Detection of li O157:H7by a Hybrid Microfluidic SPR and Molecular Imaging Cytometry Device
Michael D.Zordan,1,2,3Meggie M.G.Grafton,1,2,3Ghanashyam Acharya,1,3Lisa M.Reece,2,3,4Christy L.Cooper,2,3,4Arthur I.Aronson,5Kinam Park,1,6James F.Leary 1,2,3,4*
Abstract
Current methods to screen for bacterial contamination involve using costly reagents such as antibodies or PCR reagents or time-costly growth in cultures.There is need for portable,real-time,multiplex pathogen detection technology that can predict the safety of food.Surface plasmon resonance (SPR)imaging is a nsitive,label-free method that can detect the binding of an analyte to a surface by the changes in refractive index that occur upon binding.We have designed a hybrid microfluidic biochip to perform multi-plexed detection of single-celled pathogens using a combination of SPR and fluores-cence imaging.The device consists of an array of gold spots,each functionalized with a capture biomolecule targeting a specific pathogen.This bionsor array is enclod by a polydimethylsiloxane
microfluidic flow chamber that delivers a magnetically concen-trated sample to be tested.The sample is imaged by SPR on the bottom of the biochip and epi-fluorescence on the top.The prototype instrument was successfully able to image li O157:H7bacteria by SPR and fluorescence imaging.The efficiency of capture of the bacteria by the magnetic particles was determined using spectrophotometric ferric oxide absorbance measurements.The binding of li to each spot was quantified by measuring the percent of the gold spot area upon which the bacteria was bound and analyzed using NIH ImageJ software.This hybrid imaging approach of li detection coupled with an estimate of relative infectivity is shown to be a working example of a testing device for potential foodborne pathogens.'2008International Society for Advancement of Cytometry Key terms
foodborne pathogens;microbes;detection;microfluidic;cytometry;imaging;surface plasmon li O157:H7
T HE incread incidence of fatal pathogen-contaminated food supplies places a new
emphasis on the rapid detection and quantification of the foodborne pathogens.This issue has been further compounded by the fact that many high-risk foodborne patho-gens are easily transmitted thro
ugh food supplies,thus constituting a major public health problem.Accurate and rapid identification of pathogens in food is to facilitate timely and appropriate actions in the event of a contamination.Conventional patho-gen detection methods involve enriching the sample and performing various media-bad metabolic tests (1).The detection methods are elaborate and typically require 2–7days to obtain results.Detection using magnetic beads coated with pathogen-specific antibodies or enzyme-linked immunosorbent assays still require veral hours for completing the tests (2,3).The oligonucleotide array method bad on amplification and hybridization of DNA fragments of pathogenic bacteria also takes more than veral hours (4).Some techniques also require rather expensive special instruments,such as flow cytometry (5)and real-time PCR (6).Hence,a rapid,label-free,and easy-to-u bionsor capable of detecting toxigenic bacteria in a few minutes is of immediate
importance.
1
Weldon School of Biomedical
还有的拼音
Engineering,Purdue University,West Lafayette,Indiana
2
Bindley Bioscience Center,Purdue University,West Lafayette,Indiana 3六味地黄丸副作用
Birck Nanotechnology Center,Purdue University,West Lafayette,Indiana 4
Department of Basic Medical Science,School of Veterinary Medicine,Purdue University,West Lafayette,Indiana 5
Department of Biological Sciences,Purdue University,West Lafayette,Indiana
6
Department of Industrial and Physical Pharmacy,Purdue University,West Lafayette,Indiana
Received 15August 2008;Revision Received 31October 2008;Accepted 5November 2008
Additional Supporting Information may be found in the online version of this article.Grant sponsor:USDA (through the Center for Food Safety Engineering at Purdue University);Grant number:
102145*Correspondence to:James F.Leary,Birck Nanotechnology Center,Purdue University,Room 2021,1205W.State Street,West Lafayette,IN 47907,USA Email:jfleary@purdue.edu
Published online 5December 2008in Wiley InterScience (www.)
