Highly Sensitive Laser Scanning of Photon ... - ACS Publications

17 downloads 6 Views 4MB Size Report
Dec 24, 2015 - UCNPs covered by a dye-doped silica shell were separated by agarose gel electrophoresis and ... luminescent nanoparticles such as quantum dots are widely ..... encapsulated with a 20 nm silica layer and functionalized with.

This is an open access article published under an ACS AuthorChoice License, which permits copying and redistribution of the article or any adaptations for non-commercial purposes.

Article pubs.acs.org/ac

Highly Sensitive Laser Scanning of Photon-Upconverting Nanoparticles on a Macroscopic Scale Andreas Sedlmeier,† Antonín Hlavácě k,‡ Lucia Birner,† Matthias J. Mickert,† Verena Muhr,† Thomas Hirsch,† Paul L. A. M. Corstjens,§ Hans J. Tanke,§ Tero Soukka,∥ and Hans H. Gorris*,† †

Institute of Analytical Chemistry, Chemo- und Biosensors, University of Regensburg, 93040 Regensburg, Germany Central European Institute of Technology (CEITEC), Masaryk University, Brno 625 00, Czech Republic § Department of Molecular Cell Biology, Leiden University Medical Center, 2300 RC Leiden, The Netherlands ∥ Department of Biochemistry/Biotechnology, University of Turku, 20520 Turku, Finland ‡

S Supporting Information *

ABSTRACT: An upconversion laser scanner has been optimized to exploit the advantages of photon-upconverting nanoparticles (UCNPs) for background-free imaging on a macroscopic scale. A collimated 980 nm laser beam afforded high local excitation densities to account for the nonlinear luminescence response of UCNPs. As few as 2000 nanoparticles were detectable, and the linear dynamic range covered more than 5 orders of magnitude, which is essentially impossible by using conventional fluorescent dyes. UCNPs covered by a dye-doped silica shell were separated by agarose gel electrophoresis and scanned by a conventional fluorescence scanner as well as the upconversion scanner. Both optical labels could be detected independently. Finally, upconversion images of lateral flow test strips were recorded to facilitate the sensitive and quantitative detection of disease markers. A marker for the parasitic worm Schistosoma was used in this study.


The focus on the luminescence enhancement, however, has overshadowed the instrument development.10 The readout of UCNPs is intrinsically more challenging compared to conventional luminescent reporters because two or more photons must be absorbed for each emitted photon of higher energy. The power-law-dependent emission of UCNPs requires high near-IR excitation intensities if one aims for the highest detection sensitivity. Laser scanning microscopy can achieve very high power densities in the confocal spot of the laser11 and has frequently been used for in vitro studies of cells labeled with UCNPs.12−16 The long lifetimes of UCNPs, however, slow down the scanning speed. Consequently, wide-field microscopy is preferred for in vitro studies that involve fast intracellular processes.17 If the laser beam is collimated on an area of about 60 × 60 μm2 in the focal plane of the sample, the excitation density is sufficient for detecting single UCNPs by wide-field microscopy.18 In vivo studies and small animal imaging19−21 require an imaging area that is much larger than the field of view of a microscope and are mainly based on self-built instruments equipped with a near-IR laser beam expander to enlarge the excitation area.22 While the beam expander leads to a decrease in the excitation density, the load of UCNPs applied to the

uminescence-based techniques belong to the most important tools in analytical chemistry because they are highly sensitive, selective, and versatile and provide a high spatial resolution. 1,2 Organic fluorophores and various luminescent nanoparticles such as quantum dots are widely used as reporters in bioimaging.3 Conventional luminescent reporters, however, suffer from excitation by ultraviolet (UV) or visible light, which is also absorbed by surrounding materials and leads to autofluorescence, light scattering, and photodamage. Consequently, these conventional labels result in a high background noise. Photon-upconverting nanoparticles (UCNPs) emit light of shorter wavelengths under near-infrared (near-IR) excitation, which hardly interacts with biological materials. The anti-Stokes emission avoids background interference, and the large antiStokes shifts allow for an excellent separation of excitation and emission channels. Additionally, UCNPs do not photobleach or blink. 4,5 The energy transfer upconversion (ETU) of lanthanide-doped nanocrystals is based on long-lived energy states in inner f-orbitals and is several orders of magnitude more efficient than two-photon excitation or second harmonic generation.6 In recent years, great research efforts have been made to enhance the luminescence of UCNPs, e.g., by optimizing the crystalline host matrix or the lanthanide dopant composition or by designing novel core−shell architectures.7,8 Hexagonal NaYF4:Yb,Er belong to the brightest UCNPs known to date.9 © 2015 American Chemical Society

Received: November 2, 2015 Accepted: December 24, 2015 Published: December 24, 2015 1835

