sensors Article
Electron-Beam-Lithographed Nanostructures as Reference Materials for Label-Free Scattered-Light Biosensing of Single Filoviruses Anant Agrawal 1, * 1 2
*
ID
, Joseph Majdi 1 , Kathleen A. Clouse 2 and Tzanko Stantchev 2
Division of Biomedical Physics, Office of Science and Engineering Laboratories, Center for Devices and Radiological Health, Food and Drug Administration, Silver Spring, MD 20993, USA;
[email protected] Division of Biotechnology Review and Research I, Office of Biotechnology Products, Office of Pharmaceutical Quality, Center for Drug Evaluation and Research, Food and Drug Administration, Silver Spring, MD 20993, USA;
[email protected] (K.A.C.);
[email protected] (T.S.) Correspondence:
[email protected]; Tel.: +1-301-796-2496
Received: 22 March 2018; Accepted: 20 May 2018; Published: 23 May 2018
Abstract: Optical biosensors based on scattered-light measurements are being developed for rapid and label-free detection of single virions captured from body fluids. Highly controlled, stable, and non-biohazardous reference materials producing virus-like signals are valuable tools to calibrate, evaluate, and refine the performance of these new optical biosensing methods. To date, spherical polymer nanoparticles have been the only non-biological reference materials employed with scattered-light biosensing techniques. However, pathogens like filoviruses, including the Ebola virus, are far from spherical and their shape strongly affects scattered-light signals. Using electron beam lithography, we fabricated nanostructures resembling individual filamentous virions attached to a biosensing substrate (silicon wafer overlaid with silicon oxide film) and characterized their dimensions with scanning electron and atomic force microscopes. To assess the relevance of these nanostructures, we compared their signals across the visible spectrum to signals recorded from Ebola virus-like particles which exhibit characteristic filamentous morphology. We demonstrate the highly stable nature of our nanostructures and use them to obtain new insights into the relationship between virion dimensions and scattered-light signal. Keywords: optical biosensing; electron beam lithography; label-free; filovirus; reference materials; scattered light; interferometry
1. Introduction Label-free optical biosensors are being developed to fill the need for rapid, inexpensive, and field-deployable detection of specific pathogenic microorganisms like Ebola virus (EBOV), the cause of one of the most severe outbreaks of deadly hemorrhagic fever during the years 2014–2016. In principle, these biosensors accept a small sample of body fluid and the target pathogens present in the sample are directly captured, optically interrogated, and identified with minimal operator interaction. One broad category of optical biosensors uses scattered light to detect and characterize single virions [1–10], which as dielectric nanoparticles yield very weak scattering, but the tiny optical field is measurable and information-rich. This “scattered-light signal” (from here, referred to simply as “signal”) is strongly influenced by the three primary physical properties of a dielectric nanoparticle—shape, size, and refractive index—all of which facilitate agile pathogen identification. At various stages of the life cycle of a biosensor—from design to deployment—reference materials are needed to compare prototype sensor configurations, calibrate sensor parameters, verify
Sensors 2018, 18, 1670; doi:10.3390/s18061670
www.mdpi.com/journal/sensors
Sensors 2018, 18, 1670
2 of 12
sensor quality at manufacture, and ensure consistent sensor performance over time in the field. For single-nanoparticle optical biosensing strategies, spherical polymer nanoparticles (i.e., polystyrene or silica) have been the biohazard-free reference material of choice to calibrate and evaluate biosensor performance [1,3,6,7,10]. These nanoparticles roughly match the refractive index of the target microorganism and are commercially available in the relevant sizes. However, when the target is non-spherical in nature, as is the case with filoviruses, reference materials with similar nanoscale morphology can greatly benefit biosensor calibration, evaluation, and refinement. In this work, we have fabricated and characterized a new type of reference material realized as nanostructures closely resembling the morphology and refractive index of filoviruses, namely EBOV. Fabrication was achieved with electron beam lithography (EBL) of hydrogen silsesquioxane (HSQ) on