DOI:10.1002/cyto.a.20692©2008International Society for Advancement of Cytometry
Original Article
Cytometry Part A 75A:155À162,2009
楚州城
The ability to screen for multiple pathogens in one assay is of great importance.It is possible to accomplish this using an array of immobilized capture biomolecules.In this tup, the pathogen will bind to and be captured at a specific spatial position in the array,corresponding to the appropriate capture molecule.There are many types of biomolecules that can be ud to specifically bind pathogens,such as antibodies,short peptides,and aptamers.Further,the capture biomolecules can be spotted on a surface in a spatially parated array that will allow for the simultaneous multiplex detection of different pathogens in the same sample.Bionsor arrays such as the have been demonstrated using peptides(7),aptamers(8),and antibodies(9).In addition,surface plasmon resonan
ce(SPR) imaging has been ud previously to measure simultaneous binding events on microarrays(9–12).Additionally,surface-bad detection requires lower quantities of the costly capture biomolecules than bulk methods.
B ASI
C P RINCIPLES OF S URFACE P LASMON
R ESONANCE I MAGING
SPR imaging can detect the prence of molecules or cells or pathogens bound to the bionsor surface by measuring the changes in the local refractive indices(13,14).SPR imaging involves the measurement of the intensity of light reflected at a dielectric covered by a ,gold)layer of$50nm thickness.The charge-density propagating along the interface of the thin metal layer and the dielectric is compod of sur-face plasmons.The surface plasmons are excited by an eva-nescent field typically generated by total internal reflection via a prism coupler.The wave vector of the surface plasmons is dependent upon the properties of the prism,the gold layer, and the surrounding dielectric medium(glass slide)(15). Under appropriate conditions,the free electrons come in reso-nance with the incident light and a surface plasmon is gener-ated.At this resonance co
ndition,the reflection decreas sharply to a minimum becau incident photons will induce surface plasmons instead of being reflected.Changes in dielec-tric ,thickness or refractive index,of the sur-rounding medium lead to changes in the wave vector and con-quently there is a shift of plasmon resonance minimum of the reflected light(15).
The binding of biomolecules or bacteria to the surface is nsitively detected,as the plasmon resonance is extremely nsitive to surface dielectric properties and the fact that reso-nance occurs only in a small range of conditions including re-fractive index,wavelength,and angle of incidence(10,15,16). Resonance angle measurements have been ud for chemical and biochemical nsing(16).Only p-polarized light in plane of incidence with the electric field vector oscillating in the plane of the metal film is able to couple to the plasmon mode. The s-polarized light,with its electric field vector oriented per-pendicular to the metal film,does not excite plasmons but is reflected by the metal surface and can be ud as a reference signal to improve the nsitivity(17,18).Thus,in SPR ima-ging,the reflectivity change resulting from biomolecular and cellular binding on the bionsor surface is measured before and after analyte binding.M OLECULAR I MAGING TO P ROVIDE A DDITIONAL
‘‘P ATHOGEN S TATE’’I NFORMATION
In some instances,it is also important to not only detect the prence of pathogens by SPR but also the state of the pathogen.For example,the prence of pathogenic bacteria is a uful information,but that information would be of much greater u if we also obtain a measure of the‘‘pathogen state.’’For this reason,simultaneous fluorescence molecular imaging contained in a hybrid SPR/molecular imaging device is signifi-cantly more powerful.The pathogen state information can extend from simple viability measures to fluorescent molecu-lar probes of metabolic or functional capabilities of the detected pathogen(19–22).This‘‘functional’’information may prove more valuable when combined with structural in-formation.This pathogen state information usually correlates more cloly with conventional microbiological assays bad on growth in culture,which in turn relates more cloly to in-fectious potential and relative health risks.However,if enough dead bacteria are prent,endotoxins such as lipopolysaccha-ride can still po a significant risk factor to human health. For this reason,it is important to try to obtain not only a rela-tive number of live/dead bacteria but also a measure of total number.
It is also easier to quantify the absolute number of patho-gens in different pathogen states using a multicolor fluores-cence approach.The combination of SPR and molecular ima-ging also provides powerful proof that each measurement is a true-positive,rather than fal-positive leading to greater
detection accuracy.This imaging need not optically resolve individual pathogens.Relative numbers can be determined by total fluorescence measured,which gives an absolute number of pathogens within the variability of fluorescence staining of individual pathogens.