DOI: 10.1021/acs.analchem.5b04147 Anal. Chem. 2016, 88, 1835−1841


Analytical Chemistry tissue or animal is typically high enough for detection.14,23−25 Bioanalytical applications of UCNPs face similar challenges. While some microarrays can be imaged under a microscope,26 a different readout scheme is required for microarrays on the bottom of a microtiter well.27,28 Lateral flow assays29 and gel electrophoresis are further examples that demonstrate the need for an imaging area beyond the field of view of a microscope. Here, we have optimized a scanner (Figure 1) for the highly sensitive detection and imaging of UCNPs on a macroscopic

type NaYF4:Yb3+,Er3+,Gd3+ (12 or 25 nm in diameter) were silica-coated at room temperature by water-in-oil (reverse) microemulsion. Oleate-coated UCNPs were diluted to a volume of 730 μL in cyclohexane, which resulted in UCNP concentrations of 0.62 mg/mL (diameter, 25 nm) or 0.31 mg/ mL (diameter, 12 nm), respectively. Then, 42 mg of polyoxyethylene (5) nonylphenyl ether (Igepal CO-520, Sigma-Aldrich) and different amounts (1.0 or 2.0 μL) of the fluorescent silanization reagent were added and stirred vigorously (vortex, 600 rpm) for 10 min. The microemulsion was formed by adding 6.1 μL of 14% aqueous ammonia solution and stirring again for 180 min. Subsequently, different amounts (0.8, 1.7, or 3.4 μL) of 25% aqueous carboxyethylsilanetriol (CEST; abcr, www.abcr.de) were added. The mixture was first sonicated for 15 min and then stirred for 40 h. COOH-FITC-UCNPs were precipitated by adding 1.0 mL of acetone and washed three times (centrifugation at 10000g and resuspension in 1.0 mL of acetone). Finally, COOH-FITCUCNPs were dispersed in bidistilled water to yield concentrations of 4.5 mg/mL (25 nm diameter) or 10 mg/mL (12 nm), respectively. The dispersions were stored at 4 °C and were stable for at least 1 year. Laser Scanning. (1) A conventional laser scanner (PharosFX, Bio-Rad, www.bio-rad.com) was set to record the emission of fluorescein (λex = 488 nm/λem = 532 nm) with a lateral resolution of 50 μm. (2) A commercial CHAMELEON microtiter plate reader from Hidex (www.hidex.com) was adapted for upconversion scanning as shown in Figure 1: 980 nm excitation light from a continuous 980 nm laser (4 W; Wavespectrum, www.wavespectrum-laser.com) was transmitted via a multimode fiber waveguide (200 μm in diameter). The laser was temperature controlled to ensure a stable excitation power. The laser output at the end of the fiber was 2.8 W as determined by a thermopile sensor (LaserPAD laser power analysis system with LM-10 thermal sensor, Coherent, www. coherent.com). After passing through a collimator and longpass filter (λcut‑on, 850 nm), the excitation light was reflected by an aluminum mirror on the focal spot of the laser. The parabolic laser profile covered an area of approximately 0.5 mm2 (800 μm in diameter) which resulted in an average power density of 560 W/cm2. The excitation time was set to a maximum of 500 ms to avoid thermal damage to the sample. The emission light of UCNPs was collected via optical lenses and passed through filter combinations including an 825 nm short-pass filter and a 535 ± 25 nm bandpass filter for the green emission of Er-doped UCNPs. Further details on the selection of optical filters are provided in the Supporting Information, Table S1 and Figure S1. Er3+-doped UCNPs enable a better separation of the signal from the background than Tm3+-doped UCNPs and were used throughout this study. The emission light was recorded on a photomultiplier tube (R4632 from Hamamatsu, www.hamamatsu.com). A software-controlled stage was moved automatically in x- and y-directions to scan the sample. The tightly collimated laser beam resulted in a high local excitation intensity, which was important since the emission signal increases nonlinearly with the excitation power. While the laser spot had a diameter of approximately 0.8 mm, smaller scanning steps were feasible as the highest excitation intensity was found in the center of the collimated beam. The precision of the automated stage used to move the sample under the laser spot limited the scanning steps to a minimum of 0.1 mm.

Figure 1. Schematic view of the upconversion scanner. The beam of the continuous wave laser diode (4 W) passes through a long-pass filter (λcut‑on, 850 nm) and is reflected by a mirror on the object. The upconversion emission passes a set of lenses and filters and is recorded on a photomultiplier tube.

scale. Upconversion scanning represents essentially the opposite approach of small animal (wide-field) imaging because the beam expander is replaced by a collimated laser to achieve high local excitation densities and scan the sample in the focal plane of the laser.

