Ambient Stable Quantitative PCR Reagents for the Detection of Yersinia pestis

Background Although assays for detecting Yersinia pestis using TaqMan probe-based real-time PCR have been developed for years, little is reported on room-temperature-stable PCR reagents, which will be invaluable for field epidemic surveillance, immediate response to public health emergencies, counter-bioterrorism investigation, etc. In this work, a set of real-time PCR reagents for rapid detection of Y. pestis was developed with extraordinary stability at 37°C. Methods/Principal Findings TaqMan-based real-time PCR assays were developed using the primers and probes targeting the 3a sequence in the chromosome and the F1 antigen gene caf1 in the plasmid pMT1of Y. pestis, respectively. Then, carbohydrate mixtures were added to the PCR reagents, which were later vacuum-dried for stability evaluation. The vacuum-dried reagents were stable at 37°C for at least 49 days for a lower concentration of template DNA (10 copies/µl), and up to 79 days for higher concentrations (≥102 copies/µl). The reagents were used subsequently to detect soil samples spiked with Y. pestis vaccine strain EV76, and 5×104 CFU per gram of soil could be detected by both 3a- and caf1-based PCR reagents. In addition, a simple and efficient method for soil sample processing is presented here. Conclusions/Significance The vacuum-dried reagents for real-time PCR maintain accuracy and reproducibility for at least 49 days at 37°C, indicating that they can be easily transported at room temperature for field application if the machine for performing real-time PCR is available. This dry reagent is of great significance for routine plague surveillance.


Introduction
Yersinia pestis, the causative pathogen of the plague, mainly resides in rodents and can be transmitted to humans by infected fleas [1]. As Y. pestis is highly virulent and infectious, it has always been recognized as one of the classical biological warfare agents [2] and was classified as a Category A pathogen by the U. S. Center for Disease Control and Prevention (http://www.bt.cdc. gov/agent/agentlist-category.asp) [3].
Y. pestis was traditionally identified by bacterial isolation and microscopy observation [4], the phage lysis assay [5] and animal experiments, which was termed as a ''four-step'' protocol in China. Although it is time-consuming and laborious, this protocol is still a gold standard for laboratory confirmation of Y. pestis infections. Immunological methods were also developed for the detection of F1 antigen and antibodies against Y. pestis [6,7,8]. Immunological biosensors, based on fiber optics, magnetic and upconverting phosphor technology, were recently applied in detection of antigen and antibodies of Y. pestis as well [9,10,11]. These methods have played important roles in fighting plague, however, nucleic acid-based detection techniques could be an even powerful alternative for detecting Y. pestis.
Conventional PCR-gel electrophoresis method has been developed for detecting Y. pestis in fleas and other specimens [12,13,14]. A handful of real-time quantitative PCR assays in various formats were also established for detecting and identifying Y. pestis [15,16,17,18,19,20,21,22,23]. Real-time PCR assays provide greater specificity, and they require less time and labor to complete than conventional PCRs. The techniques applied include SYBR Green [24], molecular beacon [25], TaqMan probes [18,20] and minor groove binding (MGB) probes [22] ect., targeting specific sequences on the chromosome and (or) plasmids. Although real-time PCR has been successfully used in detecting and identifying Y. pestis, the relevant reagents need to be transported under low temperature in dry ice in order to keep the activities of enzymes and labeled probes. In this report we developed a room-temperature stable reagent for real-time PCRs which targeted the 3a sequence [26] in chromosome and the caf1 gene [14] in the plasmid pMT1. This reagent could be stable during transportation at room temperature and thus be reliably applied for on site detection of target microorganisms if the thermal cycler is available.

Genomic DNAs
Genomic DNAs of four biovars (Microtus, Orientalis, Antiqua and Mediaevalis) of Y. pestis were stored in our lab. Closely related or other genomic DNAs used in this study include 9 species of Yersinia (Y. enterocolitica, Y. intermedia, Y. aldovae, Y. bercovieri, Y. frederiksenii, Y. kristensenii, Y. mollaretii, Y. rohdei, Y. ruckeri), 16 different serotypes of Y. pseudotuberculosis, Brucella, Francisella tularensis, Bacillus anthracis, and E. coli DH5a; and DNAs from human blood and mouse. All bacterial strains used in this study were listed in Table 1.

