key: cord-0692774-mh5rkhsg
authors: Yan, Zhongqiang; Zhou, Lei; Zhao, Yongkai; Wang, Jing; Huang, Lihua; Hu, Kongxin; Liu, Haihong; Wang, Hong; Guo, Zhaobiao; Song, Yajun; Huang, Huijie; Yang, Ruifu
title: Rapid quantitative detection of Yersinia pestis by lateral-flow immunoassay and up-converting phosphor technology-based biosensor
date: 2006-12-07
journal: Sens Actuators B Chem
DOI: 10.1016/j.snb.2006.01.029
sha: 5b01c0f218d069ed312bf2ab678306f047da7734
doc_id: 692774
cord_uid: mh5rkhsg

Up-converting phosphor technology (UPT)-based lateral-flow immunoassay has been developed for quantitative detection of Yersinia pestis rapidly and specifically. In this assay, 400 nm up-converting phosphor particles were used as the reporter. A sandwich immumoassay was employed by using a polyclonal antibody against F1 antigen of Y. pestis immobilized on the nitrocellulose membrane and the same antibody conjugated to the UPT particles. The signal detection of the strips was performed by the UPT-based biosensor that could provide a 980 nm IR laser to excite the phosphor particles, then collect the visible luminescence emitted by the UPT particles and finally convert it to the voltage as a signal. V(T) and V(C) stand for the multiplied voltage units for the test and the control line, respectively, and the ratio V(T)/V(C) is directly proportional to the number of Y. pestis in a sample. We observed a good linearity between the ratio and log CFU/ml of Y. pestis above the detection limit, which was approximately 10(4) CFU/ml. The precision of the intra- and inter-assay was below 15% (coefficient of variation, CV). Cross-reactivity with related Gram-negative enteric bacteria was not found. The UPT-LF immunoassay system presented here takes less than 30 min to perform from the sample treatment to the data analysis. The current paper includes only preliminary data concerning the biomedical aspects of the assay, but is more concentrated on the technical details of establishing a rapid manual assay using a state-of-the-art label chemistry.

Yersinia pestis, a small Gram-negative rod belonging to the family Enterobacteriaceae, is the etiological agent of plague. This organism is one of the three human pathogenic species, including Y. pestis, Yersinia pseudotuberculosis, and Yersinia enterocolitica, in the genus Yersinia [1] . Plague occurs predominantly in three different forms, that is, bubonic, pneumonic, and septicaemic, depending on the route of exposure to the pathogen. Bubonic plague usually involves one group of lymph nodes, contracted following the bite by a flea that has previously fed on an infected rodent and developed a blocked proventriculus, and the infection can spread to the lungs (pneumonic plague) when the suitable treatment is not administered in time. Septicaemic plague is the result of haematogenous dissemination of the bacteria from a bubo. Pneumonic plague can also be transmitted by airborne droplets. The infection takes place within a few hours and causes bronchopneumonia. What is more, the Y. pestis is the most dangerous bacterial agent that could potentially be used for biological warfare or bioterrorism [2] .

Plague is a typical zoonosis distributed in Asia, Africa and America and it was redefined by the World Health Organization (WHO) as a reemerging infectious disease due to the increased 0925 outbreak of plague around the world [3] . The rapid detection of Y. pestis is critical for controlling the spread of plague. The laboratory detection of Y. pestis is based on bacteriological, serological and nucleic acid-based methods [1] . The bacterial culture is not well suitable for rapid detection of Y. pestis due to the time delay. The passive haemagglutination assay (PHA), which is used to detect Y. pestis-specific antibodies, has repeatedly shown to lack sensitivity and specificity [4] . Therefore, rapid and sensitive assays are desirable for the detection of Y. pestis, its specific antigens or nucleic acid sequences. At present, sensitive PCR (polymerase chain reaction)-based detection methods have been described to identify Y. pestis [5] [6] [7] . Especially, real-time PCR is highly specific and less prone to cross-contamination [8] . Although offering reasonable sensitivity, these assays are time consuming (include both DNA/RNA preparation and the longer detection procedure), and not readily available for onsite detection. Detection of F1 capsular antigen (F1 CA) of Y. pestis by immunoassay methods such as immunogold chromatographic dipstick assay and ELISA (enzyme-linked immunosorbent assay) has been developed [9, 10] .