EXPERIMENTAL SECTION Synthesis of UCNPs. Hexagonal-phase UCNPs were synthesized in oleic acid/1-octadecene at 300 °C as previously described:30 NaYF4:Yb3+, Er3+, Gd3+ (approximately 12, 25, and 30 nm in diameter) or NaYF4:Yb3+,Tm3+ (approximately 23 nm in diameter). The smaller sizes of UCNPs were adjusted by the amount of Gd3+-doping. This synthesis resulted in monodisperse UCNPs coated with oleate. The size and monodispersity of the samples were determined by transmission electron microscopy (TEM, Philips CM12, 120 kV; www.fei. com).30 Er3+-doped UCNPs show typical emission bands in the green (521/543 nm) and red (657 nm) and Tm3+-doped UCNPs in the blue (475 nm) and near-IR (800 nm) (Supporting Information Figure S1). A dispersion of 1 mg/ mL in cyclohexane was stored for several months at 4 °C. Silica Coating and Silanization of UCNPs. A solution of 6.4 mM (3-aminopropyl)triethoxysilane (APTES, 98%; SigmaAldrich, www.sigma-aldrich.com) and 3.2 mM fluorescein 5(6)isothiocyanate (FITC, >90%; Sigma-Aldrich) was prepared in dried N,N-dimethylformamide (p.a.; Penta, www. pentachemicals.eu) and incubated for 2 h at room temperature. The FITC−APTES adduct was mixed with tetraethyl orthosilicate (TEOS, 98%; Sigma-Aldrich) in a ratio of 23:80 (v/v) to obtain a fluorescent silanization reagent that renders UCNPs visible by conventional laser scanners.31 UCNPs of 1836

DOI: 10.1021/acs.analchem.5b04147 Anal. Chem. 2016, 88, 1835−1841


Analytical Chemistry

Figure 2. Sensitivity and dynamic range of upconversion scanning. (A) UCNPs (30 nm in diameter) are serially diluted, and six replicate spots are prepared on cellulose pads (3 mm in diameter) to obtain amounts in the range of 100 ng to 1 pg UCNP per spot. Upconversion images (scanning step size, 500 μm) are plotted in a logarithmic scale and shown in pseudocolors. (B) The graph shows a linear dynamic range over 5 orders of magnitude. Each data point represents the mean and standard deviation of at least 10 spots and two independent dilution series. The hatched line indicates the very low LOD (102.54 counts), which enables the detection of 2 000 nanoparticles in the exaction area of the laser.

and a gel was cast (2% agarose, 83 mm × 70 mm, 1 mm thick). The silica-coated UCNPs were diluted in water and 50% (w/w) aqueous glycerol solution to yield a final glycerol concentration of 5%, which was loaded onto the agarose gel. After electrophoresis at 100 V for 30 min, gels were dried on a flat poly(methyl methacrylate) surface at 50 °C under air flow. The dried gels were scanned in the upconversion and downconversion mode. Lateral Flow Assay for Schistosoma Circulating Anodic Antigen. Lateral flow assay (LFA) test strips consisted of a test (T) line containing a circulating anodic antigen (CAA)-specific monoclonal antibody (MαCAA, No. 147-3G4, LUMC Department of Parasitology) to capture CAA as well as a flow control (FC) line with a goat-antimouse polyclonal antibody (GαM) to bind all MαCAA UCP reporter particles that passed the T line. The T line containing 200 ng of MαCAA per 4 mm and the FC line with 100 ng of polyclonal goat-antimouse antibody (GαM, Sigma-Aldrich, M8642) were prepared on a laminated nitrocellulose membrane (Millipore, www.merckmillipore.com) by using a CAMAG Automatic TLC Sampler 4 (ATS4; BCON Instruments B.V. www.bconinstruments.nl). The laminated nitrocellulose, a glass fiber sample application pad, and a paper absorbent pad were attached to a plastic backing and cut into strips of 4 mm as described earlier.34,35 The LF strips were stored dry in the presence of a desiccant (silica) and were stable for at least 1 year. The MαCAA antibody was also conjugated to Y2O2S:Yb3+,Er3+ UC reporter particles (400 nm in diameter, encapsulated with a 20 nm silica layer and functionalized with C10-carboxyl groups; OraSure Technologies, www.OraSure. com). 3 6 Using a standard two-step 1-ethyl-3-(3-

Limit of Detection and Dynamic Range of the Upconversion Scanner. Oleic acid-coated UCNPs (Er3+doped, 30 nm in diameter) were diluted in cyclohexane to obtain concentrations in the range of 100 ng/μL to 1 pg/μL. For each concentration, six circular cellulose pads (3 mm in diameter) were punched out and attached to a planar polystyrene surface via double-faced adhesive tape. A volume of 1 μL was applied to each pad and completely absorbed. Only cyclohexane was applied to the cellulose pads (background (bg) of nine replicates) to determine the limit of detection (LOD = meanbg + 3 × standard deviationbg). After drying, the upconversion luminescence was analyzed on the areas defined by the cellulose pads (collecting time, 500 ms; scanning step size, 500 μm). Images were rendered by ImageJ (http://imagej. nih.gov). The mass of a single UCNP (NaYF4:Yb,Er) was calculated as follows: UCNPs have a spherical shape, and the density (ρ) of NaYF4:Yb,Er is equated with NaYF4.32 4230 kg/m 3

density (ρ) of NaYF4 radius (r )

