Ultra-Fast and Sensitive Detection of Non-Typhoidal Salmonella Using Microwave-Accelerated Metal-Enhanced Fluorescence (“MAMEF”)

Certain serovars of Salmonella enterica subsp. enterica cause invasive disease (e.g., enteric fever, bacteremia, septicemia, meningitis, etc.) in humans and constitute a global public health problem. A rapid, sensitive diagnostic test is needed to allow prompt initiation of therapy in individual patients and for measuring disease burden at the population level. An innovative and promising new rapid diagnostic technique is microwave-accelerated metal-enhanced fluorescence (MAMEF). We have adapted this assay platform to detect the chromosomal oriC locus common to all Salmonella enterica subsp. enterica serovars. We have shown efficient lysis of biologically relevant concentrations of Salmonella spp. suspended in bacteriological media using microwave-induced lysis. Following lysis and DNA release, as little as 1 CFU of Salmonella in 1 ml of medium can be detected in <30 seconds. Furthermore the assay is sensitive and specific: it can detect oriC from Salmonella serovars Typhi, Paratyphi A, Paratyphi B, Paratyphi C, Typhimurium, Enteritidis and Choleraesuis but does not detect Escherichia coli, Pseudomonas aeruginosa, Klebsiella pneumoniae, Streptococcus pneumoniae, Haemophilus influenzae or Acinetobacter baumanii. We have also performed preliminary experiments using a synthetic Salmonella oriC oligonucleotide suspended in whole human blood and observed rapid detection when the sample was diluted 1∶1 with PBS. These pre-clinical data encourage progress to the next step to detect Salmonella in blood (and other ordinarily sterile, clinically relevant body fluids).

Invasive Salmonella spp. are routinely detected by standard blood culture techniques. Culturing blood specimens has become much faster and easier, since the advent of continuously monitoring blood culture instruments such as the BACTEC 9000 systems (Becton Dickinson, Cockeysville, MD, USA) and BacTAlert (BioMerieux, Durham, NC, USA). However, it still takes several days to detect and identify Salmonella [23,24]. Due to the time required for blood culture identification and the fact that many diagnostic laboratories are unable to serotype Salmonella spp. themselves, alternative methods of identification of Salmonella are being sought [25]. In particular, DNA detection methods such as the polymerase chain reaction (PCR) have been investigated. The food industry routinely uses PCR to detect Salmonella in food [26,27]. There are many reports of PCR primers designed to detect S. Typhi from the blood of enteric fever patients [28][29][30][31][32][33][34][35][36]. Furthermore, the sensitivity of PCR has often been found to be higher than that of blood culture [30,31,33,34]. However, PCR has not yet become an established method for diagnosis of typhoid fever [25]. One reason for this may be that although some reports claim high sensitivity, with detection of as few as 10 CFU/ml of blood [35], Wain et al.'s prospective study of the concentration of S. Typhi in blood of typhoid fever patients showed a median value of 0.3 (range of 0.1 to 399) CFU/ml, well below current PCRbased detection limits [37]. Interestingly, in another study, Wain et al. [38] showed that 63% of the S. Typhi cells were located in the buffy coat layer (presumably in monocytes and polymorphonuclear leukocytes) and the mean number of bacteria per infected leukocyte was 1.3 CFU/cell. These quantitative studies of Wain et al. corroborate a classic early study of Watson [39], who showed a median of 6 CFU/ml of S. Typhi in 15 patients with typhoid fever. Gordon et al. [40] have shown that NTS in bacteremic patients are present at a similarly low concentration (1 CFU/ml).
A promising new rapid diagnostic technique is microwaveaccelerated metal-enhanced fluorescence (MAMEF) [41,42], which integrates metal-enhanced fluorescence using surface deposited silver nanoparticles (to amplify fluorescence signatures) with low power microwave heating (to accelerate biomolecular recognition events kinetically). MAMEF has detected DNA from Bacillus anthracis spores and vegetative cells within 1 minute, which included a ,30 second spore lysing and sample preparation time [43,44]. The target was a highly conserved region within the gene encoding protective antigen (PA) [43]. The process was accelerated by using low-power microwave heating. MAMEF has also detected DNA from less than 100 CFU/ml of Chlamydia trachomatis in 40 seconds [45]. MAMEF is an ultra-fast, sensitive, and specific assay, using relatively simple but cost-effective technology that can be performed in a 96-well plate [46] and that can be multiplexed [47].
There is a pressing need for a sensitive and specific rapid diagnostic test to detect NTS bacteremia. In this paper, we describe adaption of the ultra-fast, highly sensitive MAMEF technology to detect the oriC locus, that we currently use as a Salmonella-specific target in PCRs [11], of Salmonella spp. which have been rapidly lysed by focused microwave radiation. This work establishes the proof of concept that MAMEF can be used to detect Salmonella directly from blood.