This device integrates an SPR imaging system with an epi-fluorescence molecular imaging system,both to determine what fraction of pathogenic bacteria are live or dead and to confirm the SPR results(23).A microfluidic chip enclos a bionsor array that will capture the sample.The array is func-tionalized with ligands specific to the pathogens that are to be detected.
I MPORTANCE OF‘‘P RECONCENTRATING’’THE P ATHOGENS FOR S UBSEQUENT A NALYSIS IN THE D EVICExls表格
Since microfluidic devices by definition can only sample small amounts of fluid,it is important and necessary to pre-concentrate all possible pathogens prent in large volumes of fluid prior to microfluidic analysis.U of specific antibo-dies or other capture molecules,such as peptides or apta-mers,works well but requires specific reagents and creation of a multiplexed magnetic capture molecule system.The nanoparticle-labeled pathogens can be captured and held against a surface while excess fluid is discarded.This trans-lates to very large improvements in sampling statistics. Th
ere remains a need to develop more general capture methods that can easily lect a number of different patho-gens simultaneously.
ORIGINAL ARTICLE
156Hybrid Imaging Pathogen Detector
M ATERIALS AND M ETHODS
Overall Design of SPR Device
This hybrid,multicomponent device contains (1)a front-end magnetic concentrator to capture magnetic micro-or nanoparticle-labeled microbes and increa their concentra-tion into a smaller volume suitable for a microfluidic flow/imaging device,(2)a surface plasmon imaging subsystem to detect captured microbes on a patterned grid of gold contact spots,(3)a molecular imaging epi-fluorescence subsystem to determine viability and functional status of the captured microbes,and (4)small imaging cameras to capture SPR and molecular imaging data coupled to a portable computing de-vice (initially a laptop computer,but eventually a much smal-ler PDA-type device).This computing device can contain automated image analysis and other software to do completely autom
ated analysis for pathogen detection.A schematic of the overall conceptual design of this portable foodborne pathogen detection system is shown in Figure 1.
Bacterial Strains,Growth,and Staining
Two strains li ,li O157:H7(Castel-lani and Chalmers strain,ATCC,Manassas,VA),and the li DH5-a (provided by Arthur Aronson,PhD,Department of Biological Sciences,Purdue University,West Lafayette,IN)were ud for proof-of-concept experiments.The bacteria were streaked onto an LB (Luria-Bertani)plate and incubated at 378C overnight.Single isolated colonies were aptically harvested from the LB plate and allowed to grow in culture tubes containing 10mL of LB broth,in an agitating in-cubator t to 200rpm overnight at 378C.Bacterial cultures were harvested when they reached an OD of 0.66at 670nm for the live DH5-a and 0.467for the live O157:H7bacteria.
To make sure that our growth conditions for our bacterial strains were optimized and to ensure that we had viable bacte-ria for our experiments,we performed a simple fluorescence method for live/dead bacteria determinations.Bac Light TM Bacterial Viability Kits (Invitrogen,Carlsbad,CA)provides a nsitive,single-step,fluorescence-bad assay for bacterial cell
viability employing two nucleic acid stains—the green-fluores-cent SYTO 19stain and the red-fluorescent propidium iodide (PI)stain (19–21,24).When ud alone,the SYTO 19stain labels both live and dead bacteria.In contrast,PI penetrates only bacteria with damaged membranes,reducing SYTO 19fluorescence when both dyes are prent.This assay is not a universal viability indicator for all species of bacteria,becau it has been found that in multiple bacterial species will uptake PI during the exponential growth stage resulting in a fal dead signal (24).Therefore,the efficacy of Bac Light TM kit must be verified for the desired pathogen strain.When the assay is applicable for the pathogen species,it is desirable to u becau it can be completed in minutes and does not require wash steps.This assay was demonstrated to success-fully discriminate between live and li O157:H7cells (e Supp.Info.Fig.S1).For this assay and all subquent binding li O157:H7bacteria were incubated with 7.16l M SYTO-9and 17.14l M PI for 10min at room temperature per manufacturer’s protocol.
Magnetic Preconcentration of Bacterial Sample
Magnetic preconcentration was accomplished using superparamagnetic carboxyl-functionalized Biomag 11.5l m iron oxide beads (Bang’s Labs,Fishers,IN)coupled with anti-bodies specific to a membrane antigen li O157:H7and then incubated with li O157:H7(shown in Supp.
Info.