Immunochromatographic assay, also called lateral-flow (LF) immunoassay, with benefits of low-cost, easy-to-use, rapid and sensitive detection of various analytes, has been developed for many years and mainly been used to detect drugs of abuse and for pregnancy testing at the early stage [11] . Now, it has been surging in infectious disease diagnostics [12, 13] . Jung et al. developed a colloidal gold particles-based LF strip for detecting Escherichia coli O157 at a minimum of 1.8 × 10 5 CFU/ml [22] . LF procedure utilizes the specific interaction between antigens and antibodies and provides rapid detection of various analytes. In the field of biomedical diagnostics, the search for the increased detection sensitivity and the possibility of quantitative detection using simple inexpensive assays is an ongoing challenge. In this respect, LF assays have become the popular diagnostic tools in a variety of settings because they are sensitive, simple to perform, inexpensive to manufacture, and perhaps most important, well suitable for rapid on-site detection and can be carried out nearly anywhere by non-technical personnel [14] .

UPT particles are submicrometer-sized, lanthanide-doped ceramic particles that possess anti-Stokes shift emission by upconverting infrared excitation light (980 nm) to emit visible light. Up-converting phosphors have also been used to realize quantitative assays and increase sensitivity. UPT reporters are 10-to 100-fold more sensitive than assays using conventional reporter systems such as colloidal gold or colored latex beads [15] . They have attracted considerable attention as the relatively novel luminescent labels, especially in LF assays. UPT has already been demonstrated for the sensitive detection of nucleic acids in microarray [16] , single-stranded nucleic acids in a sandwich-hybridization assay [17] and Streptococcus pneumonia in an amplification-free hybridization-based DNA [18] ; cell and tissue surface antigens by immunocytochemistry [19] ; human chorionic gonadotropin [15] , drugs of abuse, pathogenic E. coli by lateral-flow immunoassay [20] . In this study, up-converting phosphor is used as a reporter in rapid LF immunoassay for quantitative detection of Y. pestis.

Affinity-purified polyclonal antibody from rabbit against F1 antigen of Y. pestis and affinity-purified antibody from goat against rabbit IgG were prepared in our laboratory. Materials used in preparation of reagents, including HNO 3 (concentrated), tetraethylorthosilicate (TEOS), NH 4 ·OH, isopropanol, triethoxyaminopropylsilane (APES), CHCl 3 , ninhydrin, glycine, glutaraldehyde, sodium cyanoborohydride (NaCNBH 3 ), polyoxyethylene bis-amino (NH 2 -PEG-NH 2 , with an average molecule weight of 3350), Na 2 CO 3 , NaHCO 3 , NaCl, Na 2 HPO 4 , NaH 2 PO 4 , NaN 3 , albumin bovine V from bovine serum (BSA), p-nitrophenyl phosphate (pNPP), Tween-20, NP-40, sucrose, etc., were all obtained from Sigma Chemical Co. 

UPT particles (400 nm in diameter) were used to conjugate with the antibody against F1 antigen of Y. pestis. The biological molecules cannot be directly coupled to the untreated UPT particles. Therefore, a series of surface modifications and activations of UPT particles were required to realize covalent coupling of antibodies to them [14] . Firstly, UPT particles were coated with a thin layer of silica by using TEOS. Then the silica-coated surface of UPT particles can be functionalized with amino-, aldehyde-functional groups using APES, polyoxyethylene bis-amino and glutaraldehyde, respectively. Eventually these aldehyde-functionalized UPT particles can be conjugated to the antibody directly in a 4 • C-prechilled Na 2 CO 3 -NaHCO 3 buffer (50 mM, pH 9.5) under stirring. Uncoupled antibodies to the UPT particles were separated from the conjugated ones by a series of washing steps involving centrifugation and resuspension of pelleted UPT particles in a phosphate buffer of 30 mM pH 7.2. After several washes, UPT conjugates (1 mg/ml) were stored at 4 • C in 30 mM phosphate buffer [pH 7.2, containing 0.1% (w/v) BSA, 0.05% (v/v) Tween-20, 0.02% (w/v) NaN 3 ].

