatures. It is interesting to note that the Y. pseudotuberculosis strains that were deemed phage susceptible by Gunnison et al. (16) differed markedly in their susceptibility to the phage, with many of the strains showing lysis only when “spot tested” using undiluted or 10−1 serially diluted phage, suggesting that a very high multiplicity of infection was needed; this susceptibility is in stark contrast to that of Y. pestis, since in that case all 47 strains were susceptible to highly diluted (10−6) phage (20). In our research, five Y. pseudotuberculosis strains tested positive for bioluminescence at 37°C. This signal was either absent or significantly reduced at 26°C, indicating a temperature-differential response as expected. Moreover, the “reporter phage-positive” strains produced an attenuated bioluminescent signal response at the restrictive temperature that was approximately 100-fold lower than that obtained with Y. pestis under similar conditions. The reason for this attenuated signal is not known, but it is likely due to a reduced ability of the reporter phage (and parental phage) to infect and grow on Y. pseudotuberculosis rather than to a reduced ability to express and accumulate LuxAB.
Due to the ability of A1122 to grow on nearly all Y. pestis isolates (2, 15, 16) and the inability of A1122 to grow on Yersinia species other than Y. pestis under selective growth conditions, the A1122::luxAB reporter phage detection system has the potential to be a highly specific test for Y. pestis. The mechanisms for determining parental A1122 specificity and A1122::luxAB reporter phage specificity, however, are not analogous. This is because A1122 specificity is determined by the ability of the phage to “attack” the host, which encompasses phage infection, phage replication, and cell lysis, with the presence of plaques or “host clearing” being indicative of host susceptibility. In contrast, A1122::luxAB host susceptibility is measured by a bioluminescent signal response which requires fewer steps, i.e., phage infection and luxAB expression (not multiplication and lysis). Thus, the requirements and events that necessitate a bioluminescent signal are not the same as those that mediate cell lysis. In addition, there may be host inhibitory factors that can prevent phage replication and lysis but may not inhibit luxAB expression, such as the presence of a prophage in the cell, DNA restriction-modification, lysis resistant mutants, and even phage-specific inhibition genes (abortive infection) (7). Thus, it is possible that the reporter phage can have a detection range that is wider that the parental phage cell lysis host range (19, 27). Our results indicate that the reporter phage did not transduce a bioluminescent signal to E. coli K-12 or E. coli B, but 1 of the 10 Y. enterocolitica strains (CDC 497-70) tested produced an attenuated bioluminescent signal (Table (Table2).2). The bioluminescent signal response obtained with the Y. enterocolitica strain was approximately 100-fold lower than that for Y. pestis. This strength of signal was similar to the signal obtained with the “reporter phage-positive” Y. pseudotuberculosis strains and again suggests a reduced capacity of the phage to grow on Yersinia species other than Y. pestis.
The reporter phage was able to detect Y. pestis at a concentration of 4 × 103 CFU/ml within 60 min. Since only 200 μl was measured per assay, the amount of cells detected corresponded to a sensitivity of 820 CFU. These results were obtained (i) without collecting cells and artificially putting them in close proximity to the phage, (ii) without a preenrichment step for amplifying the target cells, and (iii) with a multidetection reader that reads absorbance and fluorescence as a primary function and luminescence as a secondary function. The last aspect is very important since the technology exists for the detection of bioluminescence from a single cell, which would significantly improve the detection system. In addition, it may be possible to increase luxAB expression by changing the expression signals and/or changing the location of the reporter in the phage genome. Commercially available diagnostics such as the Plague BioThreat Alert test strips are based on the detection of the fraction 1 capsular antigen. The sensitivity limits of detection are between 6 × 103 and 7 × 103 CFU/ml, which is similar to the detection sensitivity described in this report (33). However, both of these sensitivity limits pale in comparison with that of real-time PCR, which can detect between 0.01 and 1 pg of extracted DNA; assuming a 100% extraction efficiency, this translates to approximately 1 CFU (24, 32). Nevertheless, the reporter phage technology has a number of desirable traits compared to PCR-based assays. These include simplicity/minimal processing (the phage is mixed with the samples and bioluminescence is measured), portability, and low cost, which would permit routine surveillance. In addition, since active cell metabolism is required for reporter phage detection, a potential advantage of the reporter phage in comparison to PCR is that the latter methodology detects the presence of DNA but offers no information on whether the cell is intact and potentially infectious. This aspect is particularly relevant for environmental samples.
The A1122 phage was chosen for this study because it is currently used by the CDC in a phage lysis assay for the confirmed identification of Y. pestis. Y. pestis is a highly pathogenic select agent. Therefore, to overcome the safety and regulatory issues governing the use of this virulent species, we used the attenuated Y. pestis A1122 strain in order demonstrate the feasibility of the reporter phage as a diagnostic tool. Y. pestis A1122 is an exempt select-agent strain (26) which lacks the pgm locus and the pCD1 plasmid, both of which encode critical virulence factors. Attenuated strains are commonly used for research purposes in order to demonstrate diagnostic detection feasibility (21, 31, 33). Although we used an attenuated strain in this study, there is strong evidence to indicate that the results obtained will translate to fully virulent Y. pestis strains. This is because (i) the A1122 phage is able to infect both attenuated and fully virulent Y. pestis strains (20), and moreover, according to the CDC, the phage can lyse all but two of thousands of natural Y. pestis isolates in the CDC collection (M. C. Chu, CDC, unpublished observations noted in reference 15); (ii) A1122 phage infection and lysis are not dependent on the virulence status of the Y. pestis strain; and (iii) the luxAB reporter genes are functional in virulent Yersinia species (14, 18).
In summary, we generated a recombinant luxAB-tagged phage and demonstrated that the reporter phage can rapidly transduce a bioluminescent signal response to Y. pestis. The phage displays promise as a tool for the detection of cultivated isolates or within clinically relevant matrices.
This research was supported by National Institutes of Health (National Institute of Allergy and Infectious Diseases) grant 1R43AI082698-01 awarded to D.A.S. of Guild Associates, Inc.
We thank Alexander Sulakvelidze (University of Florida) and Michael Schmidt (Medical University of South Carolina) for their advice and support. We also thank the Medical University of South Carolina Biotechnology Resource laboratory for sequencing data.
Published ahead of print on 14 October 2009.