Sage Journals: Discover world-class research

Abstract

The culture-independent and automated detection of bacteria in the environment is a scientific and technological challenge. For detection alone, a number of sensitive methods are known (e.g., PCR, enzyme-linked immunosorbent assay [ELISA], fluorescent in situ hybridization) but a major problem remaining is the enrichment and separation of the bacteria that usually occur at low concentrations. Here, we present an automated capturing and separation system, which can easily be combined with one of the sensitive detection techniques.

We have developed a method for enrichment and detection of Legionella pneumophila in liquid media. Concentrated microorganisms were either detected by PCR or by sandwich ELISA. The limit of detection with the immunological assay was about 750 bacteria. Using PCR, the equivalent of about 2000 genomes could be detected.

The assays were then transferred to a laboratory prototype for automated processing. It was possible to automatically enrich L. pneumophila by immunomagnetic separation (IMS), and again, the bacteria were detected by sandwich ELISA and PCR amplification of the ompS gene. As a novel aspect, ompS gene was used for the first time as a target for the detection of L. pneumophila on magnetic beads.

The aim of this work was to develop an automated procedure and a device for IMS of bacteria. With Legionella as a model organism, we could show that such a novel fully automated system can be an alternative to timeconsuming conventional cultivation methods for detecting bacteria or other microorganisms.

Keywords

immunomagnetic separation Legionella pneumophila automated processing sandwich ELISA ompS

Introduction

Legionella pneumophila is a thin, rod-like aerobic gram-negative bacterium with a thickness of 0.5–0.7 μm and a length of 2–20 μm. It belongs to the group of aerobic gram-negative bacteria and can be isolated from many terrestrial and aquatic habitats as well as from patients suffering from legionellosis. This disease, caused by L. pneumophila, derives its name from the fact that it was first recognized as a disease entity from an outbreak of pneumonia that occurred during a convention of the “American Legion” in the summer of 1976. Studies have shown that Legionella is an inhabitant of potable water systems, cooling towers of air conditioning systems, and waste-water. An infection occurs by inhalation of bioaerosols generated from such sources.^1,2

The standard detection methods for viable Legionella are time consuming and require a cultivation step of several days. Specific media are required for the cultivation, but it has been shown that the presence of background microorganisms can limit the sensitivity of detection. Fast detection of Legionella without cultivation but with high sensitivity is a scientific challenge and very important for applications, such as water monitoring. PCR methods are alternative tools for detection of microorganisms compared with the conventional culturing methods. A number of PCR techniques for detection of Legionella have been described in the literature. Most popular target genes for PCR amplification are the 5S rRNA gene, 16S rRNA gene, and the macrophage infectivity potentiator gene called mip. ^3,4 Another target that was used in combination with immunomagnetic separation is the defective organelle trafficking gene (dotA). In an assay based on the specific amplification by real-time PCR, a fragment of 80 bp was amplified to detect 50 different strains of L. pneumophila and additional 12 Legionella species.⁵ This study indicates that, beside the classic targets, other genes could also be a powerful tool for the detection of Legionella.

Fluorescent in situ hybridization is a sensitive and rapid detection method for microorganisms. Water samples can be filtered, and the captured bacteria are hybridized with a fluorescence-labeled probe complementary to 16S rRNA.^6,7

For rapid enumeration of different strains of viable Legionella, a double-staining method has been developed. It combines the specific detection by fluorescently labeled antibody together with cell staining with a viability marker.⁸

However, all these detection methods are not sensitive enough to detect bacteria at very low concentrations. One solution to this problem is an enrichment step by filtering the water sample, but often, the bacteria stick to the filter membrane, and it is not possible to completely recover them from the filter surface. Micromechanical filters have successfully been used as a tool for efficient filtration and enrichment of bacteria. Because of their absolutely flat surface, bacteria can be recovered and transferred to a biodetector system.⁹