The shape of the test strip is rectangle with dimension of 74 mm × 4 mm. The components of the strip were the same as those used in immunogold chromatographic dipstick assay (Fig. 1) . To prepare LF strips, the sample pad (15 mm, (3), an absorbent pad (4), and a laminating card (5) . During the manufacture, a test line (6), a control line (7), and an ending index line (8) are disposed on the strip and other components are pasted on the laminating card with proper overlapping.

glass-fiber GFCP20300, Millipore), the conjugate release pad (10 mm, GFCP20300, Millipore), the nitrocellulose membrane (25 mm, Hi-Flow nitrocellulose, SRHF09002, Millipore) and the absorbent pad (30 mm, paper no. 470, Scheicher and Schuell) were mounted on the laminating card with proper overlaps. That is, the sample pad overlaps the conjugate release pad with 3 mm, the latter overlaps the nitrocellulose membrane with 1 mm, and the absorbent pad overlaps the nitrocellulose membrane with 2 mm.

In this study, UPT-antibody conjugates (1.0 mg/ml) were dispersed into the glass-fiber in a conjugate diluting phosphate buffer [pH 7.2, containing 1% (w/v) BSA, 10% (w/v) sucrose, 1% (v/v) Tween-20]. The conjugate release pad was incubated at 37 • C for 3 h. Using the dispenser (XYZ3050, Bio-Dot), the nitrocellulose was provided simultaneously with a capture line of rabbit polyclonal antibody IgG against F1 antigen of Y. pestis (1 mg/ml) and a control line of goat against rabbit IgG (1 mg/ml) at a dispensing rate of 0.1 l/cm. The nitrocellulose membrane was allowed to dry at 37 • C for 2 h. After the nitrocellulose membrane, the conjugate release, the sample and the absorbent pads were adhered on laminating card with special overlapping seriation, an automatic cutter (CM4000, Bio-Dot) was used to cut the assembled strips into 4 mm in width. The strips were used immediately or stored dry at room temperature until used. Finally, the strip was put into a disposable plastic cartridge, which was designed to fit to the stage of the UPT-based biosensor.

The portable biosensor (UPT-based biosensor) was codeveloped by Shanghai Institute of Optics and Fine Mechanics, Chinese Academy of Sciences and our institute (Fig. 2 ). Fig. 3 illustrates a design of the UPT-based biosensor, which is similar, at least partly, to what was presented by Zijlmans et al. and Niedbala et al. [19, 20] . The mainboard PC104 installed inside the biosensor was used instead of the accessory computer to control the whole system and the convenient Windows 98 system was employed as the platform to support the running of biosensor software programmed by us. Therefore, just equipped with the keyboard and mouse, the UPT-based biosensor can work independently, and the user can accomplish the scanning process just by pressing the icons displayed on the LCD of biosensor. During detection, firstly the strip was inserted in the biosensor and put on the stage, and then PC104 through multifunction DAS (Data Acquisition Station) card made the laser diode emit the infrared light with wavelength about 980 nm. This light irradiated on the dichroic mirror A, and a part of this light passed through it to the photodiode and gave an electric signal to collimate the instability of the system. The rest light was refracted and irradiated on the dichroic mirror B. This dichroic mirror made the infrared light refracted totally on the focusing lens A and through it focused on the strip. UPT particles on the strip were excited by this infrared light and then emitted visible light with wavelength about 541 nm. This visible light through focusing lens A irradiated on the dichroic mirror B and was transmitted completely. After filtrated away the light with other wavelength by a filter, the visible light emitted by UPT particles was focused on the PMT and gave multiplied voltage (relative fluorescence unit) as a signal. At the ending of data collection on this point, the stage held the strip to slide along X-axis to the next point of the nitrocellulose membrane and then another data-collecting process began. In the whole process of strip-scanning, the biosensor set 50 datacollecting points per millimeter of the nitrocellulose membrane, that is, 2000 points per strip. Therefore, the integrated scan- ning curve representing the diffusion of UCP particles on the nitrocellulose membrane was obtained. Voltage unit was plotted against the strip length increment to identify the height of peaks.

In conclusion, this UPT-based biosensor with an infrared laser (980 nm) was used to scan the resulting test strips. Upon IR excitation, the green light with the wavelength of 541 nm emitted by UPT particles was collected and detected by the UPT-based biosensor. Luminescence was measured, compiled and displayed as voltage units in a 2D plot as a function of the position of the strip. The software used for analysis included algorithms for background subtraction and peak-height determination.