Bacterial strains, blood and genomic DNA
We constructed an attenuated Salmonella Enteritidis strain CVD 1940 (pGEN206) carrying a deletion in the chromosomal guaBA operon of wild-type strain R11 and also carrying plasmid pGEN206 for developing and optimizing microwave lysis and the MAMEF assay (unpublished results). CVD 1940 is unable to synthesize guanine nucleotides and is highly attenuated, therefore making it safer to handle than wild-type S. Enteritidis. The parent strain, S. Enteritidis R11, was isolated from the blood of a child in Mali [11]. Plasmid pGEN206, which expresses the fluorescent reporter protein GFPuv, encoded by gfpuv, was used to monitor efficient lysis of bacteria by fluorescence microscopy. Other Salmonella strains tested included S. Typhi (2 strains), S. Paratyphi A (2 strains), S. Paratyphi B (2 strains), S. Paratyphi C (1 strain), S. Typhimurium (2 strains), S. Enteritidis (2 strains), S. Dublin (2 strains), S. Choleraesuis sensu stricto (1 strain), S. Choleraesuis var. kunzendorf (1 strain) and S. Newport (1 strain) that came from Salmonella collections at the CVD or the Salmonella Reference Laboratory at the Centers for Disease Control and Prevention (Atlanta, GA, USA). Non-Salmonella strains used to determine specificity include Escherichia coli (2 strains), Pseudomonas aeruginosa (1 strain), Klebsiella pneumoniae (1 strain) and Streptococcus pneumoniae (1 strain each of serovars 6b, 14, 19F and 23) and Haemophilus influenzae type b (2 strains). The non-Salmonella strains were all from collections stored at the CVD. Salmonella spp., E. coli, P. aeruginosa and K. pneumoniae were grown in an animal product-free (APF) LB Lennox medium (APF-LB; Athena Environmental Sciences, Baltimore, MD, USA) at 37uC. Media were supplemented with guanine (0.001% w/v) and 50 mg/ml carbenicillin for growth of CVD 1920 (pGEN206). S. pneumoniae was grown on Columbia agar containing 5% Sheep Blood (BD, Franklin Lakes, NJ, USA) or in Brain Heart Infusion (BHI) broth at 37uC with 5% CO 2 .
H. influenzae was grown on BHI containing 10 mg/ml nicotinamide adenine dinucleotide (NAD) and 10 mg/ml hemin anaerobically using the AnaeroPack System (Mitsubishi Gas Chemical Co., Tokyo, Japan). Whole human blood containing sodium heparin as an anticoagulant was purchased from Innovative Research (Novi, MI, USA). Genomic DNA from Acinetobacter baumanii isolates 10 and 11 were obtained from BEI Resources (Manassas, VA, USA).

DNA methods
Genomic DNA was isolated using the WIZARD SV Genomic DNA Purification system (Promega, Madison, WI, USA) according to the manufacturer's instructions. DNA was visualized by agarose gel electrophoresis and staining with ethidium bromide. DNA released from lysed bacteria was visualized by agarose gel electrophoresis following concentration by ethanol precipitation. Briefly, 450 ml of lysate were concentrated by ethanol precipitation. The DNA pellet was resuspended in 20 ml TE (10 mM Tris and 1 mM EDTA [pH 8.0]) and the entire volume was electrophoresed on a 0.9% agarose gel.

Deposition of gold triangles on glass substrates to lyse Salmonella
Glass microwave slides were covered with a mask (12.5 mm in size with a 1 mm gap between two triangles), leaving a bowtie region exposed. Equilateral gold triangles of 12.5 mm were subsequently deposited onto glass microscope slides through the mask using a BOC Edwards 306 vacuum deposition with vacuum 3.0610 26 Torr, with a deposition rate of ,1 A/s. Four layers of self-adhesive silicon isolators (D 2.5 mm) were placed over the gold bowtie region to create a sample well (see Figure 1A).