Another technique for capturing microorganisms is immunomagnetic separation (IMS). IMS uses small superparamagnetic particles or beads coated with antibodies against surface antigens of pro- and eukaryotic cells. Many applications of IMS can be found in the literature (e.g., Refs. 10 and 11). This technique is also used in diagnostic applications to detect a wide spectrum of pathogens. After capturing the organisms with the beads, nearly all bioanalytical methods can be applied for detection, such as immunoassay, hybridization techniques, and PCR.^12,13 Furthermore, immunomagnetic cell sorting is also possible. For extracting mesenchymal stem cells from peripheral blood, an immunomagnetic cell sorter in combination with a mixer was developed by Inokuchi et al.¹⁴ In the first step, the cell suspension is mixed with antibody-coated paramagnetic beads. During mixing, the paramagnetic beads interact with the stem cells by antigen—antibody reaction. In the next step, the captured cells are separated from the medium by a magnetic separator.

A different capturing technology is to bind the cells to cationic magnetic beads. These positively charged beads interact with the negatively charged lipopolysaccharides on the surface of gram-negative bacteria. With the commercial capture system Pathatrix Enterobacter sakakii were isolated from contaminated infant milk formula. After the enrichment, the separated gram-negative bacteria were plated onto agar, and the number of colony-forming units (cfu) was determined. With this method, 1–5 cfu of E. sakakii in 500 g of powdered milk formula were detected.¹⁵

In this article, we present a novel automated system for capturing and enrichment of bacteria. Paramagnetic beads coated with a specific antibody against L. pneumophila were mixed in a vessel together with defined amounts of bacteria. Then the bacteria—bead complex was collected by a removable magnet as a novel element for processing paramagnetic beads. After collection, the pathogens were manually detected by sandwich ELISA and PCR. For PCR assay, we use the gene of the major outer membrane protein (ompS), which encodes for a surface protein composed of two subunits (28 and 31 kDa).^16,17 It was shown that the ompS DNA sequence is highly conserved among the different serogroups of L. pneumophila. This makes the gene a perfect target for detection of this pathogen. It was the first time to use ompS gene as a marker gene to detect L. pneumophila.

Materials and Methods

Bacteria, Genomic DNA, and Determination of Cell Numbers

Experiments were performed using inactivated L. pneumophila in a mixture of serogroups 1–14 (PRO-LAB Diagnostics, Richmond Hill, Canada) and L. pneumophila from InvivoGen (Toulouse, France). Cells were stained by gram staining to check for the shape and the quality of the cells (results not shown). Cell numbers were counted under the microscope using an improved Neubauer counting chamber. For PCR experiments, standard genomic DNA from L. pneumophila (Minerva Biolabs GmbH, Berlin, Germany) was diluted in a 10 × dilution series and used as PCR template. After PCR, the reaction products were loaded onto a 2% agarose gel, and DNA bands of the unknown sample were compared with the bands of the dilution series.

Immunoblot Assay (Western Blot) of Legionella pneumophila

A suspension of heat-killed L. pneumophila was denatured and subjected to sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE; 12% separating gel), as described by standard protocol.¹⁸ The separated proteins were electroblotted (semi-dry) onto a nitrocellulose membrane, which was then blocked overnight with 5% milk powder in Tris Buffered Saline (TBS-Tween) (20 mM Tris-HCl at pH 7.5, 0.5 M NaCl, 0.05% Tween-20). Polyclonal rabbit anti—L. pneumophila (Acris Antibodies, Herford, Germany; directed against whole cells of L. pneumophila, 4–5 mg/mL), in a dilution of 1:1000 (in TBS-Tween), were allowed to react with the immobilized proteins for 2 h at room temperature. After three washes with TBS-Tween, a goat anti-rabbit polyclonal antibody labeled with horseradish peroxidase (HRP) (Novus Biologicals, Littleton, USA; 1 mg/mL; 1:50,000 diluted in TBS-Tween) was added. After 30 min, the blot was washed four times with TBS-Tween and developed with diaminobenzidine, 0.5% NiCl₂ × 7H₂O, and hydrogen peroxide.

