Accepted Manuscript Title: Fast label-free detection of Legionella spp. in biofilms by applying immunomagnetic beads and Raman spectroscopy Author: Dragana Kusi´c Petra R¨osch J¨urgen Popp PII: DOI: Reference:
S0723-2020(16)00016-3 http://dx.doi.org/doi:10.1016/j.syapm.2016.01.002 SYAPM 25746
To appear in: Received date: Revised date: Accepted date:
19-10-2015 18-1-2016 21-1-2016
Please cite this article as: D. Kusi´c, P. Rddotosch, J. Popp, Fast labelfree detection of Legionella spp. in biofilms by applying immunomagnetic beads and Raman spectroscopy, Systematic and Applied Microbiology (2016), http://dx.doi.org/10.1016/j.syapm.2016.01.002 This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
Fast label-free detection of Legionella spp. in biofilms by applying immunomagnetic beads and Raman spectroscopy Dragana Kusić1, Petra Rösch1,3* and Jürgen Popp1,2,3 Institut für Physikalische Chemie and Abbe Center of Photonics, Friedrich-Schiller-Unive-
ip t
1
rsität Jena, Helmholtzweg 4, D-07743 Jena, Germany
Leibniz Institute of Photonic Technology (IPHT), Albert-Einstein-Straße 9, D-07745 Jena,
cr
2
Germany
InfectoGnostics Forschungscampus Jena e.V., Zentrum für Angewandte Forschung,
us
3
an
Philosophenweg 7, D-07743 Jena, Germany Short title:
Ac ce p
Corresponding author
te
*
d
M
Identification of legionellae by IMS and Raman spectroscopy
Petra Rösch
E-mail: [email protected] Tel. 03641/9-48381; Fax 03641/9-48302
1 Page 1 of 25
Abstract Legionellae colonize biofilms, can form a biofilm by itself and multiply intracellularly within
ip t
the protozoa commonly found in water distribution systems. Approximately half of the known species are pathogenic and have been connected to severe multisystem Legionnaires’ disease. The detection methods for Legionella spp. in water samples are still based on
cr
cultivation, which is time consuming due to the slow growth of this bacterium. Here, we
us
developed a cultivation-independent, label-free and fast detection method for legionellae in a biofilm matrix based on the Raman spectroscopic analysis of isolated single cells via immunomagnetic separation (IMS). A database comprising the Raman spectra of single
an
bacterial cells captured and separated from the biofilms formed by each species was used to build the identification method based on a support vector machine (SVM) discriminative
M
classifier. The complete method allows the detection of Legionella spp. in 100 min. Crossreactivity of Legionella spp. specific immunomagnetic beads to the other studied genera was tested, where only small cell amounts of P. aeruginosa, K. pneumoniae and E. coli compared
d
to the initial number of cells were isolated by the immunobeads. Nevertheless, the Raman
te
spectra collected from isolated non-targeted bacteria were well-discriminated from the Raman spectra collected from isolated Legionella cells, whereby the Raman spectra of the
Ac ce p
independent dataset of Legionella strains were assigned with an accuracy of 98.6%. In addition, Raman spectroscopy was also used to differentiate between isolated Legionella species.
Keywords: Raman spectroscopy, Immunomagnetic separation, Biofilms, Legionella spp.
2 Page 2 of 25
Introduction Legionella is a bacterium that replicates intracellularly within a wide range of protozoa and persists in the biofilms of water storage and distribution systems. About half of the members
ip t
of the genus Legionella have been documented to be pathogenic, including L. pneumophila, which is the clinically most relevant species as it has been found to invade alveolar macrophages of the human lung causing the severe pneumonia of Legionnaires’ disease [6].
cr
Colonization of existing biofilms in water distribution systems by L. pneumophila appears to
us
be dependent on the presence of other microbial species. For example, due to its intracellular lifestyle within protozoa, it has been shown to persist in biofilms without replication in the absence of amoebae such as Hartmannella vermiformis [21]. Other studies
an
have suggested the extracellular growth of L. pneumophila within biofilms in the absence of amoebae [26]. In these studies, L. pneumophila has been considered in the context of pre-
M
grown biofilms. However, the ability of this bacterium to form a biofilm itself in rich complex medium has been investigated by Mampel and colleagues [18]. It was found that clinical as well as environmental isolates of L. pneumophila were capable of forming biofilms, where
d
the formation of a biofilm was found to be independent of the serogroup of L. pneumophila
te
and widespread among L. pneumophila strains. Biofilm formation by Legionella species in nutrient-poor environments such as tap water has been presented in a recent study [16].