Measurement of temperature over the gold bowtie
Since sapphire transmits infrared radiation, it is an ideal substrate for thermal imaging experiments. A slide coated with gold triangles and containing water in the sample well was covered by a sapphire plate. Using this sapphire and metal sandwich configuration, we determined the average temperature increase of the water in proximity to the metal. The optical configuration consisted of a microwave cavity with a 1 inch diameter opening at the base, a gold mirror, and a thermal imaging camera (Silver 420 M; Electrophysics Corp, Fairfield, NJ, USA) that was equipped with a lens that provided a resolution of approximately 300 mm ( Figure S1). The clear sapphire plate of the sandwich geometry was fixed to the base of the microwave cavity opening. A gold mirror was positioned such that the image of the opening was reflected onto the thermal camera. Thermal imaging data was recorded at 100 frames/second before, during and after the application of microwave pulses.

Lysis by microwave radiation
Bacteria were lysed using gold bowtie deposits on a glass slide and heating in a GE microwave Model No. JE2160BF01, kW:1.65 (M/ W), for 13 seconds on maximum power. 2 ml bacterial suspensions were placed into wells sterilized by rinsing with 70% ethanol and air-drying. Overnight cultures of bacteria were diluted to the required concentration in APF-LB and lysed. Viable counts of unlysed samples were performed by spread-plating up to 200 ml of bacterial suspension on agar plates or by using the pour plate method (5 ml of the bacterial suspension was added to 20 ml of molten agar and poured into a petri dish). Viable counts of lysed samples were performed by spread-plating 100-200 ml on APF-LB and incubating the remaining suspension at 37uC overnight. A sample was considered to be fully lysed if the broth showed no turbidity.

Transmission Electron Microscopy (TEM)
TEM images of lysed bacteria were taken using an Electron Microscope Tecnai T12 microscope. Samples were drop cast onto Formvar carbon films on copper grids (400 mesh) by placing a droplet of a 10-ml aqueous sample solution on a grid. The grid was air-dried for 24 h.

Formation of silver island films (SiFs) on glass substrates
SiFs, which provide the metal component of 'metal-enhanced fluorescence', were prepared as previously published [48]. In a typical SiFs preparation, a solution of silver nitrate (0.5 g in 60 ml of deionized water) was put in a clean 100 ml glass beaker. 200 ml of freshly prepared 5% (w/v) sodium hydroxide solution and 2 ml of ammonium hydroxide were added to a continuously stirred silver nitrate solution at room temperature. Subsequently, the solution was cooled to 5uC by placing the beaker in an ice bath, followed by soaking the Silane-prep TM glass slides in the solution and adding a fresh solution of D-glucose (0.72 g in 15 ml of water). The temperature of the mixture was then allowed to warm to 40uC. As the color of the mixture turned from yellowgreen to yellowish brown, the slides were removed from the mixture, washed with water, and sonicated for 1 min at room temperature.

Anchor and fluorescent probes
We have successfully used previously published primers to detect oriC of Salmonella in several PCRs [11,49]. We therefore decided to use this region as a target for our Salmonella MAMEFbased assay. Probes specific for the oriC locus of Salmonella spp. were designed using the oriC sequence from S.

Preparation of MAMEF assay platform for detection of Salmonella DNA
SiFs-deposited glass slides were coated with self-adhesive silicon isolators, containing oval wells (2.0 mm D632 mm L619 mm W) prior to the assay fabrication and subsequent fluorescence experiments. 1 mM of thiolated anchor probe was incubated overnight at 4uC on the surface of SiFs-deposited glass slides in 16 TE Buffer. The thiolated anchor probe was covalently linked to SiFs via well-established self-assembled monolayer chemistry [50].

MAMEF-based Salmonella DNA assays
The MAMEF-based DNA capture assay was performed using 1 ml lysed Salmonella mixed with 1 ml TAMRA-labeled fluorescent probe overlaid onto the SiFs containing bound anchor probe and incubated for 30 seconds in a microwave cavity (a 0.7 cu ft, GE Compact Microwave Model: JES735BF, max power 700 W). The power setting of the microwave cavity was set to 2 (which corresponded to 140 W over the entire cavity).

Fluorescence Spectroscopy
Fluorescence emitted by the MAMEF-based DNA capture assay was measured using a 532 nm diode laser and a Fiber Optic Spectrometer (HD2000) from Ocean Optics, Inc. by collecting the emission intensity through a notch filter (532 nm).