Sensitivity testing was conducted to determine the detection limit of the UPT-LF immunoassay for Y. pestis by using pure cultures of Y. pestis EV76 as a model. For preparation of a bacterial suspension, pure cultures of Y. pestis EV76 in LB broth were centrifuged at 8000 rpm for 10 min, then the cell pellet was washed once, resuspended and serially diluted in PBS (137 mM NaCl, 2.7 mM KCl, 4.3 mM Na 2 HPO 4 , 1.4 mM KH 2 PO 4 , pH 8.0). In order to determine the viable cell number (CFU/ml) in the diluted samples, aliquots of 100 l from 10 −4 , 10 −5 , 10 −6 and 10 −7 dilution tubes were plated in duplicate onto LB agar plates. For detection by the UPT-LF strip, 100 l of each sample containing different concentrations of Y. pestis ranging from 10 2 to 10 9 CFU/ml was mixed with 900 l of a dilution buffer [2% (w/v) BSA, 0.5% (v/v) Tween-20, 0.5% (v/v) NP-40, in PBS] by gently stirring for 15 min in 1.5-ml microcentrifuge tubes. One hundred microlitres of each sample was applied to the sample window of the test strip and the same amount of the dilution buffer was added as blank control in triplicate. The LF immunoassay lasted for approximately 15 min at room temperature, and then the strip was scanned by the UPT-based biosensor for the results. The cut-off value of this assay was calculated as the mean ratio V T /V C of the blank controls plus three standard deviations.

The bacterial strains of Y. enterocolitica, Y. pseudotuberculosis and other Gram-negative enteric bacteria (E. coli and S. choleraesuis) were used to test the specificity of the UPT-LF assay. Overnight cultures of these bacteria were centrifuged and diluted either in PBS to determine CFU/ml by plate counting or in the dilution buffer for the UPT-LF assay. Cultures diluted in the dilution buffer from 10 0 to 10 7 CFU/ml were tested as described above. 10 8 CFU/ml of Y. pseudotuberculosis spiked with 10 4 to 10 8 CFU/ml of Y. pestis EV76 in the dilution buffer was used for evaluating the interference of other bacteria on the sensitivity and specificity of the assay.

To quantify the cell number in an unknown sample, serial 10fold dilutions of pure cultures of Y. pestis EV76 ranging from 10 8 to 10 4 CFU/ml in the dilution buffer were assayed. Each cell concentration was detected in six repeats, and then the mean V T /V C ratio was plotted against the logarithmic cell number for constructing the standard curve.

Blind experiment was applied to examine the accuracy of the UPT-LF further. Ten different dilutions of pure cultures of Y. pestis EV76 in PBS were prepared and coded randomly for an assay operator by another person. The UPT-LF was performed by the same batch of the strips that were used to construct the standard curve and each dilution was detected in six repeats.

Balb/c mice were infected through intranasal inoculation of 10 l of a phosphate buffer 30 mM pH 7.2 containing 10 7 CFU/ml of Y. pestis strain 201. Y. pestis 201 was isolated from Microtus brandti in Inner Mongolia, China. It has major phenotypes, such as F1 + (able to produce fraction 1 antigen), VW + (presence of V antigen), Pst + (able to produce pesticin) and Pgm + (pigmentation on Congo-red media). Strain 201 has an LD 50 of less than 100 cells for mice by subcutaneous challenge. It belongs to a newly established Y. pestis biovar, microtus [21] . Bacterial CFU were enumerated by plating serial dilutions of lung tissue homogenates from experimentally infected Balb/c mice on LB agar at 28 • C for 36 h. Fifty-one lung tissue homogenates from experimentally infected Balb/c mice were tested in triplicate by the UPT-LF immunoassay.

In this study, our aim was to develop and optimize a rapid, specific and sensitive UPT-LF immunoassay to quantitatively detect the whole cells of Y. pestis using pure cultures of Y. pestis EV76 as a model. Optimal results were obtained with the purified rabbit polyclonal antibody IgG against F1 antigen of Y. pestis immobilized on the nitrocellulose membrane as a capture antibody and the same UPT-antibody conjugate as a detector antibody in a sandwich immunoassay.