Ac ce p
Compared to planktonic cells, sessile cells of L. pneumophila in biofilms accumulate polysaccharides as an extracellular polymer matrix as well as a greater amount of lipids. This finding has confirmed a distinct biofilm metabolism developed by L. pneumophila under low nutrient conditions. The ability of this pathogen to form biofilms in which the cells are surrounded by a self-produced matrix of diverse extracellular compounds may increase its survival in harsh environments and protect from harmful compounds. Molecular detection tools such as Raman microspectroscopy or secondary ion mass spectrometry have been recently combined with isotope-labelling techniques for the analysis of a bacterial community to allow the examination of the metabolic properties of microorganisms at a single cell level [33]. Furthermore, many different rapid detection methods based on the separation and subsequent identification of bacteria in environmental samples have been developed. For instance, immunomagnetic separation (IMS) based on the specific antigen-antibody reaction in combination with other detection techniques such as 3 Page 3 of 25
double fluorescent staining, flow cytometry [10] or real-time PCR amplification [35] have been used for the detection of L. pneumophila in water. These techniques in combination with IMS require the use of highly specific monoclonal antibodies to detect the unique pathogen with a chosen specificity. Otherwise, accompanying cross-reactions of antibodies
ip t
to similar epitopes of microbial species commonly found in environmental specimens limit the application of these methods for environmental work. In a recent study by Díaz-Flores and colleagues, a method based on IMS has been compared with the quantitative
cr
polymerase chain reaction (qPCR) as well as with traditional culture methods for the
us
detection of legionellae in complex water samples [5]. The reported results have shown that the IMS method was less influenced by matrix effects and is thus more suitable for the analysis of complex water samples compared to both culture and qPCR [5]. IMS in
an
combination with cultivation and flow cytometry was also used to investigate the occurrence of legionellae in both water and biofilm samples collected from groundwater wells [25]. It
M
has been shown that the IMS method could be used to partially reduce background microorganisms thus allowing a more efficient isolation of legionellae from environmental samples [25].
d
The potential of vibrational spectroscopy for bacterial characterization has been
te
documented [15; 27; 28; 29]. So far, Raman microspectroscopy has been used as a fast and label-free method to identify isolated or extracted microorganisms from samples such as
Ac ce p
milk [19], meat [20], powder samples [30], urine [11], ascitic fluid [13] or sputum samples [12] without any cultivation. Here, various isolation techniques such as filtration, buoyant density centrifugation, and enzymatic milk clearing have been used to isolate bacteria from complex matrices. Furthermore, due to the high specificity of antibodies in the recognition of their target cells, immunocapture has been applied to separate and subsequently identify bacteria via Raman spectroscopy [9; 22; 34]. Likewise, surface enhanced Raman scattering (SERS) combining specific capture of bacteria by antibodies was used for the detection of pathogens [14; 32]. Furthermore, for the identification of both Salmonella enterica and Staphylococcus aureus in complex food matrices such as spinach and peanut butter, Wang and colleagues proposed a MNPs@SiO2-based SERS platform with a reported detection limit of 103 CFU/mL [34]. Knauer and colleagues have used an amino-polyethylene glycol-coated surface treated with Legionella-specific and Salmonella-specific antibodies to immobilize both L. pneumophila and S. typhimurium and to subsequently identify them with SERS. 4 Page 4 of 25
Furthermore, Jing and colleagues have demonstrated the possibility of using SERS to differentiate among infectious and mildly infectious strains of L. pneumophila [7]. The majority of diverse bacteria contain an almost identical molecular composition and display very similar Raman spectra. However, small differences in the molecular composition
ip t
of various microbial cells result in a unique Raman signature allowing bacteria to be differentiated by their Raman spectra by applying chemometric approaches. It has been shown that the Raman spectra of numerous investigated legionellae can be differentiated
cr
via a multi class support vector machine [17]. In addition, Raman spectroscopy has been
us
applied to discriminate between sessile cells of diverse Legionella species grown as monospecies biofilms [16]. However, obtaining the Raman spectra of single bacterial cells in environmental samples that can have a complex heterogeneous composition can be
an
extremely challenging. For this, the application of Raman spectroscopy to identify distinctive bacterial species in environmental samples makes sense only if it is used in combination with
M
an efficient isolation technique compatible to Raman spectroscopy [23]. In this study, an evaluation of IMS for its compatibility with Raman microspectroscopy to isolate bacteria from a biofilm matrix and their subsequent identification on a single cell
d
level by applying multivariate analysis on the Raman spectra of single bacterial cells is