Lysis of Salmonella
We tested a variety of configurations of gold deposited on glass slides and found that the gold bowtie configuration was the best in terms of its ability to effectively lyse Salmonella when heated in a microwave ( Figure 1A). Figure S1 shows that the temperature of liquid at the point where the two triangles apex is raised during microwave heating. Overnight cultures of S. Enteritidis CVD 1920 (pGEN206) were diluted to biologically relevant concentrations (up to 10 4 CFU/ml) in APF-LB and 2 ml were lysed by microwave radiation. As can be seen in Figure 1B, a range of 2 CFU/ml to 1.3610 4 CFU/ml could be efficiently lysed (98%68% lysis; mean 6 standard deviation) using gold lysing triangles and heating in a microwave. DNA released from an overnight culture (1.9610 10 CFU/ml) of CVD 1920 (pGEN206) lysed using gold lysing triangles and a microwave was fragmented into a range of sizes with most of the fragments around 100 bp, as shown in Figure 1C. Examination of lysed bacteria by electron microscopy showed bacteria with blurred edges surrounded by clumps of lysed debris; bacteria from unlysed samples showed distinct edges against a clear background ( Figure 1D).

Detection of a synthetic oriC oligonucleotide by MAMEF
As mentioned earlier, oriC is used as a target in several Salmonella PCRs [11,51,52]. Figure S2 shows the primer binding sites in S. Typhimurium LT2 (GenBank Accession No. AE006468) of two different oriC primer pairs, as well as the location in bold of the segment of DNA targeted by our fluorescent and anchor probes. This oriC target DNA is conserved in all serovars of Salmonella enterica subsp. enterica except for S. Agona, S. Gallinarum and S. Dublin ( Figure S2). We therefore designed anchor and fluorescent probes so that they can detect all Salmonella enterica subsp. enterica serovars including S. Agona, S. Gallinarum and S. Dublin. Figure 2 shows a schematic diagram of the anchor and fluorescent probes binding to the target DNA. We first tested the ability of the probes to detect a synthetic oligonucleotide target. As can be seen in Figure 3, the target DNA was detected in a concentrationdependent manner, successfully detecting oriC at concentrations ranging from 500 nM all the way down to 0.5 nM.

Detection of DNA released from lysed Salmonella
Once lysis and detection were optimized, we performed the two methods sequentially to detect DNA using MAMEF from Salmonella that had been lysed by microwave radiation. Overnight cultures of CVD 1920 (pGEN206) were diluted to biologically relevant concentrations (up to 10 3 CFU/ml) in APF-LB and 2 ml were lysed by microwave radiation. 1 ml of the lysed bacteria was tested by MAMEF. We tested serial dilutions of two independent samples, with each dilution lysed separately. The first consisted of 10-fold serial dilutions of a 10 3 CFU/ml suspension and lysis results are shown in Figure 4A. The second dilution series consisted of 2-fold serial dilutions of a 12 CFU/ml suspension, with results shown in Figure 4B. Once again, the intensity of the fluorescence signal was concentration dependent and we were able to detect S. Enteritidis at concentrations down to 1 CFU/ml ( Figure 4B).

Specificity of the Salmonella MAMEF assay
To demonstrate that the assay is specific for detecting only Salmonella DNA, we tested DNA from a variety of Salmonella serovars and other bacteria commonly isolated from blood. Sixteen Salmonella strains of various serovars were suspended in APF-LB media at 10 3 -10 4 CFU/ml and 2 ml of each suspension were lysed using microwave radiation; every strain was lysed (93%616%, mean 6 standard deviation). 1 ml of the lysed sample was then assayed by MAMEF and a fluorescent signal was observed for every strain tested (Table 1). When 2 ml of 10 4 CFU/ml suspensions of E. coli, P. aeruginosa and K. pneumoniae were lysed by microwave lysis and tested by MAMEF, no fluorescence signal was observed. We also tested genomic DNA (diluted to 100 pg/ml) from these 3 species as well as DNA from S. pneumoniae, H. influenzae and A. baumanii and did not observe any detection ( Table 1).

Detection of Salmonella DNA in blood
The overall goal of our research efforts is to detect Salmonella directly from blood. We have performed some preliminary experiments using blood spiked with the synthetic oriC oligonucleotide and were not able to observe fluorescence ( Figure 5). However, when the spiked blood sample was diluted with an equal volume of PBS, the synthetic oriC target was readily detected by MAMEF.