Figure 1.Multicomponent sch-ematic of the overall pathogen detection system showing light paths and detection system.[Color figure can be viewed in the online issue,which is avail-able at www.]
ORIGINAL ARTICLE
Cytometry Part A 75A:155À162,2009
多字157
Fig.S2).The method ud to concentrate bacteria,as described in the‘‘Results’’ction,involves u of a specific antibody against the bacterial strain that is being screened. This linked the bacteria to one or two magnetic beads.After washing with water,the coupled beads and bacteria were diluted with water into different concentrations from1:10to 1:100with a total volume of0.5mL.Each sample was vortexed to create homogeneity immediately before the spectrophot-ometer reading was taken in a UV–vis spectrophotometer (Genesys10uv,Thermo-Fisher,Waltham,MA)at350nm, which is a wavelength absorbed by iron oxide.Next,a200mT magnet was ud to draw the magnetic beads to the side of the tube so that the supernatant fluid could be removed.An equivalent amount of water was then added to the beads and shaken.The absorbance at350nm of the resuspended bead mixture was then measured in the spectrometer.The superna-
tant fluid was also measured in the spectrophotometer to check for stray magnetic beads to help determine the capture efficiency.Atomic force microscopy imaging was performed to determine that1
–4magnetic beads were bound to li cell(e Supp.Info.Fig.S3).
Microfluidic Chip Asmbly
The microfluidic chip was manufactured using conven-tional soft lithography techniques.First,the chip was designed using Ansoft HFSS v10.1software(Ansoft,Pittsburgh,PA). The design was tailored to create a sheet of fluid across a wide channel that would remain clo to the gold surface to allow for maximum capturing capabilities.The resin mold(Accura SI10polymer,3D Systems,Rock Hill,SC)for this chip was then created using a stereolithography machine(VIPER si2T SLA System by3D Systems).Stereolithography was ud becau it could easily create a multidimensional shape with one mold as compared with photolithography that would have required veral layers to achieve the height differentials needed to link the tubing(400l m)to the flat chip(50l m). Once the mold was cured with UV light,a1:10ratio of curing agent to polydimethylsiloxane(PDMS)polymer was mixed and then poured over the mold.This was allowed to cure overnight so that the pattern of the mold would be embedded onto the PDMS.Next,the PDMS was peeled off the resin mold,yielding microchannels for flow,and an inlet sample port was punched into the PDMS using a blunt-tipped28-gauge needle.The PDMS was then placed onto a clean glass microscope slide(microchannel face down)and oxidized using a Corona plasma etch system(BD20AC,Electro-Tech-nic Products,Chicago,
IL).The Corona system is a handheld device that creates a localized plasma field at room tempera-ture.This was ud to first oxidize the PDMS for$20s,and then the PDMS was presd onto the glass slide and heated on a hotplate at708C for15min to ensure a good al.The Co-rona process is important becau it does not require higher temperatures that may damage antibodies,peptides,or other capture molecules during the process of bonding the micro-fluidic structure to the gold contact-printed slide.After the PDMS chip preparation,tubing was inrted into the port cre-ated by the blunt needle and aled with liquid,uncured PDMS.The overall microfluidic chip asmbly is shown in Figure2.