a substrate specific to a promising scattered-light biosensing scheme based on interferometric reflectance imaging [1–4], though the nanostructures are readily fabricated on other substrates as well. Upon fabrication, we measured the nanostructures’ dimensions with scanning electron and atomic force microscopes for subsequent correlation to signal. We assessed the similarity of the nanostructures’ signals to those from EBOV virus-like particles (VLPs), which have morphology resembling that of the wild-type EBOV [11,12]. We observed notable consistency of the signals over a 9-month period (coefficient of variation (CV) < 2%) from a collection of nanostructures on a chip stored simply in a Petri dish on a lab bench. Finally, we used the nanostructures to reveal the polarization-dependent signal from asymmetric objects, a characteristic which can enhance detection of EBOV virions and other asymmetric microorganisms. 2. Materials and Methods 2.1. Key Virion Characteristics: Shape, Size, and Refractive Index Filoviruses such as EBOV are long and tubular in shape with length ranging from hundreds to thousands of nanometers. The shortest virions generally appear as straight line segments, occasionally with a bulbous end, while longer virions more likely have curves. Recent electron microscopy studies have reported 96–99 nm diameter for EBOV virions either frozen-hydrated [13] or treated with ionic liquid [14]. The EBOV genome size of 18.9 kilobases suggests a theoretical virion length of 980 nm, which is consistent with cryo-electron microscopy findings that most virus particles have a length of 982 ± 79 nm, though significantly longer polyploid virions were also observed [13]. We assume a filovirus refractive index of 1.42, based on a recent report of the first-ever experimental refractive index measurement of individual animal virions (HIV-1) [15]. 2.2. Electron Beam Lithography (EBL) EBL is a direct-write process to create structures with nanoscale detail by scanning a focused electron beam in a precise pattern over a surface coated with a thin film of electron-sensitive material (resist) [16]. Immersion of the film in the appropriate developer solution after electron beam exposure leads to removal of either the exposed or unexposed resist from the surface, depending on whether the resist is positive-tone or negative-tone, respectively. The EBL system we used in this work comprised the Nanometer Pattern Generation System (JC Nabity Lithography Systems, Bozeman, MT, USA) coupled to a scanning electron microscope (SEM, JSM-6400, JEOL Ltd., Tokyo, Japan). This system accepts computer-aided design (CAD) drawings of the patterns to be traced with the electron beam on the resist. We drew patterns of individual virions represented as simple rectangles of various lengths and widths, with large separation between the rectangles to easily isolate the signal from each. We chose HSQ as the resist with which to fabricate our nanostructures, in part because HSQ is negative-tone and therefore more efficient than positive-tone for fabricating the spread-out collection of nanostructures we desired. Furthermore, from reported ellipsometry measurements of HSQ films [17] and our own Abbe refractometer measurements of the liquid resist, HSQ refractive index is 1.39 ± 0.01, closely matching the assumed virion index of 1.42. Our specific EBL process is illustrated in Figure 1.
Sensors 2018, 18, x FOR PEER REVIEW Sensors 2018, 18, 1670
3 of 12 3 of 12
1.39 ± 0.01, closely matching the assumed virion index of 1.42. Our specific EBL process is illustrated in 1. We HSQ spin coated HSQ (XR-1541-002 -004, Dow Corning Corp., onto WeFigure spin coated (XR-1541-002 or -004, Dow or Corning Corp., Midland, MI, Midland, USA) ontoMI, theUSA) biosensing ◦ C for the biosensing substrate in the next section), bake on aathot at 150 °C substrate (described in (described the next section), followed by followed soft bakeby onsoft a hot plate 150plate 5 min for 5 min to evaporate solvent. To prevent troublesome charge accumulation during EBL, we spin to evaporate solvent. To prevent troublesome charge accumulation during EBL, we spin coated a coated a water-soluble conductive film (aquaSAVE, Mitsubishi Chemical water-soluble conductive film (aquaSAVE, Mitsubishi Chemical Corp., Corp., Tokyo,Tokyo, Japan)Japan) on topon oftop the of the film, HSQfollowed film, followed soft at bake for 2tomin to evaporate solvent. The CAD patterns HSQ by softbybake 100at◦ C100 for°C 2 min evaporate solvent. The CAD patterns were were written the