We set up experiments for the dose-response curve with serial dilutions of pure cultures of Y. pestis EV76 from 10 1 to 10 8 CFU/ml in the dilution buffer. The raw data (data not shown) and background subtraction data (Fig. 4) could be obtained by the UPT-based biosensor. The first peak (C) and the second peak (T) in Fig. 4 displayed the multiplied voltage units for the control and the test line, respectively, as plotted in Y-axis. The distance from the scanning window of cartridge was plotted in X-axis. It should note that absolute values of these signals depend on the physical properties of test strips and the photomultiplier tube used for detection and cannot necessarily be used to compare data from different experiments [15] . The variability observed from strip to strip at the same cell concentration attributes to 

, and the ratio of V T /V C for each scanning, and then a mean value ± S.D. was calculated for repeated tests. In order to get the quantitative result, the ratio V T /V C was treated as the final result for each assay. The mean ratio of the V T /V C values corresponding to serial dilutions of bacterial samples ranging from 10 1 to 10 8 CFU/ml was plotted against the logarithmic cell number (CFU/ml). The cut-off value of the current assay was calculated as the mean ratio (V T /V C ) of the blank controls plus three standard deviations, which corresponded to 99% confidence. The dose-response curve was obtained by plotting the ratio V T /V C as the Y-axis against logarithmic cell count at the X-axis (Fig. 5) .

The key feature of UPT is a high signal-to-noise ratio due to the elimination of autofluorescence interference. UPT particles, in conjunction with the portable UPT-based biosensor, offer increased sensitivity when used in a rapid LF format. The UPTbased biosensor provides a simple instrument to read assays by avoiding the possibility of subjective errors compared to interpreting a visual strip.

In order to test the sensitivity of the UPT-LF, serial dilutions of Y. pestis EV76 in the dilution buffer were tested with UPT-LF (Fig. 5) . The detection limit of the current assay was 10 4 CFU/ml, which was below the levels of Y. pestis that may be found in the blood of mice or human infected with Y. pestis. In mice, bacterial levels in blood may reach >10 4 CFU/ml and in humans it may reach 10 7 CFU/ml. The ratio V T /V C for 10 4 CFU/ml was easily distinguished from the mean ratio of V T /V C of the blank controls plus three standard deviations (cutoff value). The UPT-LF immunoassay is 10-fold more sensitive than the colloidal gold particles-based immunochromatography [22] .

Our data showed that the high level of other bacteria (10 8 CFU/ml of Y. pseudotuberculosis) in the dilution buffer did not interfere with the detection of low levels of Y. pestis EV76. Y. pseudotuberculosis did not cross react with antibodies specific for F1 antigen of Y. pestis at the cell concentrations tested. The specificity of the UPT-LF immunoassay was 100%, which was further demonstrated by showing no cross-reactivity to three strains of Y. pseudotuberculosis, three strains of Y. enterocolitica, two strains of E. coli and one strain of S. choleraesuis.

To demonstrate the intra-assay variance, a series of cell concentrations from 10 4 to 10 8 CFU/ml was repeatedly tested by using the same batch of test strips. The coefficient of variation ranged from 6.43% to 11.74% (n = 20). The inter-assay variance was determined by repeating the assay at six different days by the same investigator. These experiments were performed in parallel at our laboratory and at the National Academy of Inspection and Quarantine, China. The coefficient of variation for the cell concentrations above the cut-off value (10 4 CFU/ml) varied between 9% and approximately 12% (Table 1) . To evaluate the scanning process by the UPT-based biosensor, the test strip was scanned repeatedly (n = 10), resulting in an average CVs of less than 5% in peak height for test and control line reads at any time.