Primers and probes
The primers and probes for real-time PCR were designed based on the 3a sequence in the chromosome [26] and caf1 in plasmid pMT1 using Primer Express 2.0 (PE Corporation,USA). Other primers were also designed for cloning the target 3a and calf1 sequences into pGEM-T Easy Vector (Promega, USA). The primers and probes used in this study were listed in Table 2.

Standard curve
DNA fragments flanking amplicons of 3a and caf1 were amplified using primers CF and CR (Table 2), respectively, from the 91001 genomic DNA for cloning into the pGEM-T Easy Vector according to the standard protocols reported elsewhere [18,27]. The ligation products were transformed into DH5a and positive clones were identified by PCR and sequencing. Plasmids containing target fragments were purified separately and linearized by Sal I digestion.
The concentration of linearized plasmid solution was then determined by UV spectrophotometer for calculating the copy numbers of the target DNAs. The quantified plasmid solution was serially diluted by 10-fold to prepare the standard templates with known copy numbers of target DNAs. Therefore, the real-time PCR was performed using these templates for obtaining the standard curves of C t -Log concentration for 3a and caf1, respectively [18,27]. C t , the cycle threshold, refers to the cycle at which the fluorescence from a sample crosses the threshold.
Sensitivity and specificity Y. pestis live attenuated strain EV76 was cultivated in Luria-Bertani broth overnight at 26uC and then serially diluted by 10fold. The number of viable cells was determined by counting colony forming unit (CFU) on the HBI (Heart Brain Infusion) agar plate. Five ml of each sample were directly applied to PCR amplification for evaluating the sensitivity of the PCR systems. All experiments were performed in duplicate. Genomic DNAs listed in Table 1, human and mouse DNAs were employed to evaluate the specificity of the PCR systems.

Stability of real-time PCR reagent
All reagents needed for real-time PCR were mixed up at appropriate proportion and dispensed into 200 ml microcentrifuge

Author Summary
Plague, caused by Yersinia pestis, is one of the oldest and most dangerous diseases in human history, and has claimed millions of lives in the three major historical pandemics. Although panic caused by the Black Death is fading, the threat of the reemergence of plague pandemics still exists, with the additional potential of misuse in biowarfare or bioterrorism. Rapid on-site detection and identification of the pathogen is of paramount significance for timely implementation of effective countermeasures. TaqMan probe-based real-time PCR assays can give quick and accurate identification; however, the need for cold delivery and storage prevents its potential on-site application. The objective of this study was to develop a stable PCR system for easy delivery and storage under room temperature, which is vital for conventional plague surveillance and for preparedness in public health emergencies. We present a solution to this particular issue, hoping that it is helpful to future applications. tubes with 15 ml each for a single reaction. The dispensed reaction mixture was then vacuum-dried. For stability evaluation, the dry reagents were kept away from lights at 37uC up to 79 days and the reagents were taken out at different time points to perform PCR by the following method. The reagents were reconstituted by adding 10 ml of ddH 2 O, and then, 5 ml of plasmid templates containing 10 1 , 10 3 , 10 5 and 10 7 copies/ml were separately added. The realtime PCRs were performed according to the protocol mentioned above. Only the real-time PCR system for 3a primers and probe was evaluated in this experiment. The stability of the real-time PCR reagent is assessed in terms of accuracy and reproducibility. A standard curve of the PCR system was made using the plasmid templates as mentioned above, and the ''log 10 concentration'' of the freshly prepared reagents was chosen as reference values. Results of ''log 10 concentration'' of the reagents kept at 37uC for different periods were compared with the reference ones for accuracy evaluation.
Four plasmid templates (10 1 , 10 3 , 10 5 , and 10 7 copies/ml) were used to perform PCR in triplicate in parallel. The coefficient of variation (CV) of C t values was used to evaluate the reproducibility of the 3a PCR systems.