Discussion
We have developed a MAMEF-based Salmonella assay capable of lysing and detecting 1 CFU suspended in 1 ml of bacteriological medium. The time to detection (excluding processing time) was only 30 seconds. This level of speed and detection limit greatly surpasses all currently available assays. For example, Nga et al. [53] have recently described a multiplex real-time PCR assay that targets S. Typhi and S. Paratyphi A. The sensitivity of the assay on blood samples was low with only 42% sensitivity for S. Typhi and 39% sensitivity for S. Paratyphi A. This poor sensitivity is most likely due to the poor detection limit of the assay. The authors prepared serial dilutions of S. Typhi in blood and observed detection limits of 250 CFU/ml for their real-time assay and 25 CFU/ml for detection by blood culture. The authors concede that identification of invasive Salmonella by PCR is not a practical approach.
Sensitivity of nucleic acid-based detection can be markedly increased by introducing an incubation step but only at the cost of greatly increasing the duration of the assay and its adaptation to becoming a point of care diagnostic test. For example, Zhou and Pollard [54] overcame low sensitivity due to small sample volumes by including a 3-hour incubation step in tryptone soya broth containing 2.4% ox bile, prior to detection of S. Typhi by PCR. They were able to achieve a detection limit of 0.75 CFU per ml of blood. However, the entire protocol takes almost 8 hours to complete. Moreover, the need for an incubation step in culture media means that this is no longer amenable to being a point of care diagnostic test. Similarly, by incorporating an overnight enrichment step, a fluorescence in situ hybridization (FISH) method for the detection of Salmonella spp. using a novel peptide nucleic acid (PNA) probe has achieved 100% sensitivity and 100% specificity and can detect 1 CFU per 10 ml of blood [55]. Because of the lengthy (overnight) enrichment step, this assay is also neither rapid nor adaptable to become the much needed point of care diagnostic for invasive Salmonella disease.
One molecular assay that does not require an enrichment step is the Lightcycler SeptiFast Test MGRADE kit (Roche Diagnostics, Germany), which is a commercial real-time PCR assay. This kit detects and identifies the 25 most common pathogens known to cause bloodstream infections directly from whole blood in ,6 hours but does not target Salmonella [56]. In serial experiments performed on EDTA-blood samples spiked with different concentrations of bacterial and fungal reference organisms, hit rates of 70-100% were achieved for 23 out of 25 organisms at 30 CFU/ml, but for only 15 out of 25 organisms at 3 CFU/ml. These results suggest that the assay may not be as sensitive as blood culture which has a theoretical sensitivity of 1 CFU.
However, two studies indicate that the Lightcycler SeptiFast is more sensitive than blood culture, as it was able to detect target DNA in several samples which were negative by blood culture [57,58]. Disadvantages of this method are that it includes a sample preparation step that requires the use of a centrifuge and the time taken to detection.
Recently, a MAMEF method showing detection of biotinylated BSA (b-BSA) was used to establish that MAMEF can be used to detect targets in complex biological samples such as blood [59]. In this study, phosphate buffer containing b-BSA was mixed with whole blood in a 1:1 ratio, and the target protein was detected using fluorophore-labeled avidin in 1 min. These findings and our own results show that detection of protein and DNA targets in blood is possible though presently the blood needs to be diluted to enable detection. We are currently investigating various methods to lyse red and white blood cells to reduce some of the viscosity of the liquid which we believe is impeding mass transport of the biological components to the surface during microwave-acceleration, and therefore overall fluorescence detection, and to release bacteria within the white blood cells.  The key points of the MAMEF technology that make it attractive for detection of pathogens in blood include: 1) it is a rapid and highly sensitive method; 2) the technology has been multiplexed, so that in one sample well, three DNA or protein targets can presently be identified within 20-40 seconds; 3) detection of well fluorescence can be accomplished by a variety of standard inexpensive sample well-reader technologies; 4) the assay platform requires no washing steps whatsoever to remove excess fluorescent probe or labeled DNA/antibody; 5) chambers that are disposable to minimize cross-contamination; 6) absence of centrifugation steps; and 7) the assay can be made quantitative by comparing levels of fluorescence to a standard curve. These attributes make it possible for the assay to be developed ultimately into a point-of-care device that can be used by people with minimal training.
The novel findings of this study are: 1) we have developed a sensitive and specific MAMEF-based Salmonella assay; 2) we have shown detection in 1 ml sample volumes which is greater than previously reported (and which provides the proof-in-principle that large volumes can be tested by MAMEF); 3) we can lyse and detect Salmonella without any centrifugation or washing steps; and 4) detection in blood is possible. We anticipate that we will be able to develop a multiplex MAMEF-based Salmonella assay that can efficiently detect the chromosomal oriC from blood-borne Salmonella and further determine whether the serovar is S. Typhimurium or S. Enteritidis, the two most commonly isolated NTS from invasive sites.   primer binding sites of oriC primers described by Widjojoatmodjo et al. [51]. (TIF) Author Contributions