te
presented. Here, commercial polyclonal antibodies specific to Legionella species were used to remove the cells of particular species out of the biofilm formed by that particular
Ac ce p
microbial species before Raman microspectroscopic analysis. Raman spectra collected from isolated single cells were used to train the classification model which was later validated by an independent Raman dataset. In addition, the recovery efficiency of the IMS isolation of the biofilm form of Legionella species was evaluated. Finally, cross-reactivity of immunomagnetic beads to non-target genera grown in biofilms as well as the potential of Raman spectroscopy to distinguish between them and Legionella species was estimated. Materials and methods Bacterial strains and growth conditions Three subspecies of Legionella pneumophila representing serogroups 1, 4 and 5, (L. pneumophila ssp. pneumophila DSM 7513, L. pneumophila ssp. fraseri DSM 7514 and L. pneumophila ssp. pascullei DSM 7515), Legionella bozemanae DSM 16523, Legionella micdadei DSM 16640 Legionella feeleii DSM 17645, Legionella dumoffii DSM 17625 and 5 Page 5 of 25
Legionella anisa DSM 17627 obtained from the German Collection of Microorganisms and Cell Cultures (Leibniz Institute DSMZ, Germany) as well as Escherichia coli DSM 10806, Klebsiella pneumoniae ATCC 700603 and Pseudomonas aeruginosa PAO1 strain provided by the Institute of Medical Microbiology, Jena University Hospital were used. An overview of
ip t
the Legionella and other bacterial species used throughout this study is provided in Table 1. Differential buffered charcoal yeast extract (DBCYE) agar (Sigma Aldrich, D3560-500G) containing ferric pyrophosphate, L-cysteine hydrochloride and α-ketoglutarate was used to
cr
maintain bacterial cultures. All species were grown at 37 °C in a humid atmosphere (≈ 85%
us
humidity) with 5% CO2 for 24 hours (E. coli, K. pneumoniae and P. aeruginosa) or 3 days (Legionella spp.) prior the biofilm development in filter sterile tap water so that the physiological state of the replicates was each time the same. For long-term storage, bacterial
an
cultures were kept at -80 °C.
M
Formation of the biofilm samples
Bacterial colonies pre-grown on DBCYE agar were rinsed off with sterile distilled water, washed via centrifugation for 3 minutes at 5000 rpm at room temperature, and the bacterial
d
pellet was then resuspended in sterile-filtered tap water (Pall Life Sciences, Acrodisc PF
te
syringe filters with 0.2 µm Supor membrane) so that the final OD was approximately 1.5. Four mL of bacterial solution from each species was divided into 8 vessels (Lab-Tek™
Ac ce p
chambered cover glasses, 734-2060) and incubated at 37 °C for 3 days in a humid atmosphere (≈ 85% humidity) with 5% CO2 under static conditions to allow the biofilms to be formed at the borosilicate glass surface. The supernatant was afterwards carefully aspirated, while the remaining biofilms were carefully rinsed three times with distilled water. Washed biofilms were mixed thoroughly with sterile-filtered tap water by pipetting and frequently scratching the bottom of the vessel via an inoculation loop. Subsequently, the contents of each vessel were transferred to a new tube and resuspended in sterile-filtered tap water to give a total volume of 1 mL an a final OD of ~0.5. Four independently cultivated batches of biofilms for each species grown in sterile-filtered tap water were prepared as described. Coating of the magnetic beads Dynabeads® M-28 Tosylactivated (Invitrogen by Life Technologies AS, Oslo; 14203) were used as magnetic beads. A stock suspension of magnetic beads was washed with sterile 6 Page 6 of 25
0.1 M sodium phosphate buffer before coating with polyclonal rabbit anti-Legionella spp. IgG (GenWay, 18-783 78491) according to the supplier’s protocol. Washed and resuspended magnetic beads were transferred to a new tube, placed in a magnet (Roti-Mag TubeSeparator, Carl Roth, 2807.1) for 1 min, and then the supernatant was removed. The
ip t
magnetic beads and Legionella spp. specific antibodies were then resuspended so that the final concentration was 20 μg of antibodies per mg of magnetic beads. Then, both 0.1 M sodium phosphate buffer and 3 M ammonium sulphate in 0.1 M sodium phosphate buffer
cr
were added to the solution and incubated on a roller at 37 °C for 18 hours. A slightly
us
hydrophobic surface of magnetic beads supports binding of the more hydrophobic crystallisable fragment (Fc), or "tail" of the antibody, while the antigen-binding (Fab) region remains accessible to its target. Here, direct covalent binding of antibodies via primary
an
amino- or sulfhydryl groups is provided via the surface tosyl groups. Subsequently, the tube was placed on a magnet for 2 min and the supernatant was removed. In addition, 1 mL of
M
phosphate buffered saline (PBS) with 0.5% (w/v) bovine serum albumin (BSA) (Invitrogen, AM2616) was added to the tube. The tube was then incubated at 37 °C for 1 hour on a roller. Afterwards, the supernatant was removed by placing the tube on a magnet for 2 min. Finally,
d
immunomagnetic beads (MbAb) were washed twice with PBS with 0.1% (w/v) BSA,
te
resuspended and diluted in PBS with 0.1% (w/v) BSA to achieve a final desired MbAb concentration of 20 mg/mL. The MbAbs were stored in a suspension of phosphate buffered
Ac ce p
saline (PBS) pH 7.4 with 0.1% BSA at 4 °C.