15 nm

volume (V = (4/3)πr ) 1.41 × 10−23 m 3 3

mass (m = ρV )

5.98 × 10−20 kg = 5.98 × 10−5 pg

Agarose Gel Electrophoresis. COOH-FITC-UCNPs were separated by vertical agarose gel electrophoresis (MiniProtean tetra cell, Bio-Rad, www.biorad.com) as described earlier.33 Agarose LE (Lonza, www.lonza.com) was dissolved in electrophoresis buffer [45 mM tris-2-amino-2hydroxymethylpropane-1,3-diol (>99%, Sigma-Aldrich), 45 mM H3BO3 (p.a., Penta), pH 8.6, containing 1.0% (w/w) sucrose (p.a., Penta) and 1.0% (w/w) glycerol (p.a., Penta)], 1837

DOI: 10.1021/acs.analchem.5b04147 Anal. Chem. 2016, 88, 1835−1841


Analytical Chemistry

Figure 3. Agarose gel electrophoresis. UCNPs (25 nm in diameter) are silica coated by adding different amounts of TEOS (containing fluoresceinAPTES) and CEST to the microemulsion. Without further purification, each sample is loaded in duplicate into the pockets of an agarose gel (2%, 83 mm × 70 mm): (1) 1.0 μL of TEOS and 0.8 μL of CEST; (2) 1.0 μL of TEOS and 1.7 μL of CEST; (3) 1.0 μL of TEOS and 3.4 μL of CEST; (4) 2.0 μL of TEOS and 0.8 μL of CEST; (5) 2.0 μL of TEOS and 1.7 μL of CEST; (6) 2.0 μL of TEOS and 3.4 μL of CEST. After electrophoresis (100 V, 30 min), the gel is dried on a flat surface and scanned by (A) the upconversion scanner (500 μm resolution) and (B) a conventional fluorescence scanner (50 μm resolution). Excess fluorescent dye is not detectable in panel A but accounts for the largest part of fluorescence emission in panel B.

upconversion signal was integrated over the entire spot area. Figure 2B shows a linear relationship between the amount of UCNPs and the upconversion signal ranging over more than 5 orders of magnitude. The wide range of signal intensities can only be visualized by using a logarithmic scale for the image as well as for the graph. It should be noted that the highest amount of 100 ng of UCNPs is not included in the graph because the signal leads to detector saturation. If the excitation was reduced by neutral density filters, it would be possible to obtain a linear response for even higher amounts of UCNPs. This extremely broad dynamic range is a hallmark of upconversion scanning, which cannot be achieved by using organic fluorophores or other downshifting materials because both the relatively high background of autofluorescence and inner filter effects level out larger signal differences. Due to the absence of autofluorescence, the background signal was very low and homogeneous (mean, 325 counts; standard deviation, 7 counts). The detection limit of UCNPs was 3 pg, which amounts to 30000 UCNPs distributed over the area defined by the cellulose pad (7 mm2). This area was about 14 times larger than the excitation area of the collimated laser beam (diameter, 0.8 mm; area, 0.5 mm2). Thus, as few as 2000 nanoparticles can be detected by the upconversion scanner. Agarose Gel Electrophoresis. Gel electrophoresis has become an important tool for the rapid and efficient optimization of nanoparticle surfaces.38,39 Recently, we adapted agarose gel electrophoresis for the separation and characterization of UCNPs.33 Since we had no instrument for the twodimensional readout of upconversion luminescence, a fluorescent dye was added into the silica shell to locate the nanoparticles in the gel by a conventional fluorescence imaging system. The direct detection of the upconversion luminescence in a gel, however, would make gel electrophoresis more widely applicable for the optimization of UCNP surface modifications without the need for additional labels. Here, we have employed different amounts of TEOS, fluorescein-APTES, and CEST to generate silica-coated UCNPs (25 nm in diameter) equipped with a fluorescent dye and carboxyl groups to confer a negative surface potential. After agarose gel electrophoresis, the gels were dried and