Coating of Magnetic Beads. 200 μL of streptavidin-coated beads (Dynabeads My One Streptavidin; Invitrogen, Darmstadt, Germany) in a concentration of 7–12 × 10⁹ per mL was washed five times in 1 mL phosphate-buffered saline (PBS) (0.2 g KCL, 0.2 g KH₂PO₄, 1.15 g Na₂HPO₄, 8 g NaCl, water to 1000 mL, pH adjusted to 7.4) with 10% bovine serum albumin (BSA) and 0.01% Tween-20 using a magnet. After incubation overnight (“blocking”) in the same buffer, they were resuspended in 200 μL PBS + 10% BSA + 0.01% Tween-20.

Twenty microliters of biotinylated polyclonal anti- Legionella antibody (4–5 mg/mL; Acris Antibodies) was added to 200 μL of blocked beads and incubated for 1 h at room temperature. Unbound antibody was removed by washing five times with 400 μL PBS + 10% BSA + 0.01% Tween-20.

Fluorescence Microscopy of Legionella pneumophila on Magnetic Beads. 1 × 10⁷ inactivated bacteria in a volume of 36 μL were mixed with 8 μL anti—Legionella-antibody-coated beads, and the solution was incubated for 1 h. The beads were then washed five times with 400 μL PBS + 10% BSA + 0.01% Tween-20. After removing unbound bacteria, 2-μL fluorescein isothiocyanate (FITC)-labeled antibody (4–5 mg/mL; Acris Antibodies) was added. After incubation, nonbound antibodies were removed by washing five times with 400 μL PBS +10% BSA + 0.0.1% Tween-20. After this step, captured bacteria were observed by fluorescence microscopy. Uncoated magnetic beads were used as negative control under the same conditions.

Electron Micrographs of Legionella pneumophila. A volume of 8-μL anti—Legionella-antibody-coated beads were mixed with 1 × 10⁷ heat-inactivated Legionella (PRO-LAB Diagnostics, Richmond Hill, Canada) in a volume of 36 μL. After incubation for 1 h, the beads were washed five times with 400 μL PBS + 10% BSA + 0.01% Tween-20 and resuspended in 100 μL of 4% glutaraldehyde. After fixation, the beads were washed again four times with 200 μL PBS. For complete dehydration, the beads were successively transferred into 200 μL each of increasing amounts of ethanol: 30–50–70–80–90% EtOH. After dehydration, the beads were spotted onto a glass slide and dried in an oven at 80 °C overnight. The dried beads were taken up with a double-sided adhesive tape, which in turn, was fixed to the sample holder of the electron microscope.

Immunodetection of Legionella pneumophila on Magnetic Beads. 40 μL of antibody-coated beads containing 6 × 10⁸ paramagnetic particles were mixed with 36 μL of different dilutions of heat-inactivated Legionella (PRO-LAB). The maximum amount in the dilution series was 7.4 × 10⁷ (per 36 μL), and the bacteria were diluted in decade steps down to an amount of 74 bacteria. After incubation for 1 h, the beads were washed again five times with 400 μL PBS + 10% BSA. For detection, we used an HRP-conjugated polyclonal anti-Legionella antibody (courtesy of Dr. Hero Brahms, DRG Instruments GmbH, Marburg, Germany) in a dilution of 1:750. After incubation and washing, HRP activity was determined.

Colorimetric Horseradish Peroxidase Detection. Two hundred microliters of HRP buffer (30 mM potassium citrate; adjusted to pH 4.1 with potassium hydroxide) is mixed with 10 μL substrate solution (240 mg tetramethylbenzidine in 100 mL of a 1:9 acetone:ethanol mixture; supplemented with 750 μL of 30% hydrogen peroxide). Enzymatic reaction was started with 2 μL of sample containing the beads with the immunological sandwich. After incubation for 15 min, the reaction was stopped by adding 100 μL of 1 M sulfuric acid, and the absorption at 450 nm was measured in a microplate reader (BMG Labtech GmbH, Offenburg, Germany).

Primer Design and Detection of Immunocaptured Legionella pneumophila by PCR. For PCR detection, the following primer pair was designed to amplify an 849-bp fragment of the ompS gene (GeneBank accession number: M76178): ompS_Leg_forward GCGGCTGTATTTGCTCTGGGAA and ompS_Leg_reverseTAAGCCTATGTAGGGGCCAGATGC (synthesized by Eurofins MWG Operon, Ebersberg, Germany).