Immunomagnetic separation of bacteria (IMS) Immunomagnetic cell separation was performed using a magnet separator (Roti-Mag TubeSeparator, Carl Roth, 2807.1). Fifty µL of the MbAb solution containing 1 mg of antibodycoupled beads were added to 1 mL of biofilm suspension and incubated for 60 min under permanent agitation at room temperature to capture the target bacteria. Subsequently, MbAb-bacterial complexes were washed with sterile distilled water by placing the sample tube on the magnet for 2 min, and then the supernatant was pipetted. This washing step was repeated three times. Next, to remove the captured bacteria from the surface beads, the separated magnetic beads with target bacteria were resuspended in 30 µL of pepsin solution (15 μg/mL, pH 5.5 [31]) and incubated for 15 min as the sample tube was vortexed vigorously every 5 min for 20 sec. Distilled water was then added to give a total volume of 7 Page 7 of 25
100 µL. Detached bacteria were removed from the immunomagnetic beads by placing the sample tube on the magnet for 2 min. Microbial cell quantification
ip t
The number of colony forming units (CFU) was determined by the enumeration of the colonies formed on buffered charcoal yeast extract (BCYE) agar (Sigma-Aldrich, 86558500G F) enhanced with Legionella growth supplement (Sigma-Aldrich 42981-1KT). To
cr
calculate the initial concentration of bacteria in the biofilm suspensions before the IMS
us
procedure (OD~0.5), 10-fold serial dilutions were prepared in sterile distilled water and 100 µL aliquots of each dilution were spread on BCYE agar plates in triplicate. The plates were incubated at 37 °C in a humid atmosphere (≈85% humidity) with 5% CO2 for 5-7 days
an
(L. pneumophila ssp. pneumophila) or 24 hours (E. coli, K. pneumoniae and P. aeruginosa) and the number of colony counts was determined. The three non-Legionella species were
M
used to test the cross-reactivity of immunobeads specific to Legionella species. To determine the concentration of isolated biofilm cells via IMS, 10-fold serial dilutions of the final 100 µL of IMS isolated bacteria were prepared, and 100 µL aliquots of each dilution were spread on
d
BCYE agar plates in triplicate. In addition, the number of bacterial residues of L. pneumophila
te
ssp. pneumophila pulled out by the magnetic particles after IMS was also determined. A 10-fold serial dilution of the remaining “MbAb-bacterial complexes” on the side of the tube
Ac ce p
were prepared; 100 µL aliquots of each dilution was spread on BCYE agar plates and the number of CFUs was established. Spectroscopic instrumentation
A Raman microscope (Bio Particle Explorer; Rap.ID Particle Systems GmbH, Berlin, Germany) was used to record the Raman spectra of the cells isolated via IMS out of the biofilm matrix. The incident laser excitation wavelength was 532 nm from a solid-state frequency-doubled Nd:YAG laser source. An Olympus MPLFLN-BD 100 X microscope objective was used to focus the laser beam with a laser power of about 3.5 mW incident on the single bacterial cell. Afterwards the back scattered light was diffracted with a single-stage monochromator (HE 532; Horiba JobinYvon) equipped with a grating of 920 lines mm-1, while a thermoelectrically cooled CCD camera (DV401 BV; Andor Technology) was used to detect the Raman scattering. Spatial resolution was about 1 μm (spot diameter of
S0723-2020(16)00016-3 http://dx.doi.org/doi:10.1016/j.syapm.2016.01.002 SYAPM 25746
To appear in: Received date: Revised date: Accepted date:
19-10-2015 18-1-2016 21-1-2016
Please cite this article as: D. Kusi´c, P. Rddotosch, J. Popp, Fast labelfree detection of Legionella spp. in biofilms by applying immunomagnetic beads and Raman spectroscopy, Systematic and Applied Microbiology (2016), http://dx.doi.org/10.1016/j.syapm.2016.01.002 This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