Specific Pathogen Capture Process
The ba chip ud was a glass slide with a434array of 1-mm diameter gold spots(GWC Technologies,Madison, WI).Three biomolecules were ud to functionalize the gold spots.The first was an antibody that specifically li O157:H7.The cond was rabbit preimmune rum as a nega-tive control while the third was0.5%bovine rum albumin solution in water(BSA Fraction V,Sigma-Aldrich,St.Louis, MO)as a cond negative control.The array was patterned by applying1l L(at a concentration of100l g/mL)of a specific biomolecule to each gold spot,which was left to adsorb to the surface for1hour at room temperature.The chip was then washed with phosphate-buffered sali
ne(PBS),followed by 0.5%BSA to occupy any remaining active sites on the gold surface,as well as nonspecific sites on the antibodies.Two strains li O157:li DH5-a,were then lectively introduced to the array by pipetting1l L of one of the bacteria suspension onto each spot after being fluores-cently labeled with7.16l M SYTO-9and17.14l M PI(Invi-trogen).A hemacytometer was ud to determine the concen-trations of the bacteria,there was33106cells/mL li O157:H7and23108cells/mL li DH5-a.The bacteria were allowed to incubate at room temperature for10min,and unbound bacteria were washed away with PBS. Construction of Optical Components of SPR
Imaging Subsystem
Our bench-top SPR imaging system was built on an opti-cal breadboard using postmount optics,bad on the Kretsch-mann configuration(24),whereby a thin gold film is directly deposited on a slide placed on top of a SFL11equilateral prism (Edmund Optics,Barrington,NJ)that is ud to generate the necessary evanescent wave at the metal-dielectric interface by means of total internal reflection,as described earlier.An inex-pensive635-nm diode lar(Edmund Optics)was placed on top of a SFL11equilateral prism(Edmund Optics)and ud to illuminate the sample.The prism is subquently mounted on
a Figure2.Schematic of the overall microfluidic chip asmbly pro-cess showing the steps from the patterns of the microfluidic chan-nels and grid to the finished SF10glass chip:(A)SU-8photoresist is patterned to make a silicon mold;(B)PDMS is poured on silicon mold and allowed to t;(C)PDMS replica with microchannels, (D)photolithography on SF10glass chip;(E)gold evaporation and lift off;(F)PDMS chip is bound over the gold spot array using a Corona etcher.The constructed microfluidic chip(right).
ORIGINAL ARTICLE
158Hybrid Imaging Pathogen Detector
goniometer (Thorlabs,Newton,NJ),which is ud to control the incidence angle of the lar.An inexpensive computer-con-trolled charge coupled device (CCD)camera (Pt.Gray Rearch,Richmond,BC,Canada)is then ud to collect the SPR image.
Fluorescence Imaging of Bacteria
For this article,the capture of fluorescently labeled bacteria was assd using epi-fluorescence microscopy (Nikon Diaphot Inverted Fluorescence Microscope;Nikon,Melville,NY).A flu-orescence image of each spot was captured for both SYTO-9(475/20nm BPD excitation,500nm LPD,and 525/15nm BPD emission filters were ud;Omega Optics,Austin,TX)and PI (475/20nm BPD excitation,500nm LPD,and 585/30nm BPD filters;Omega Optics),and the prence of captured pathogen was quantified by using NIH ImageJ analysis software (v/ij/).The amount of captured bacteria was calcu-lated from the SYTO-9images for each spot.The images were first thresholded,using a threshold pixel value of 48(8-bit image),which was 10%above the background intensity.Then the ‘‘Analyze Particles’’function of ImageJ was ud to deter-mine the number of bacteria on each spot.Viability of the bac-teria was determined using the PI images.The same ImageJ procedure was ud to quantify the amount of PI-positive dead cells,except that the threshold ud was a pixel value of 30,becau a control sample of SYTO-9only showed a fluorescent signal maximum of 25with the PI imaging ttings.
Design and Construction of the Portable Hybrid Imaging System
水草种类
A more portable prototype of the hybrid imaging system has been subquently designed and is shown in Figure 3.This prototype utilizes the Microptic optical cage system (AF Opti-cal,Fremont,CA)to make a three-armed device.The SPR arms are also bad on the Kretschmann configuration.This SPR
imaging tup captures data for the entire probe array,includ-ing controls to detect nonspecific binding as described later in this proposal,simultaneously on a CCD camera.A BK7glass right angle prism (Thorlabs,Newton,NJ)is mounted at the center of the three arms.The prism mounts contain variable angle slots,which allow the SPR illumination arm and detection arm,to swing to create the appropriate incident angle.The first SPR imaging arm,the illumination arm,consists of a 635-nm diode lar (Thorlabs)that is then shaped by a beam expander to illuminate the whole sample.A polarizer on a rotary mount (AF Optical)is ud to generate p-polarized light.The SPR detection arm consists of a 43long working distance objective (Olympus,Nashua,NH),a focusing lens,and a CCD camera (Pt.Gray Rearch)to capture the SPR image,whereas the epi-fluorescence imaging arm us a 43objective to image the sam-ple,with the standard excitation (480/20nm band pass)dichroic (500nm long pass dichroic)and emission filter tup (515/20or 565/30nm band pass).An ultra-bright 470nm LED is ud to illuminate the sample (LumiLEDs,San Jo,CA)for molecular imaging of the fl
uorescently stained bacteria,and a monochromatic CCD camera (Pt.Gray Rearch)is ud to image the sample.Both cameras are connected to a notebook computer (Dell Inspiron 1300,Dell Computers,Round Rock,TX)where frame grabber software acquires the images (Pixel-Scope Pro,Wells Rearch,Lincoln,MA).The microfluidic chip was placed on top of the prism where it can be imaged by both SPR imaging and epi-fluorescence molecular imaging.