PCR mixtures (48 μL) contained 10 mM Tris-HCl at pH 8.3, 50 mM KCl, 1.5 mM MgCl₂, 2.5 mM each of the four deoxynucleotides, 100 pmol of each primer, and 0.02 U of Taq DNA polymerase (New England Biolabs, Ipswich, USA), to which 2 μL of beads with captured Legionella was added as template. Controls were performed with no template and with a negative control solution from PRO-LAB.

PCR assays were conducted in a RoboCycler (Stratagene; Agilent Technologies, Santa Clara, USA) with the following thermal profile: initial denaturation for 5 min at 95 °C, followed by 30 cycles of denaturation at 95 °C for 20 s, annealing at 54 °C for 20 s, extension at 72 °C for 1 min, and final extension at 72 °C for 5 min. Amplified PCR products were separated in a 2% (wt/vol) agarose gel and stained with ethidium bromide. Agarose gels were prepared and run in 1 × Tris-acetate-EDTA buffer (TAE) (50x TAE: 242 g Tris-base, 57.1 mL acetic acid, 100 mL 0.5 M EDTA (Ethylenediaminetetraacetic acid), water to 1000 mL, pH adjusted to 8.5). PCR products were compared with a 1-kbp DNA molecular size ladder (New England Biolabs, Ipswich, USA) to estimate their size. Results were scored as positive or negative by visual inspection.

Construction of the Automated Separation System. The automated capturing and separation system was designed and assembled in collaboration with the companies Plugit AG (Hallau, Switzerland) and HTI bio-X GmbH (Ebersberg, Germany). Software development and implementation of all processing steps were done by Infoteam Software GmbH (Bubenreuth, Germany) and Plugit AG.

Results

Automated Immunomagnetic Enrichment of Legionella pneumophila

We have developed an automated system for capturing bacteria using paramagnetic beads (principle shown in Fig. 1A). Two pipettors-one for larger volumes up to 10 mL and the other for small volumes up to 300 μL-are mounted on independent z axes (Fig. 1B). Both axes can be moved in the horizontal direction by a common x axis. Sample vessel, pipette tips, and waste containers for liquid waste and for used tips are placed in a turntable (Fig. 1D and E). The system allows for automated replacement of large and small pipette tips. A novel key element is a permanent magnet that can be rotated to and from the reaction vessel (Fig. 1C). With this special processing element, a controlled movement of the collected beads in the tube is possible. Although in other magnetic separators the permanent magnet is either fixed or can only be moved in on and off positions, in our system, the magnet can be freely positioned. In a first step, the magnet is placed at one side of the vessel and the beads are collected at short distance but in a relatively large area of the wall. Then, the magnet is slowly rotated down dragging along the beads to a concentrated spot close to the bottom of the vessel. Holding the beads in this position, the supernatant can be taken off by the pipette. The procedure is the same for the large vessel/large pipette or the small vessel/small pipette. The wash cycle or the concentration step is completed by rotating the magnet completely away from the tube (third position) and resuspending the beads.

Figure 1.

Illustration of the automated separation system and schematic diagram of capturing by IMS. A) Schematic illustration of the capturing and detection of Legionella pneumophila. B) Photograph of the whole automated separation system. Turn table (1), common x-axis (2), z-axes with pipettors (3), washing/heating station with reaction tubes (4), micro plate (5). C) Automated separation of magnetic beads (6) in a reaction tube (7) by a movable permanent magnet (8). D) Sample vessel (10) pipette tips (9) and waste container (11) are placed in a turn table. (E) Detailed view of the automated separation system with tip release station (12), sample vessel (13), reservoir tubes (14), pipette tips (15), reaction tubes placed in the heating/washing station (16), micro plate (17).

Using a large and a small reaction vessel, magnetic beads from a sample volume of 10 mL can be washed and concentrated into any volume between 20 μL and 1 mL. An integrated heater enables denaturation of samples. As a last step of the process, the concentrated beads are transferred into a microplate for further processing. The whole system is controlled by custom-made software with an integrated editor for structured programming.