Fast label-free detection of Legionella spp. in biofilms by applying immunomagnetic beads and Raman spectroscopy Dragana Kusić1, Petra Rösch1,3* and Jürgen Popp1,2,3 Institut für Physikalische Chemie and Abbe Center of Photonics, Friedrich-Schiller-Unive-
ip t
1
rsität Jena, Helmholtzweg 4, D-07743 Jena, Germany
Leibniz Institute of Photonic Technology (IPHT), Albert-Einstein-Straße 9, D-07745 Jena,
cr
2
Germany
InfectoGnostics Forschungscampus Jena e.V., Zentrum für Angewandte Forschung,
us
3
an
Philosophenweg 7, D-07743 Jena, Germany Short title:
Ac ce p
Corresponding author
te
*
d
M
Identification of legionellae by IMS and Raman spectroscopy
Petra Rösch
E-mail: [email protected] Tel. 03641/9-48381; Fax 03641/9-48302
1 Page 1 of 25
Abstract Legionellae colonize biofilms, can form a biofilm by itself and multiply intracellularly within
ip t
the protozoa commonly found in water distribution systems. Approximately half of the known species are pathogenic and have been connected to severe multisystem Legionnaires’ disease. The detection methods for Legionella spp. in water samples are still based on
cr
cultivation, which is time consuming due to the slow growth of this bacterium. Here, we
us
developed a cultivation-independent, label-free and fast detection method for legionellae in a biofilm matrix based on the Raman spectroscopic analysis of isolated single cells via immunomagnetic separation (IMS). A database comprising the Raman spectra of single
an
bacterial cells captured and separated from the biofilms formed by each species was used to build the identification method based on a support vector machine (SVM) discriminative
M
classifier. The complete method allows the detection of Legionella spp. in 100 min. Crossreactivity of Legionella spp. specific immunomagnetic beads to the other studied genera was tested, where only small cell amounts of P. aeruginosa, K. pneumoniae and E. coli compared
d
to the initial number of cells were isolated by the immunobeads. Nevertheless, the Raman
te
spectra collected from isolated non-targeted bacteria were well-discriminated from the Raman spectra collected from isolated Legionella cells, whereby the Raman spectra of the
Ac ce p
independent dataset of Legionella strains were assigned with an accuracy of 98.6%. In addition, Raman spectroscopy was also used to differentiate between isolated Legionella species.
Keywords: Raman spectroscopy, Immunomagnetic separation, Biofilms, Legionella spp.
2 Page 2 of 25
Introduction Legionella is a bacterium that replicates intracellularly within a wide range of protozoa and persists in the biofilms of water storage and distribution systems. About half of the members
ip t
of the genus Legionella have been documented to be pathogenic, including L. pneumophila, which is the clinically most relevant species as it has been found to invade alveolar macrophages of the human lung causing the severe pneumonia of Legionnaires’ disease [6].
cr
Colonization of existing biofilms in water distribution systems by L. pneumophila appears to
us
be dependent on the presence of other microbial species. For example, due to its intracellular lifestyle within protozoa, it has been shown to persist in biofilms without replication in the absence of amoebae such as Hartmannella vermiformis [21]. Other studies
an
have suggested the extracellular growth of L. pneumophila within biofilms in the absence of amoebae [26]. In these studies, L. pneumophila has been considered in the context of pre-
M
grown biofilms. However, the ability of this bacterium to form a biofilm itself in rich complex medium has been investigated by Mampel and colleagues [18]. It was found that clinical as well as environmental isolates of L. pneumophila were capable of forming biofilms, where
d
the formation of a biofilm was found to be independent of the serogroup of L. pneumophila
te
and widespread among L. pneumophila strains. Biofilm formation by Legionella species in nutrient-poor environments such as tap water has been presented in a recent study [16].