(dimethylamino)propyl)carbodiimide hydrochloride mediated reaction, MαCAA was conjugated to the UC reporter using a concentration of 25 μg of antibody/(mg of UC particles). For a detailed conjugation protocol see ref 37. The UC conjugate (1 mg/mL) was stored at 4 °C in UC storage buffer (50 mM glycine, 0.3% (v/v) Triton X-100, 0.1% (w/v) NaN3). CAA (obtained from a crude Schistosoma worm preparation after extraction with 4% (w/v) of trichloroacetic acid)34 was serially diluted (10 ng/mL to 10 pg/mL) in high salt lateral flow buffer (HSLF: 100 mM HEPES, pH 7.2, 270 mM NaCl containing 0.5% (w/v) Tween 20 and 1% (w/v) BSA). This standard series (20 μL) was mixed with 100 ng of UC particles in HSLF and applied to the LFA sample application pad as described earlier.34 Three dilution series were tested, each including one negative control containing buffer only. The upconversion luminescence was recorded in a dry and wet state of the LFA strip. Assay results are presented as a normalized ratio (R) value: R is the signal measured at the T line divided by the signal of the FC line (R = T/FC).37 The LOD was determined from three LFA strips containing buffer only as meanbuffer‑only + 3 × standard deviationbuffer‑only. The dry LF strips can be stored indefinitely for reanalysis. GraphPad Prism 6 (www.graphpad.com) was used for a nonlinear regression analysis with a four-parameter logistic function: ratio[CAA] =

ratiomax − ratio bg 1+


( ) [CAA] EC50

+ ratio bg

RESULTS AND DISCUSSION Sensitivity and Dynamic Range of Upconversion Scanning. The sensitivity and dynamic range of upconversion scanning was determined by depositing oleic acid-coated UCNPs (NaYF4:Yb3+,Er3+,Gd3+, approximately 30 nm in diameter) dispersed in cyclohexane on cellulose pads (3 mm in diameter). Figure 2A shows six replicate pads containing defined amounts of UCNPs including one sample without UCNPs to determine the LOD. After drying, the samples were scanned with a signal acquisition time of 500 ms and the 1838

DOI: 10.1021/acs.analchem.5b04147 Anal. Chem. 2016, 88, 1835−1841


Analytical Chemistry

Figure 4. Gel scanning in the upconversion mode (500 μm resolution). Very small (12 nm in diameter) and monodisperse UCNPs are coated with silica by adding 2.2 μL of TEOS (containing fluorescein-APTES) and 2.0 μL of CEST into the microemulsion. Without further purification, a 10× dilution series of UCNPs is loaded in duplicate into the pockets of an agarose gel (2.8%, 83 mm × 70 mm): (1) empty; (2) 0.75 ng; (3) 7.5 ng; (4) 75 ng; (5) 750 ng; (6) 7,500 ng; (7) 75,000 ng. After electrophoresis (100 V, 45 min), the gel is dried on a flat surface and scanned. The upconversion signal intensities are plotted in pseudocolors using either (A) a linear or (B) a logarithmic scale bar (units, counts per second). Only the strongest bands are observable in panel A while the logarithmic scale in panel B also reveals weaker bands. The insets in panel A show enlarged images of the left band in double lane 7 scanned with different step sizes as indicated. Decreasing the step size improves the resolution and reveals more details.

Figure 5. Upconversion readout of lateral flow assays. (A) Serial dilutions of CAA in the range of 10 ng/mL (strip 1) to 10 pg/mL (strip 7) and samples without CAA (strip 8) are applied to LFA strips. The outline of the strips is indicated by a white frame. T line and FC line can be readily distinguished from the background (log scale). (B) The assay results are plotted as the ratio (T line/FC line). Error bars indicate the standard deviation of three independent dilution series. Each LFA series is scanned in a dry (red) and a wet state (blue). The curves show a regression analysis with a four-parameter logistic function. The LODs are calculated from the three buffer-only samples and are indicated as hatched lines in respective colors.

very wide dynamic range of upconversion detection, it is not possible to display simultaneously low and high signal intensities in a linear upconversion image. Thus, only the band containing the highest load of UCNPs (double lane 7) is visible in the linear image (Figure 4A). By contrast, Figure 4B is plotted in a logarithmic scale and shows both strong and weak bands in a single image. Figure 4B reveals bands containing nanoparticle aggregates, but it should be noted that low signals are overrepresented. The same electrophoretic pattern as in double lane 7 is also observed in 5 and 6, where the amount of UCNPs is reduced by a factor of 10 or 100, respectively. In general, larger amounts (by mass) of these small UCNPs (Figure 4) are necessary to yield the same signal intensity compared UCNPs of 25 nm in diameter (Figure 3) because the upconversion luminescence strongly decreases with the nanoparticle diameter. Next, we investigated the influence of different scanning step sizes on the resolution of the image. The enlarged images of the

scanned with both the upconversion scanner and a conventional fluorescence scanner (Figure 3). Both signals can be detected independently and without interference. An identical electrophoretic pattern is evident in both images. In the upper part of the images distinct bands are visible that indicate various degrees of nanoparticle aggregation. Monodisperse nanoparticles without aggregation can be observed as a single band in double lane 3. By contrast, the gel pockets in double lane 5 mainly contain aggregates that are too large for entering the gel matrix. The most striking difference in Figure 3B, however, is the large red blot of unbound fluorophore, which is only visible in the downshifting mode and demonstrates the independent detection of both labels. Agarose gel electrophoresis was also used for investigating small silica-coated UCNPs of 12 nm in diameter that are only weakly luminescent. Various amounts of UCNPs were loaded on an agarose gel, separated by electrophoresis, and scanned in the upconversion mode (Figure 4). As a consequence of the 1839