A typical processing cycle starts with an initial sample volume of 5 mL in the large vessel. 6 × 10⁸ beads are added to the sample and mixed with the large pipette. After an incubation time of 30 min, magnetic separation starts followed by a concentration of the beads into a volume of 1 mL. The beads are washed three times with 1 mL washing buffer (PBS with 10% BSA and 0.01% Tween-20) using the large pipette. The beads are collected again and resuspended in 1 mL washing buffer containing detection antibody conjugated with HRP (1:1000), followed by an incubation time of 20 min, including mixing every 5 min with the small pipette. The beads are collected again, resuspended in 300 μL buffer and transferred to the small reaction vessel. To remove unbound antibody, beads are washed six times with a volume of 300 μL. In the last processing step, beads are concentrated to a final volume of 100 μL and fractionated into five aliquots of 20 μL each in a microplate. These samples were either subjected to HRP reaction (immunoassay) or to PCR detection. The whole processing cycle for sandwich ELISA was about 2.5 and 2 h for processing for detection with subsequent PCR (without addition of detection antibody). This results in a total time of about 3 h for immunodetection (2.5 h washing and enrichment plus ca. 0.5 h HRP reaction and reading). In the case of PCR, about 4–5 h (including analysis of reaction products on a gel) have to be taken into account. The conventional detection of L. pneumophila by culturing takes between 4 and 11 days.² Comparing this long detection time with the short processing and immunodetection cycle of 3 h, the automated system has a great advantage over the culturing method. Even when using PCR, the automated system is still faster than conventional methods.

To estimate the sampling efficiency, a small aliquot of beads was taken out after each concentration step and counted under the microscope with a Neubauer improved counting chamber. From an initial amount of 6 × 10⁸ beads, 2.5 × 10⁸ beads were remaining after all concentration and washing steps, that is, 45% of the magnetic beads were left over for detection by sandwich ELISA or PCR. The loss of beads of 55% seems to be very high, but it results from the high number of washing steps. We found out that the major loss (approx. 40%) comes from the washing with the large pipette. Further improvements of the system are underway to reduce this loss. In summary, the sample volume of 5000 μL was reduced down to 100 μL (50-fold concentration). From this, a fraction of only 2 μL was used for the assays (i.e., HRP reaction or PCR).

Characterization of the Antigen for Assay Development. In a first experiment, an immunoblot assay (Western blot) with Legionella has been carried out. Inactivated Legionella were obtained from two different suppliers, InvivoGen and PRO-LAB. Samples from both sources were boiled in SDS-loading buffer and loaded in volumes of 5, 10, and 20 μL on an SDS gel together with a negative control from PRO-LAB. After separation and transfer of the proteins onto nitrocellulose membrane, the polyclonal antibody (raised against whole cells of Legionella) reacted with the surface proteins of L. pneumophila. The Western blot exhibited a strong recognition of the cells from PRO-LAB, but the antibody did not react with Legionella from InvivoGen. We assume that the bacterial antigen obtained from InvivoGen was modified by its inactivation process. Thus, the antibody could not recognize the target surface proteins from L. pneumophila. A weak cross-reaction was observed by the negative control, but the unspecific interaction can be neglected compared with the positive signal (Fig. 2). For further experiments and assay development, we only used the antigen from PRO-LAB.

Figure 2.

Definition of the antigen with the method of the Western blots. Immunoblot assay (Western blot) of Legionella pneumophila antigens from different suppliers, InvivoGen and PRO-LAB. Lysed cells in a volume of 5, 10 and 20 μL were separated in 12.5% SDS-PAGE and transferred onto nitrocellulose membrane. The antibody mainly recognised the bacterial antigen from PROLAB (middle lanes and arrow). M: molecular protein marker.

Fluorescence Microscopy and Electron Microscopy of Legionella pneumophila. For direct bacterial identification and as a further control for specific binding, Legionella was examined by fluorescence microscopy. Bacteria were captured with magnetic beads followed by incubation with a fluorescence-labeled antibody. Under the fluorescence microscope, Legionella shows the typical morphological characteristic of this species. Beads without Legionella were used as negative control under the same condition as the sample (Fig. 3A and B).