Detection of Y. pestis in spiked soil samples
A culture of Y. pestis EV76 was serially diluted by 10-fold, and the number of viable cells was determined. For each of serial dilutions, 100 ml was added to 0.2 g soil samples, and the spiked samples were inoculated at 4uC over night for complete adsorption of the bacteria to soil. The soil samples were collected from the yard of our institute, and they were air dried at room temperature. The dried soil samples were directly spiked with dilution of Y. pestis cultures without any pretreatment for mimicking the real situation that is usually confronted in practice.
The spiked soil samples were vortexed thoroughly and centrifuged at 2, 000 rpm for 4 min and the supernatants were collected. Then, 1 ml of dH 2 O was added to the pellet, and the mixture was vortexed completely and centrifuged again at 2, 000 rpm for 4 min. The supernatant was transferred to the tube containing the supernatants collected in the first step. This washing step was repeated once for collecting as many bacteria in the soil as possible. The supernatants collected above were mixed together for centrifuging at 8,000 rpm for 5 min to collect the target bacteria. The pellets were washed by dH 2 O once, and 100 ml of dH 2 O was added to each sample to resuspend the pellet for boiling in a water bath for 5 min. Five microlitres of this solution were used as templates for real-time PCR. We evaluated the sensitivity and specificity of the PCR system for detecting Y. pestis in the spiked soil samples and a blind experiment was performed for evaluating the feasibility of applying this technique in detection of Y. pestis directly from soil samples.

Sensitivity and specificity for detecting Y. pestis in soil samples
The spiked soil samples with different concentrations of Y. pestis were prepared and processed following the procedures described above, and the washed supernatants were used for sensitivity evaluation. As for the specificity test, genomic DNA (100 mg each) of the four biovars of Y. pestis, and other bacterial genomic DNAs listed in Table 1 (100 mg each) were all spiked into soil samples separately, and the real-time PCR was performed to test the specificity of the primers and probes. All experiments were performed in triplicates. Blind detection of Y. pestis in the spiked soil samples Soil samples spiked with different concentrations of EV76 were prepared by our colleague 1, and disordered and renumbered by our colleague 2, and then, the real-time PCR were performed by the first two authors of this article. After reporting the results by the first two authors, the three sides involved in this blind experiment sat together for evaluating agreement of the results to the original preparations. The blind test was performed in triplicates.

Standard curves
As shown in Fig. 1A and Fig. 1B, standard curves of both 3a and caf1 sequences, based on 10-fold serial dilutions of 10-10 8 copies/ml, have a linear relationship between the log 10 concentration and the C t s, with slopes of 23.3687 and 23.5154, respectively. The efficiency of each reaction is 98% and 93%, and Y-intercepts of 36.298 and 40.317, respectively, indicating a good quantitative determination of the PCR system. The r 2 (coefficient of determination) are 0.9996 and 0. 9997, for 3a-and caf1-based PCRs, respectively, implying the accuracy of serial dilution and the precision of sampling, and the ideal linearity of our PCR system as well.

Sensitivity and specificity
The sensitivity of real-time PCR system based on the 3a and the caf1 sequences is 1.5 and 15 CFU per system, respectively, using dilutions of EV76 as template (Table 3). For both 3a and caf1 primers and probes, all four biovars of Y. pestis gave positive signals, and the other genomic DNAs in Table 1, human and mouse DNAs gave negative results, indicating the good specificity of these PCR systems.

Stability of real-time PCR reagent
The vacuum-dried reagents were put in a 37uC incubator and taken out at different time points. We used the ratio of ''log 10 concentration'' of them against the original ''log 10 concentration'' of the standards as a criterion to evaluate the stability of the reagents. A value close to 1 indicates good stability of the reagents, and we set the ratio of 0.900,1.100 as the standard of acceptable stability. As shown in Table 4, only the results at the days 18 and 79 for lower concentration (10 copies/ml) were beyond this threshold. However, for the days 23, 32 and 49 at this concentration, the results were good enough to indicate the stability of the reagents. Therefore, we could reasonably attribute the biased results at day 18 for the lower concentration to the experimental errors. For concentration of template higher than 10 copies/ml, the vacuum-dried real-time PCR reagents could remain effective for at least 79 days at 37uC.
For evaluating the reproducibility of the PCR reagents, four concentrations of standard templates (10, 10 3 , 10 5 , 10 7 copies/ml) were chosen for performing real-time PCR in triplicates. As illustrated in Table 5, the coefficients of variations (CV) of C t value for all concentrations were less than 3% in the 79-day's storage at 37uC, indicating an excellent reproducibility.
The results above suggest that the vacuum-dried real-time PCR reagent could be transported at room temperature by normal post or express mailing system.