DOI: 10.1021/acs.analchem.5b04147 Anal. Chem. 2016, 88, 1835−1841


Analytical Chemistry

compared to the dried strip. By contrast, the tightly collimated laser of the upconversion scanner ensures a high excitation power density which always keeps a high proportion of Yb3+ in the excited state. These excited sensitizer ions are then available for energy transfer to the activator Er3+ even if there is a constant deactivation due to water quenching. The relatively large size of the upconversion reporter particles (400 nm in diameter) further reduces the surface to volume ratio and concomitant quenching effects. Figure 5B shows that there is only a relatively small difference in the readout of the dry (LOD, 30 pg/mL CAA) and wet (LOD, 100 pg/mL CAA) LFA strips.

left band in double lane 7 (Figure 4A) were recorded by increasing the scanning step size from 180 to 500 μm. With a resolution of 180 μm, the upconversion scanner comes close to conventional fluorescence scanners that typically feature a scanning step size of 50 μm. Due to the limited scanning speed, however, it takes a very long time to scan the entire area with a high resolution, and for most analytical applications a step size of 500 μm is sufficient. Lateral Flow Assay for Schistosoma Circulating Anodic Antigen. Lateral Flow Assays (LFA) are an important tool for point-of-care (POC) testing because they are fast and easy to perform, and cost-efficient.40 Most current LFAs rely on the direct readout of colloidal gold by eye. As this detection scheme cannot provide the highest sensitivity, other approaches have been developed for detecting analytes of low abundance, which can be particularly important for disease markers. For example, it has been demonstrated that upconversion provides the ability to improve the detection sensitivity by a factor of 100 compared to immunogold labeling.36 In this context a distinction has to be made between the instrumental sensitivity, i.e., the amount of UCNPs detectable above the background, and the assay sensitivity, i.e., the concentration of a disease marker detectable above the background. Furthermore, the assay sensitivity does not only depend on optical background interference, which is very low for photon-upconversion as shown earlier, but also on nonspecific binding and random deposition of the reporter particles during immunochromatography. Previously, such LFA strips were read out in one dimension by line-scanning along the flow direction.29,34−37 The results of line-scanning are shown in Supporting Information Figure S3. Here, we have taken images of the entire test strips. An upconversion LFA was used for detecting a distinct antigen produced and regurgitated by the parasitic worm Schistosoma as described earlier.34 The CAA is an important disease marker with a glycosaminoglycan-like structure of repeating disaccharide units and indicates an ongoing active infection.29 CAA was serially diluted from 10000 to 10 pg/mL, mixed with MαCAA UCP and applied to the LFA strips (Figure 5A). The FC line is always clearly detectable while the signal of the T line decreases with the amount of analyte. However, even the LFA strip containing buffer only shows a T line signal higher than the background, which can be attributed to nonspecific binding of the UC reporter particles. Additionally, UC reporter particles deposit randomly during immunochromatography, which leads to a significant but slightly decreasing background signal along the flow direction. Consequently, it is no longer the optical detection but rather further assay parameters that limit the sensitivity of upconversion-based LFAs. The luminescent signals of the T line and FC line were background-corrected and converted to ratiometric values (T line/FC line) that are independent of absolute signal intensities.41 The LOD was calculated based on three LFA strips containing buffer only. Approximately 30 pg/ mL of CAA was detectable on a dry strip above the LOD (Figure 5B), which is equivalent to about 100000 molecules of CAA (MW, ∼100 kDa; sample volume, 20 μL). For POC applications, it is commonly desirable to read out LFA strips directly after immunochromatography, i.e., when the strip is still wet. Water, however, exerts a strong quenching effect on the upconversion luminescence due to multiphonon deactivation of the sensitizer Yb3+,42 and the absolute signal intensities are typically much lower in an aqueous environment

CONCLUSION We have demonstrated a highly sensitive macroscopic scanning approach for UCNPs. The unique photoluminescent features of UCNPs combined with high local power densities in the focal spot of the laser enable the detection of as few as 2000 UCNPs as well as a very wide dynamic range of more than 5 orders of magnitude. Essentially, imaging in the upconversion mode breaks both the lower detection limit of conventional (downshifting) fluorophores because there is no optical background interference and the upper limit because there are no inner filter or self-quenching effects. In the past, these features were a hallmark of radioactive screening. UCNPs, however, only require short signal acquisition times, provide a higher resolution, and create no radioactive waste. Consequently, our scanning approach exploits the best of luminescent and radionuclide labels.