Figure 3.

Fluorescence and electron microscopy of Legionella pneumophila bound to beads. Photographs of Legionella pneumophila bound to paramagnetic beads. (A) Legionella pneumophila captured with magnetic beads and visualised with FITC labelled antibody. (B) Negative control. (C) electron microscopy photographs of Legionella pneumophila attached to magnetic beads. (D) Lysed Legionella pneumophila. A large fraction of the cells showed damaged cell walls.

In addition, electron micrographs were taken from heat-inactivated Legionella bound to paramagnetic beads (Fig. 3C). Pictures indicate that Legionella as antigen interacts excellently with antibody-coated beads. However, a large number of cells showed damaged membranes (Fig. 3D). Electron micrographs were taken from beads as negative control without Legionella (data not shown).

Colorimetric Sandwich ELISA of Legionella pneumophila on Magnetic Beads. Streptavidin beads were coated with biotinylated antibody against L. pneumophila. Initial experiments suggested that the antigen interacts unspecifically with the magnetic beads (data not shown). After blocking with 10% BSA, unspecific binding could not be observed any more. Captured Legionella were detected with a second polyclonal antibody conjugated with HRP and 3,3′,5,5′-tetramethylbenzidine as substrate. The limit of detection was 740 bacteria (Fig. 4A).

Figure 4.

Detection of Legionella pneumophila with Sandwich ELISA. Immunological detection ofLegionella pneumophila. A) Dilution series of Legionella. To determine the limit of detection, we take the mean signal of the blank (i. e. without bacteria) plus 3 times the standard deviation (red horizontal line). The number of bacteria corresponding to this value is 740. The green curve serves as guide to the eyes only. B) Detection of 1.7 × 10⁸ cells after automated capturing by the automated separation system.

In a next step, the assay was transferred to the automated separation system. After binding of Legionella to paramagnetic beads and several washing steps, the separated bacteria could again be detected by sandwich ELISA. The amount of Legionella used in this experiment with our automated laboratory prototype was approximately 1.7 × 10⁸. We did not determine the limit of detection (LOD) in this case (Fig. 4B).

PCR of Legionella pneumophila on Magnetic Beads. A conventional PCR assay was developed for the detection of L. pneumophila captured by paramagnetic beads. As molecular marker, we use the gene for the membrane protein ompS. From this target gene, an internal PCR product with a length of 849 bp was amplified. For capturing experiments, a dilution series of Legionella was made. The diluted bacteria were captured with the paramagnetic beads, and after stringent washing, 2 μL of the bead suspension was subjected to PCR reaction. As negative control, paramagnetic beads without template were used. To estimate the limit of detection, a 10x dilution series of standard genomic Legionella DNA was made. From these dilutions, the same 849-bp-long PCR fragment was amplified using the same master mix as for the immunocaptured bacteria. PCR products from genomic DNA and from the beads with captured bacteria were loaded in same volumes onto an agarose gel.

After staining, the bands from paramagnetic beads and from the genomic DNA dilution series were compared with each other, and the number of genomic DNA equivalents was estimated. In our PCR assay, the limit of detection was between 200 and 2000 genomes (Fig. 5A).

Figure 5.