Ac ce p
Compared to planktonic cells, sessile cells of L. pneumophila in biofilms accumulate polysaccharides as an extracellular polymer matrix as well as a greater amount of lipids. This finding has confirmed a distinct biofilm metabolism developed by L. pneumophila under low nutrient conditions. The ability of this pathogen to form biofilms in which the cells are surrounded by a self-produced matrix of diverse extracellular compounds may increase its survival in harsh environments and protect from harmful compounds. Molecular detection tools such as Raman microspectroscopy or secondary ion mass spectrometry have been recently combined with isotope-labelling techniques for the analysis of a bacterial community to allow the examination of the metabolic properties of microorganisms at a single cell level [33]. Furthermore, many different rapid detection methods based on the separation and subsequent identification of bacteria in environmental samples have been developed. For instance, immunomagnetic separation (IMS) based on the specific antigen-antibody reaction in combination with other detection techniques such as 3 Page 3 of 25
double fluorescent staining, flow cytometry [10] or real-time PCR amplification [35] have been used for the detection of L. pneumophila in water. These techniques in combination with IMS require the use of highly specific monoclonal antibodies to detect the unique pathogen with a chosen specificity. Otherwise, accompanying cross-reactions of antibodies
ip t
to similar epitopes of microbial species commonly found in environmental specimens limit the application of these methods for environmental work. In a recent study by Díaz-Flores and colleagues, a method based on IMS has been compared with the quantitative
cr
polymerase chain reaction (qPCR) as well as with traditional culture methods for the
us
detection of legionellae in complex water samples [5]. The reported results have shown that the IMS method was less influenced by matrix effects and is thus more suitable for the analysis of complex water samples compared to both culture and qPCR [5]. IMS in
an
combination with cultivation and flow cytometry was also used to investigate the occurrence of legionellae in both water and biofilm samples collected from groundwater wells [25]. It
M
has been shown that the IMS method could be used to partially reduce background microorganisms thus allowing a more efficient isolation of legionellae from environmental samples [25].
d
The potential of vibrational spectroscopy for bacterial characterization has been
te
documented [15; 27; 28; 29]. So far, Raman microspectroscopy has been used as a fast and label-free method to identify isolated or extracted microorganisms from samples such as
Ac ce p
milk [19], meat [20], powder samples [30], urine [11], ascitic fluid [13] or sputum samples [12] without any cultivation. Here, various isolation techniques such as filtration, buoyant density centrifugation, and enzymatic milk clearing have been used to isolate bacteria from complex matrices. Furthermore, due to the high specificity of antibodies in the recognition of their target cells, immunocapture has been applied to separate and subsequently identify bacteria via Raman spectroscopy [9; 22; 34]. Likewise, surface enhanced Raman scattering (SERS) combining specific capture of bacteria by antibodies was used for the detection of pathogens [14; 32]. Furthermore, for the identification of both Salmonella enterica and Staphylococcus aureus in complex food matrices such as spinach and peanut butter, Wang and colleagues proposed a MNPs@SiO2-based SERS platform with a reported detection limit of 103 CFU/mL [34]. Knauer and colleagues have used an amino-polyethylene glycol-coated surface treated with Legionella-specific and Salmonella-specific antibodies to immobilize both L. pneumophila and S. typhimurium and to subsequently identify them with SERS. 4 Page 4 of 25
Furthermore, Jing and colleagues have demonstrated the possibility of using SERS to differentiate among infectious and mildly infectious strains of L. pneumophila [7]. The majority of diverse bacteria contain an almost identical molecular composition and display very similar Raman spectra. However, small differences in the molecular composition
ip t
of various microbial cells result in a unique Raman signature allowing bacteria to be differentiated by their Raman spectra by applying chemometric approaches. It has been shown that the Raman spectra of numerous investigated legionellae can be differentiated
cr
via a multi class support vector machine [17]. In addition, Raman spectroscopy has been
us
applied to discriminate between sessile cells of diverse Legionella species grown as monospecies biofilms [16]. However, obtaining the Raman spectra of single bacterial cells in environmental samples that can have a complex heterogeneous composition can be
an
extremely challenging. For this, the application of Raman spectroscopy to identify distinctive bacterial species in environmental samples makes sense only if it is used in combination with
M
an efficient isolation technique compatible to Raman spectroscopy [23]. In this study, an evaluation of IMS for its compatibility with Raman microspectroscopy to isolate bacteria from a biofilm matrix and their subsequent identification on a single cell
d
level by applying multivariate analysis on the Raman spectra of single bacterial cells is