S Supporting Information *

The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acs.analchem.5b04147. Selection of optical filter sets and line-scanning experiments of LFAs (PDF)


Corresponding Author

*Tel.: +49-941-943-4015. Fax: +49-941-943-4064. E-mail: [email protected] Notes

The authors declare no competing financial interest.

ACKNOWLEDGMENTS We thank Ville Haaslahti from Hidex Inc. and Juho Terrijärvi for the support with the scanning instrument and Claudia de Dood from the Leiden University Medical Center for preparing the LFA test strips. We acknowledge financial support from the COST action CM1403 “The European Upconversion Network: From the Design of Photon-Upconverting Nanomaterials to Biomedical Applications”. A.H. acknowledges funding by the Program of “Employment of Newly Graduated Doctors of Science for Scientific Excellence” (Grant CZ.1.07/2.3.00/ 30.0009).


(1) Wolfbeis, O. S. Chem. Soc. Rev. 2015, 44, 4743−4768. (2) Schäferling, M. Angew. Chem., Int. Ed. 2012, 51, 3532−3554. (3) Demchenko, A. P. Methods Appl. Fluoresc. 2013, 1, 022001.


DOI: 10.1021/acs.analchem.5b04147 Anal. Chem. 2016, 88, 1835−1841


Analytical Chemistry

(35) van Dam, G. J.; de Dood, C. J.; Lewis, M.; Deelder, A. M.; van Lieshout, L.; Tanke, H. J.; van Rooyen, L. H.; Corstjens, P. L. A. M. Exp. Parasitol. 2013, 135, 274−282. (36) Corstjens, P. L.; Li, S.; Zuiderwijk, M.; Kardos, K.; Abrams, W. R.; Niedbala, R. S.; Tanke, H. J. IEE Proc.: Nanobiotechnol. 2005, 152, 64−72. (37) Corstjens, P.; Zuiderwijk, M.; Brink, A.; Li, S.; Feindt, H.; Neidbala, R. S.; Tanke, H. Clin Chem. 2001, 47, 1885−1893. (38) Beskorovaynyy, A. V.; Kopitsyn, D. S.; Novikov, A. A.; Ziangirova, M.; Skorikova, G. S.; Kotelev, M. S.; Gushchin, P. A.; Ivanov, E. V.; Getmansky, M. D.; Itzkan, I.; Muradov, A. V.; Vinokurov, V. A.; Perelman, L. T. ACS Nano 2014, 8, 1449−1456. (39) Sapsford, K. E.; Tyner, K. M.; Dair, B. J.; Deschamps, J. R.; Medintz, I. L. Anal. Chem. 2011, 83, 4453−4488. (40) Ngom, B.; Guo, Y. C.; Wang, X. L.; Bi, D. R. Anal. Bioanal. Chem. 2010, 397, 1113−1135. (41) Gorris, H. H.; Wolfbeis, O. S. Angew. Chem., Int. Ed. 2013, 52, 3584−3600. (42) Arppe, R.; Hyppänen, I.; Perälä, N.; Peltomaa, R.; Kaiser, M.; Würth, C.; Christ, S.; Resch-Genger, U.; Schäferling, M.; Soukka, T. Nanoscale 2015, 7, 11746−11757.