PCR Detection of Legionella pneumophila with ompS as target gene. Qualitative and quantitative detection of Legionella pneumophila by PCR. A 849 bp long fragment of the ompS gene was amplified. (A) upper gel: Dilution series from a defined genome concentration: 2 × 10⁶ genomes (1), 2 × 10⁵ genomes (2), 2 × 10⁴ genomes (3), 2 × 10³ genomes (4), 2 × 10² genomes (5), 2 × 10¹ genomes (6), negative control: PCR without template (7), lower gel: Dilution series of Legionella pneumophila and estimation of the amount of bacteria. Undiluted Legionella, with signal similar to lane 3 of the upper gel, corresponding to approx. 2 × 10⁴ genomes (8) 1:10 dilution of Legionella, signal strength between signal of lane 4 and 5 of the upper gel, corresponding to approx. 2 × 10³ genomes (9), 1:100 dilution (10), 1:1000 dilution (11), 1:10.000 dilution (12), 1:100.000 dilution (13), negative control (PCR without template DNA) (14). (B) Detection of Legionella pneumophila by PCR after automatic separation. Dilution series from a defined genome concentration: 2 × 10⁶ genomes (1), 2× 10⁵ genomes (2), 2 × 10⁴ genomes (3), 2 × 10³ genomes (4), 2 × 10² genomes (5), 2 × 10¹ genomes (6), 2 × 10⁰ genomes (7), negative control: PCR without template (8), detection of approx. 2 × 10⁴ bacteria by automated enrichment (9), negative control: PCR without template DNA (10), negative control: PCR with PBS + 10% BSA instead of template (11).

Again, we used the automated separator for the immunomagnetic capturing step. Bacteria were washed several times and concentrated from 5-mL volume down to a volume of 100 μL. For detection, 2 μL of bead suspension was transferred to the PCR reaction, and the amount of amplification product was compared with PCR products from a dilution series of the Legionella genome. The amount of bacteria used in this experiment was 2 × 10⁴. Again, detection was possible using the automated enrichment system (Fig. 5B).

Conclusions

We have developed an automated capturing system for microorganisms. Its practical use could be demonstrated using L. pneumophila as an example. With the system, it is possible to automatically enrich bacteria by IMS. The prokaryotes can then be detected by sandwich ELISA or PCR.

We have also developed a sensitive PCR detection method for L. pneumophila by using an 849-bp-long fragment of the ompS gene as a target. Our capturing system is flexible and user friendly and can be expanded to a fully automated system to capture and detect the bacteria on the same platform.

Capturing and enrichment of bacteria and other microorganisms are necessary in environmental analysis, such as water and air monitoring. Our device is prepared for an aerosol collector being integrated. The complete system will then be able to concentrate airborne pathogens contained in several cubic meters of air into a small volume of about 20 μL of liquid.

Another advantage of using IMS is the possibility of multiplexing. By combining several types of beads coated with different antibodies, the number and type of target organisms can easily be adapted to the particular application.

The system can be regarded as a major step toward an automated standalone analyzer for environmental monitoring (water and air). It should also be possible to use it with other liquid-sample matrices, such as in food analysis (milk, beer, and others) and in diagnostics (blood, serum, urine, saliva, and others).

ACKNOWLEDGMENTS

The authors gratefully acknowledge that a part of this work has been supported within the frame of the German BMBF project ZentriLab (funding agency: VDI/VDE-IT GmbH, Germany). They appreciate the support from Dr. Müller and Dr. Waltenberger (MicroCoat Biotechnologie GmbH, Bernried, Germany) and, especially, from Dr. Hero Brahms (DRG Instruments GmbH, Marburg, Germany), who provided the HRP-conjugated polyclonal anti-Legionella antibody. The authors thank Plugit AG, HTI bio-X GmbH, and Infoteam Software GmbH for construction, hardware assembly, and software development.

Competing Interests Statement: The authors certify that they have no relevant financial interests in this manuscript, and that all financial and material support for this research and work are clearly identified in the manuscript.

References

Brenner

D. J.

Steigerwalt

A. G.

McDade

J. E.

Classification of the Legionnaires' disease bacterium: Legionella pneumophila, genus novum, species nova, of the family Legionellaceae, familia nova. Ann. Intern. Med. 1979, 90(4), 656–658.

Blackmon

J. A.

Chandler

F. W.

Cherry

W. B.

England

A. C.

3rd Feeley

J. C.

Hicklin

M. D.

McKinney

R. M.

Wilkinson

H. W.

Legionellosis. Am. J. Pathol. 1981, 103(3), 429–465.

Jonas

Rosenbaum

Weyrich

Bhakdi

Enzyme-linked immunoassay for detection of PCR-amplified DNA of legionellae in bronchoalveolar fluid. J. Clin. Microbiol. 1995, 33(5), 1247–1252.