te
presented. Here, commercial polyclonal antibodies specific to Legionella species were used to remove the cells of particular species out of the biofilm formed by that particular
Ac ce p
microbial species before Raman microspectroscopic analysis. Raman spectra collected from isolated single cells were used to train the classification model which was later validated by an independent Raman dataset. In addition, the recovery efficiency of the IMS isolation of the biofilm form of Legionella species was evaluated. Finally, cross-reactivity of immunomagnetic beads to non-target genera grown in biofilms as well as the potential of Raman spectroscopy to distinguish between them and Legionella species was estimated. Materials and methods Bacterial strains and growth conditions Three subspecies of Legionella pneumophila representing serogroups 1, 4 and 5, (L. pneumophila ssp. pneumophila DSM 7513, L. pneumophila ssp. fraseri DSM 7514 and L. pneumophila ssp. pascullei DSM 7515), Legionella bozemanae DSM 16523, Legionella micdadei DSM 16640 Legionella feeleii DSM 17645, Legionella dumoffii DSM 17625 and 5 Page 5 of 25
Legionella anisa DSM 17627 obtained from the German Collection of Microorganisms and Cell Cultures (Leibniz Institute DSMZ, Germany) as well as Escherichia coli DSM 10806, Klebsiella pneumoniae ATCC 700603 and Pseudomonas aeruginosa PAO1 strain provided by the Institute of Medical Microbiology, Jena University Hospital were used. An overview of
ip t
the Legionella and other bacterial species used throughout this study is provided in Table 1. Differential buffered charcoal yeast extract (DBCYE) agar (Sigma Aldrich, D3560-500G) containing ferric pyrophosphate, L-cysteine hydrochloride and α-ketoglutarate was used to
cr
maintain bacterial cultures. All species were grown at 37 °C in a humid atmosphere (≈ 85%
us
humidity) with 5% CO2 for 24 hours (E. coli, K. pneumoniae and P. aeruginosa) or 3 days (Legionella spp.) prior the biofilm development in filter sterile tap water so that the physiological state of the replicates was each time the same. For long-term storage, bacterial
an
cultures were kept at -80 °C.
M
Formation of the biofilm samples
Bacterial colonies pre-grown on DBCYE agar were rinsed off with sterile distilled water, washed via centrifugation for 3 minutes at 5000 rpm at room temperature, and the bacterial
d
pellet was then resuspended in sterile-filtered tap water (Pall Life Sciences, Acrodisc PF
te
syringe filters with 0.2 µm Supor membrane) so that the final OD was approximately 1.5. Four mL of bacterial solution from each species was divided into 8 vessels (Lab-Tek™
Ac ce p
chambered cover glasses, 734-2060) and incubated at 37 °C for 3 days in a humid atmosphere (≈ 85% humidity) with 5% CO2 under static conditions to allow the biofilms to be formed at the borosilicate glass surface. The supernatant was afterwards carefully aspirated, while the remaining biofilms were carefully rinsed three times with distilled water. Washed biofilms were mixed thoroughly with sterile-filtered tap water by pipetting and frequently scratching the bottom of the vessel via an inoculation loop. Subsequently, the contents of each vessel were transferred to a new tube and resuspended in sterile-filtered tap water to give a total volume of 1 mL an a final OD of ~0.5. Four independently cultivated batches of biofilms for each species grown in sterile-filtered tap water were prepared as described. Coating of the magnetic beads Dynabeads® M-28 Tosylactivated (Invitrogen by Life Technologies AS, Oslo; 14203) were used as magnetic beads. A stock suspension of magnetic beads was washed with sterile 6 Page 6 of 25
0.1 M sodium phosphate buffer before coating with polyclonal rabbit anti-Legionella spp. IgG (GenWay, 18-783 78491) according to the supplier’s protocol. Washed and resuspended magnetic beads were transferred to a new tube, placed in a magnet (Roti-Mag TubeSeparator, Carl Roth, 2807.1) for 1 min, and then the supernatant was removed. The
ip t
magnetic beads and Legionella spp. specific antibodies were then resuspended so that the final concentration was 20 μg of antibodies per mg of magnetic beads. Then, both 0.1 M sodium phosphate buffer and 3 M ammonium sulphate in 0.1 M sodium phosphate buffer
cr
were added to the solution and incubated on a roller at 37 °C for 18 hours. A slightly
us
hydrophobic surface of magnetic beads supports binding of the more hydrophobic crystallisable fragment (Fc), or "tail" of the antibody, while the antigen-binding (Fab) region remains accessible to its target. Here, direct covalent binding of antibodies via primary
an
amino- or sulfhydryl groups is provided via the surface tosyl groups. Subsequently, the tube was placed on a magnet for 2 min and the supernatant was removed. In addition, 1 mL of
M
phosphate buffered saline (PBS) with 0.5% (w/v) bovine serum albumin (BSA) (Invitrogen, AM2616) was added to the tube. The tube was then incubated at 37 °C for 1 hour on a roller. Afterwards, the supernatant was removed by placing the tube on a magnet for 2 min. Finally,
d
immunomagnetic beads (MbAb) were washed twice with PBS with 0.1% (w/v) BSA,
te
resuspended and diluted in PBS with 0.1% (w/v) BSA to achieve a final desired MbAb concentration of 20 mg/mL. The MbAbs were stored in a suspension of phosphate buffered
Ac ce p
saline (PBS) pH 7.4 with 0.1% BSA at 4 °C.