R ESULTS
Immunomagnetic Preconcentration and Capture Efficiency
门冬氨酸氨氯地平片from
挑拨离间是什么意思
Figure 3.A schematic drawing of the next generation portable SPR imaging hybrid imaging system with associated microfluidic chip (left).The components of the SPR system consists of a 635-nm lar diode (A ),a beam expander (B ),a polarizer (C ),a prism (D ),continuous angle mounting plate (E ),43objective (F ),and a CCD camera.The microfluidic chip is inrted on top of the prism,and molecular imaging is performed from the top side of the chip using an LED illuminated epifluorescence unit.A picture of the constructed portable SPR ima-ging hybrid imaging system (right).The mini-optical rail system gives flexibility and structural integrity to the device so that it can be lf-supporting and portable.
ORIGINAL ARTICLE
Cytometry Part A 75A:155À162,2009
159
a spectrophotometer as described previously.To demonstrate the physical capture ability of bound bacterial cells,as well as decrea the overall sample volume,the magnetic beads were expod to a
magnetic field that forced them to the side of the sample tube allowing for the removal of the supernatant (con-taining unbound bacteria as determined by microscopy).For all concentrations tested,there was greater than 90%recovery as shown in Table 1.The remaining 10%of the beads were likely lost during the wash and resuspension steps.There was
no indication of magnetic beads left in the supernatant fluid bad on spectrophotometer readings (data not shown).A small volume of the purified sample was obrved under the microscope.The fluorescence of the stained bacteria indicated a successful linkage since the beads do not fluoresce (e Supp.Info.).
Specific Capture li O157:H7on Biochip
The biochip was patterned with one of three biomole-cules on each gold spot.The spots were either functionalized with li O157:H7antibody or with one of the negative controls:rabbit preimmune rum or 1%BSA (Fraction V,Sigma Chemical,St.Louis,MO)as shown in Figure 4.This diagram also shows which spots were expod to E.coli O157:H7and which ones were expod to the negative control strain li DH5-a .To demonstrate specific capture li O157:H7,bacteria should only be prent on the gold spots functionalized li O157:H7antibodies that were expod to
The binding of pathogens to each spot was quantified by image analysis using NIH ImageJ software.The results of this analysis are shown in Figure 5.The only conditions where a significant amount of coverage occurred were on gold spots
Table 1.The recovery of li complexes measured by spectrophotometry at 350nm
INITIAL DILUTION
INITIAL A 350nm
A 350nm AFTER MAGNETIC RECOVERY
PERCENTAGE RECOVERY (%)
1:100.600A 0.547A 911:200.199A 0.185A 931:500.105A 0.097A 92.41:1000.065A 0.062A 95
A spectrophotometer was ud to measure the absorbance at 350nm of veral dilutions (1mL volume)magnetic bead bound li complexes.After an initial absorbance at 350nm was measured,a magnet was ud to concentrate the complexes against the side of the tube,and then the supernatant was removed.The samples were then resuspended in 1mL of distilled water and the absorbance was measured again.Using Beer’s law,the percent recovery of the li complexes was greater than
90%.
Figure 4.Top left:The pattern of functionalization of the gold array.Gold spots were functio-nalized with either E.coli O157:H7antibody,rabbit pre-immune rum,or 1%BSA.Then either E.coli O157:H7(black background)or E.coli DH5-a (white background)were added to each spot.Top right:A fluorescence image of the gold array demonstrating the lective capture of pathogens.Bottom:Fluorescent images of one spot (row 1,column 3)showing the SYTO-9signal (left)that shows how many cells are bound,and the PI sig-nal (right)that shows the level of dead cells.The li O157:H7was bound to be 97.7%viable.
ORIGINAL ARTICLE
160Hybrid Imaging Pathogen Detector