Detection of Y. pestis in artificially contaminated soil samples
For the spiked soil with different concentrations of Y. pestis EV76 for evaluating sensitivity of our PCR system, the results demonstrated that 5610 4 CFU per gram of soil could be detected by both 3a-and caf1-based PCR systems when C t .35 was considered as a negative result. Genomic DNAs of four biovars of Y. pestis and its closely related and non-related bacterial genomic DNAs were all spiked separately into soil samples for evaluating the specificity of the PCR systems, and only the samples spiked with the four biovars of Y. pestis DNA showed positive results while all the samples spiked with other DNAs gave negative results. In the blind test, all the spiked samples with different amount of Y. pestis EV76 were detected correctly. As expected, the CFU values determined by the real-time PCR were 10 to 100 times lower than the ones spiked into the soil samples. These results implied that the real-time PCR systems could detect Y. pestis sensitively and specifically from soil samples.

Discussion
In vitro amplification of nucleic acids by PCR has been widely used for both research and clinical diagnosis. Although quantitative real-time PCR of the TaqMan technology has been developed and applied in detection of pathogenic microorganisms from different clinical samples for many years, the reagents or diagnostic kits should be transported under dry ice for protecting enzymes from activity loss. The reverse-transcriptase PCR kit for detecting dengue virus was once evaluated by storing at room temperature, refrigerator and freezer, and the results indicated that the test kits could only be stored above its recommended storage temperature of 220uC for no more than 3 days [28]. Ramanujam et al. reported that the PCR reaction mixtures could be stabilized in carbohydrate polymers by forming glassy matrices that provide room-temperature stability [29]. This kind of stabilized reagent could be stable at 22uC for 6 weeks. Wolff et al. modified a nested PCR system using single-tube, preformulated mixes embedded in a trehalose matrix, which allowed the reagents to be stored for .6 months at ambient temperature [30]. Carbohydrates (trehalose, inulin or dextran) were also employed to stabilize influenza subunit vaccine in the dry state, making the vaccine stable for at least 26 weeks at room temperature [31]. It has been recognized that some yeast cells tend to accumulate high concentrations of trehalose when submitted to heat shock, which can lead to a more heatstable conformation of the enzyme in yeast [32]. Trehalose has been used to stabilize methanol dehydrogenase for the detection of methanol [33]. In this study, we initially wanted to compare different kinds of carbohydrates in PCR mixture for their ability in stabilizing the PCR reagents. Fortunately, when we tested our first   ), however, the one with stabilizers, as shown in this report, could be stable for a much longer period at 37uC. This result indicated the stabilizing effect of the added carbohydrates on PCR reagents. As shown by our results, this set of reagents can be stable at 37uC for up to 79 days when the concentration of template is higher than 10 copies/ml. For lower concentration, it can be stable at least for 49 days at 37uC. These results reveal that the stabilized PCR reagent can be transported by normal post mailing or express mailing system without dry-ice protection. To test its feasibility of delivering the reagents to different places without cold preservation, we mailed this reagent to different labs in Xinjiang, Shanghai, Tibet, Sichuan, Liaoning, Guangdong, Hubei Provinces by domestic couriers. It usually took 2 to 4 days for the receivers to get the reagent. The receivers tested them and confirmed that the sensitivity of the reagent met the requirement (data not shown).
Plague is one of the most important natural focus-related zoonotic diseases. Routine surveillances are of paramount significance for disease case reporting and controlling. In some of the surveillance stations, dead animals without any information of its dead time and reason and their spoiled soil in the natural plague foci are common samples for detecting and isolating Y. pestis. It usually takes 10 to 20 days to get final results. If PCR could be used in routine surveillances, it would improve the surveillance efficiency [12,34]. However, the surveillance stations are usually set up in the field, some of which even do not have freezers or refrigerators. A room-temperature-stable PCR reagent is therefore urgently demanded. After developing such a kind of reagent and testing its stability at 37uC, we chose the spiked soils as clinical samples to develop a protocol for detecting Y. pestis because contaminated soils by dead animals are common samples for Y. pestis detection [35] and Y. pestis could be persisted in soils for a long time, up to 40 weeks [36]; and also because soils contain rich compounds that inhibit enzymes, including Taq polymerase used in PCR [37,38,39,40]. A protocol for soil processing can be readily modified to treat the animal specimens by adding a step of proteinase K digestion. During soil treatment, we once compared the different lysis methods of direct heating lysis and NaI lysis. Direct heating method gave a better sensitivity than NaI lysis and hence we employed the former in our final protocol.
In the future application, we will distribute these reagents to different plague surveillance stations in China for their evaluation of its feasibility in field.