Immunomagnetic separation of bacteria (IMS) Immunomagnetic cell separation was performed using a magnet separator (Roti-Mag TubeSeparator, Carl Roth, 2807.1). Fifty µL of the MbAb solution containing 1 mg of antibodycoupled beads were added to 1 mL of biofilm suspension and incubated for 60 min under permanent agitation at room temperature to capture the target bacteria. Subsequently, MbAb-bacterial complexes were washed with sterile distilled water by placing the sample tube on the magnet for 2 min, and then the supernatant was pipetted. This washing step was repeated three times. Next, to remove the captured bacteria from the surface beads, the separated magnetic beads with target bacteria were resuspended in 30 µL of pepsin solution (15 μg/mL, pH 5.5 [31]) and incubated for 15 min as the sample tube was vortexed vigorously every 5 min for 20 sec. Distilled water was then added to give a total volume of 7 Page 7 of 25
100 µL. Detached bacteria were removed from the immunomagnetic beads by placing the sample tube on the magnet for 2 min. Microbial cell quantification
ip t
The number of colony forming units (CFU) was determined by the enumeration of the colonies formed on buffered charcoal yeast extract (BCYE) agar (Sigma-Aldrich, 86558500G F) enhanced with Legionella growth supplement (Sigma-Aldrich 42981-1KT). To
cr
calculate the initial concentration of bacteria in the biofilm suspensions before the IMS
us
procedure (OD~0.5), 10-fold serial dilutions were prepared in sterile distilled water and 100 µL aliquots of each dilution were spread on BCYE agar plates in triplicate. The plates were incubated at 37 °C in a humid atmosphere (≈85% humidity) with 5% CO2 for 5-7 days
an
(L. pneumophila ssp. pneumophila) or 24 hours (E. coli, K. pneumoniae and P. aeruginosa) and the number of colony counts was determined. The three non-Legionella species were
M
used to test the cross-reactivity of immunobeads specific to Legionella species. To determine the concentration of isolated biofilm cells via IMS, 10-fold serial dilutions of the final 100 µL of IMS isolated bacteria were prepared, and 100 µL aliquots of each dilution were spread on
d
BCYE agar plates in triplicate. In addition, the number of bacterial residues of L. pneumophila
te
ssp. pneumophila pulled out by the magnetic particles after IMS was also determined. A 10-fold serial dilution of the remaining “MbAb-bacterial complexes” on the side of the tube
Ac ce p
were prepared; 100 µL aliquots of each dilution was spread on BCYE agar plates and the number of CFUs was established. Spectroscopic instrumentation
A Raman microscope (Bio Particle Explorer; Rap.ID Particle Systems GmbH, Berlin, Germany) was used to record the Raman spectra of the cells isolated via IMS out of the biofilm matrix. The incident laser excitation wavelength was 532 nm from a solid-state frequency-doubled Nd:YAG laser source. An Olympus MPLFLN-BD 100 X microscope objective was used to focus the laser beam with a laser power of about 3.5 mW incident on the single bacterial cell. Afterwards the back scattered light was diffracted with a single-stage monochromator (HE 532; Horiba JobinYvon) equipped with a grating of 920 lines mm-1, while a thermoelectrically cooled CCD camera (DV401 BV; Andor Technology) was used to detect the Raman scattering. Spatial resolution was about 1 μm (spot diameter of