A 10-fold serial dilution of pure cultures of Y. pestis EV76 ranging from 10 8 to 10 4 CFU/ml in the dilution buffer was assayed and the ratio V T /V C was plotted against the logarithmic cell number for the standard curve. A reliable correlation of coefficient (r) was observed between them (r = 0.97857), and a good linearity was displayed throughout the entire measuring range (Fig. 6) . By applying the corresponding standard curve, the ratio V T /V C of each unknown sample could be converted into the cell number in the sample. To examine the accuracy further, we tested ten different samples by blind experiment. The results Table 1 Inter-assay variation of the UPT-LF immunoassay system

Mean ratio (V T /V C ) ± S.D. of quantitative analysis showed a good correlation (r = 0.96842, p < 0.001). We also tested 51 lung tissue homogenates from the experimentally infected Balb/c mice by the UPT-LF and calculated bacterial CFUs by converting the ratio into the cell count. The correlation between the UPT-LF assay and the plate counting was evaluated with these 51 lung tissue homogenates. There was a significant correlation (r = 0.92406, p < 0.001) between the two assays (Fig. 7) .

This study demonstrates the feasibility of combining a UPTbased biosensor with an LF immunoassay to create a rapid, on-site testing method for quantitative detection of the whole cells of Y. pestis. We developed and evaluated this assay by using pure cultures of Y. pestis EV76 as a model. By using a sandwich format, we observed a good linearity between the signal ratio (V T /V C ) and log CFU/ml of Y. pestis above the detection limit, which was approximately 10 4 CFU/ml. So the quantitative detection of Y. pestis could be easily achieved by the UPT-LF and it was confirmed by applying it to testing the lung tissue homogenates from the experimentally infected Balb/c mice by Y. pestis. The precision of the intra-and inter-assay was below 15% (coefficient of variation, CV). The specificity of this assay was also satisfactory because no cross-reactivity with related Gram-negative enteric bacteria was found. Furthermore, 10 4 to 10 8 CFU/ml of Y. pestis could be specifically and sensitively detected when 10 8 CFU/ml of Y. pseudotuberculosis were mixed together in the sample. The assay, which does not need washing and incubation steps, can be completed within 30 min.

To our knowledge, this is the first report of the combined use of a UPT-based biosensor and an LF immunoassay for quantitative detection of the whole cells of Y. pestis. The method described here is rapid, specific and sensitive; and it has the potential to be optimized for greater sensitivity for on-site detection. Therefore, the described method can be used for detection and monitoring of plague in combination of an LF immunoassay and a UPT-based biosensor. Moreover, the UPT-LF assay strips can be easily adapted for detecting other microorganisms, as we have applied this assay to detection of E. coli O157:H7, Bacillus anthracis spores and SARS-CoV etc. 

Yersinia pestis-etiologic agent of plague

Potential biological weapons threats

Comparison of passive haemagglutination and enzyme-linked immunosorbent assay for serodiagnosis of plague

A new method for plague surveillance using the polymerase chain reaction to detect Yersinia pestis in fleas

5 Nuclease PCR assay to detect Yersinia pestis

A combination of different polymerase chain reaction (PCR) assays for the presumptative identification of Yersinia pestis

Rapid detection of Yersinia pestis with multiplex real-time PCR assays using fluorescent hybridization probes

Early diagnosis of bubonic plague using F1 antigen capture ELISA assay and rapid immunogold dipstick

Evaluation of a standardized F1 capsular antigen capture ELISA test kit for the rapid diagnosis of plague

Enzyme immunochromatography-a quantitative immunoassay requiring no instrumentation

An immunochromatogarphic serological assay for the diagnosis of Mycobacterium tuberculosis

The establishment of highly sensitive monoclonal antibodies for influenza viruses and development of the new lateral-flow test strip device

Multiphoton up-converting phosphors for use in rapid immunoassays

In-Vitro Diagnostic Instrumentation

Up-converting phosphor reporters in immunochromatographic assays

Up-converting phosphor reporters for nucleic acid microarrays

Lateral-flow and up-converting phosphor reporters to detect single-stranded nucleic acids in a sandwich-hybridization assay

An amplification-free hybridization-based DNA assay to detect Streptococcus pneumoniae utilizing the up-converting phosphor technology

Detection of cell and tissue surface antigens using up-converting phosphors: a new reporter technology

Detection of analytes by immunoassay using up-converting phosphor technology

Genetics of metabolic variations between Yersinia pestis biovars and the proposal of a new biovar, microtus

Development of a rapid immunochromatographic strip for detection of Escherichia coli O157

We highly appreciate Ms. Jin Wang, Hongxia Wang and Xiangxiu Qin for their assistant in antibody preparation. This work was financially supported by the National High