(4) Kuno, M.; Fromm, D. P.; Hamann, H. F.; Gallagher, A.; Nesbitt, D. J. J. Chem. Phys. 2000, 112, 3117−3120. (5) Michalet, X.; Pinaud, F.; Lacoste, T. D.; Dahan, M.; Bruchez, M. P.; Alivisatos, A. P.; Weiss, S. Single Mol. 2001, 2, 261−276. (6) Auzel, F. Chem. Rev. 2004, 104, 139−173. (7) Haase, M.; Schäfer, H. Angew. Chem., Int. Ed. 2011, 50, 5808− 5829. (8) Zhou, J.; Liu, Q.; Feng, W.; Sun, Y.; Li, F. Chem. Rev. 2015, 115, 395−465. (9) Heer, S.; Kömpe, K.; Güdel, H. U.; Haase, M. Adv. Mater. 2004, 16, 2102−2105. (10) Liu, X. W.; Deng, R. R.; Zhang, Y. H.; Wang, Y.; Chang, H. J.; Huang, L.; Liu, X. G. Chem. Soc. Rev. 2015, 44, 1479−1508. (11) Wu, S. W.; Han, G.; Milliron, D. J.; Aloni, S.; Altoe, V.; Talapin, D. V.; Cohen, B. E.; Schuck, P. J. Proc. Natl. Acad. Sci. U. S. A. 2009, 106, 10917−10921. (12) Nyk, M.; Kumar, R.; Ohulchanskyy, T. Y.; Bergey, E. J.; Prasad, P. N. Nano Lett. 2008, 8, 3834−3838. (13) Wang, M.; Mi, C. C.; Wang, W. X.; Liu, C. H.; Wu, Y. F.; Xu, Z. R.; Mao, C. B.; Xu, S. K. ACS Nano 2009, 3, 1580−1586. (14) Cao, T. Y.; Yang, Y.; Gao, Y. A.; Zhou, J.; Li, Z. Q.; Li, F. Y. Biomaterials 2011, 32, 2959−2968. (15) Mi, C. C.; Zhang, J. P.; Gao, H. Y.; Wu, X. L.; Wang, M.; Wu, Y. F.; Di, Y. Q.; Xu, Z. R.; Mao, C. B.; Xu, S. K. Nanoscale 2010, 2, 1141− 1148. (16) Li, L. L.; Zhang, R. B.; Yin, L. L.; Zheng, K. Z.; Qin, W. P.; Selvin, P. R.; Lu, Y. Angew. Chem., Int. Ed. 2012, 51, 6121−6125. (17) Nam, S. H.; Bae, Y. M.; Park, Y. I.; Kim, J. H.; Kim, H. M.; Choi, J. S.; Lee, K. T.; Hyeon, T.; Suh, Y. D. Angew. Chem., Int. Ed. 2011, 50, 6093−6097. (18) Park, Y. I.; Lee, K. T.; Suh, Y. D.; Hyeon, T. Chem. Soc. Rev. 2015, 44, 1302−1317. (19) Rao, J. H.; Dragulescu-Andrasi, A.; Yao, H. Q. Curr. Opin. Biotechnol. 2007, 18, 17−25. (20) Pinaud, F.; Michalet, X.; Bentolila, L. A.; Tsay, J. M.; Doose, S.; Li, J. J.; Iyer, G.; Weiss, S. Biomaterials 2006, 27, 1679−1687. (21) Heilemann, M.; van de Linde, S.; Schuttpelz, M.; Kasper, R.; Seefeldt, B.; Mukherjee, A.; Tinnefeld, P.; Sauer, M. Angew. Chem., Int. Ed. 2008, 47, 6172−6176. (22) Xiong, L. Q.; Chen, Z. G.; Tian, Q. W.; Cao, T. Y.; Xu, C. J.; Li, F. Y. Anal. Chem. 2009, 81, 8687−8694. (23) Chatterjee, D. K.; Rufaihah, A. J.; Zhang, Y. Biomaterials 2008, 29, 937−943. (24) Xiong, L. Q.; Chen, Z. G.; Yu, M. X.; Li, F. Y.; Liu, C.; Huang, C. H. Biomaterials 2009, 30, 5592−5600. (25) Tian, Z.; Chen, G. Y.; Li, X.; Liang, H. J.; Li, Y. S.; Zhang, Z. G.; Tian, Y. Laser Med. Sci. 2010, 25, 479−484. (26) van de Rijke, F.; Zijlmans, H.; Li, S.; Vail, T.; Raap, A. K.; Niedbala, R. S.; Tanke, H. J. Nat. Biotechnol. 2001, 19, 273−276. (27) Päkkilä, H.; Ylihärsilä, M.; Lahtinen, S.; Hattara, L.; Salminen, N.; Arppe, R.; Lastusaari, M.; Saviranta, P.; Soukka, T. Anal. Chem. 2012, 84, 8628−8634. (28) Ylihärsilä, M.; Valta, T.; Karp, M.; Hattara, L.; Harju, E.; Hölsä, J.; Saviranta, P.; Waris, M.; Soukka, T. Anal. Chem. 2011, 83, 1456− 1461. (29) Corstjens, P. L. A. M.; De Dood, C. J.; Kornelis, D.; Fat, E. M. T. K.; Wilson, R. A.; Kariuki, T. M.; Nyakundi, R. K.; Loverde, P. T.; Abrams, W. R.; Tanke, H. J.; Van Lieshout, L.; Deelder, A. M.; Van Dam, G. J. Parasitology 2014, 141, 1841−1855. (30) Wilhelm, S.; Kaiser, M.; Würth, C.; Heiland, J.; Carrillo-Carrion, C.; Muhr, V.; Wolfbeis, O. S.; Parak, W. J.; Resch-Genger, U.; Hirsch, T. Nanoscale 2015, 7, 1403−1410. (31) Van Blaaderen, A.; Vrij, A. Langmuir 1992, 8, 2921−2931. (32) Hund, F. Z. Anorg. Chem. 1950, 261, 106−115. (33) Hlavácě k, A.; Sedlmeier, A.; Skládal, P.; Gorris, H. H. ACS Appl. Mater. Interfaces 2014, 6, 6930−6935. (34) Corstjens, P. L. A. M.; van Lieshout, L.; Zuiderwijk, M.; Kornelis, D.; Tanke, H. J.; Deelder, A. M.; van Dam, G. J. J. Clin. Microbiol. 2008, 46, 171−176. 1841

DOI: 10.1021/acs.analchem.5b04147 Anal. Chem. 2016, 88, 1835−1841

Suggest Documents