Stolhaug

Bergh

Identification and differentiation of Legionella pneumophila and Legionella spp. with real-time PCR targeting the 16S rRNA gene and species identification by mip sequencing. Appl. Environ. Microbiol. 2006, 72(9), 6394–6398.

Yanez

M. A.

Carrasco-Serrano

Barbera

V. M.

Catalan

Quantitative detection of Legionella pneumophila in water samples by immunomagnetic purification and real-time PCR amplification of the dotA gene. Appl. Environ. Microbiol. 2005, 71(7), 3433–3441.

Baudart

Coallier

Laurent

Prevost

Rapid and sensitive enumeration of viable diluted cells of members of the family enterobacteriaceae in freshwater and drinking water. Appl. Environ. Microbiol. 2002, 68(10), 5057–5063.

Declerck

Verelst

Duvivier

Van Damme

Ollevier

A detection method for Legionella spp in (cooling) water: fluorescent in situ hybridisation (FISH) on whole bacteria. Water Sci. Technol. 2003, 47(3), 143–146.

Delgado-Viscogliosi

Simonart

Parent

Marchand

Dobbelaere

Pierlot

Pierzo

Menard-Szczebara

Gaudard-Ferveur

Delabre

Delattre

J. M.

Rapid method for enumeration of viable Legionella pneumophila and other Legionella spp. in water. Appl. Environ. Microbiol. 2005, 71(7), 4086–4096.

Reidt

Chauhan

Müller

Molz

Lindner

Wolf

Friedberger

Reproducible filtration of bacteria with micromechanical filters. J. Rapid Methods Autom. Microbiol. 2008, 16(4), 337–350.

10.

Lund

Hellemann

A. L.

Vartdal

Rapid isolation of K88+ Escherichia coli by using immunomagnetic particles. J Clin Microbiol 1988, 26(12), 2572–2575.

11.

Stark

Reizenstein

Uhlen

Lundeberg

Immunomagnetic separation and solid-phase detection of Bordetella pertussis. J. Clin. Microbiol. 1996, 34(4), 778–784.

12.

Olsvik

Popovic

Skjerve

Cudjoe

K. S.

Hornes

Ugelstad

Uhlen

Magnetic separation techniques in diagnostic microbiology. Clin. Microbiol. Rev. 1994, 7(1), 43–54.

13.

Roberts

Hirst

Immunomagnetic separation and PCR for detection of Mycobacterium ulcerans. J. Clin. Microbiol. 1997, 35(10), 2709–2711.

14.

Inokuchi

H. N.

Suzuki

Kasagi

Shikazono

In Evaluation of Lamination Micro Mixer for Micro Immunomagnetic Cell Sorter, Microtechnologies in Medicine and Biology; 9–12 May 2006; pp 131–134 Okinawa, Japan.

15.

Mullane

N. R.

Murray

Drudy

Prentice

Whyte

Wall

P. G.

Parton

Fanning

Detection of Enterobacter sakazakii in dried infant milk formula by cationic-magnetic-bead capture. Appl. Environ. Microbiol. 2006, 72(9), 6325–6330.

16.

Hoffman

P. S.

Ripley

Weeratna

Cloning and nucleotide sequence of a gene (ompS) encoding the major outer membrane protein of Legionella pneumophila. J. Bacteriol. 1992, 174(3), 914–920.

17.

Hoffman

P. S.

Seyer

J. H.

Butler

C. A.

Molecular characterization of the 28- and 31-kilodalton subunits of the Legionella pneumophila major outer membrane protein. J. Bacteriol. 1992, 174(3), 908–913.

18.

Burnette

W. N.

“Western blotting”: Electrophoretic transfer of proteins from sodium dodecyl sulfate polyacrylamide gels to unmodified nitrocellulose and radiographic detection with antibody and radioiodinated protein A. Anal. Biochem. 1981, 112(2), 195–203.

Automated Immunomagnetic Processing and Separation of Legionella Pneumophila with Manual Detection by Sandwich ELISA and PCR Amplification of the ompS Gene