chip following system parameters, which settledupon uponafter aftera then then written ontoonto the chip withwith the the following system parameters, which wewe settled anumber numberofoftrials: trials:30 30kV kVacceleration accelerationvoltage, voltage,50 50µm µmaperture, aperture,16 16mm mmworking working distance, distance, 4625 4625 µC/cm µC/cm22 energy dose. After Afterelectron electronbeam beam exposure, briefly immersed in distilled water to energy dose. exposure, thethe chipchip waswas briefly immersed in distilled water to remove remove the aquaSAVE film and thenwith driedawith a purified N2 stream. Finally, the chip immersed the aquaSAVE film and then dried purified N2 stream. Finally, the chip was was immersed in a in a developer solution of tetramethylammonium 25% tetramethylammonium hydroxide (TMAH) in (331635, water (331635, Sigmadeveloper solution of 25% hydroxide (TMAH) in water Sigma-Aldrich Aldrich St.MO, Louis, MO,for USA) s to remove unexposed with distilled and Inc., St. Inc., Louis, USA) 23 sfor to 23 remove unexposed resist, resist, rinsedrinsed with distilled waterwater and dried dried with with N 2 . N2. We We imaged imaged the the fabricated fabricated nanostructures nanostructures on on the the chip chip with with field field emission emission SEM SEM (SU-70, (SU-70, Hitachi Hitachi High HighTechnologies Technologies America, America, Inc., Inc., Schaumburg, Schaumburg, IL, IL, USA) USA) to to measure measure their their length length and and width, width, and and we we used usedan anatomic atomicforce force microscope microscope (AFM, (AFM, MFP-3D, MFP-3D, Asylum Asylum Research Research Inc., Inc., Santa Santa Barbara, Barbara, CA, CA, USA) USA) to to measure thickness. SEM images were analyzed with ImageJ [18] and MATLAB (Mathworks, Natick, measure thickness. SEM images were analyzed with ImageJ [18] and MATLAB (Mathworks, Natick, MA, MA,USA), USA),while whileAFM AFMimages imageswere wereanalyzed analyzedwith withIgor IgorPro Pro(Wavemetrics, (Wavemetrics,Lake LakeOswego, Oswego,OR, OR,USA) USA) integrated integrated into into that that microscope’s microscope’s software software interface. interface. aquaSAVE
HSQ
SiO2
Silicon
Spin coat
Optical biosensing substrate
Spin coat
Conductive film applied to resist
Resist applied to substrate
e-beam
H2O
TMAH
Unexposed resist dissolved
Conductive film dissolved
Patterned e-beam exposure
Figure Figure1. 1.Electron Electronbeam beamlithography lithographyprocess. process.Dimensions Dimensionsare arenot notto toscale. scale.
2.3. 2.3. Scattered-Light Scattered-LightBiosensing Biosensing with with Interferometric Interferometric Reflectance Reflectance Imaging Imaging Interferometric sensing enables enablesdetection detectionofof astoundingly small optical fields, by mixing the Interferometric sensing astoundingly small optical fields, by mixing the small small fielda with muchreference larger reference field. the weaklylight reflected from nanoparticle a spherical field with muchalarger field. When theWhen weakly reflected from alight spherical nanoparticle interferes with reference light from a strong reflector, the detected optical intensity (I) interferes with reference light from a strong reflector, the detected optical intensity (I) relates to particle 3 3 relates I ∝ r , which is a dramatically sensitivity to the smallest radius to (r) particle as I ∝ r radius , which(r) is aasdramatically increased sensitivityincreased to the smallest particles over direct particles detection of a nanoparticle’s light, for whichisthe is illustrated I ∝ r6 [1,19]. detectionover of adirect nanoparticle’s reflected light, for reflected which the relationship I ∝relationship r6 [1,19]. As in As illustrated in Figure 2a, this sensing arrangement can be realized as a common-path interferometer Figure 2a, this sensing arrangement can be realized as a common-path interferometer by placing a by placing a on nanoparticle on top substrate, of a two-layer substrate, with top layertothickness similar to the nanoparticle top of a two-layer with top layer thickness similar the particle size and the particle size and the bottom layer effectively semi-infinite, and illuminating from above. Light bottom layer effectively semi-infinite, and illuminating from above. Light reflecting from the particle reflecting from the particle measurably interferes with the reflected lightbetween from the measurably interferes with the reflected light from the underlying interface theunderlying substrate’s interface between the serving substrate’s topreference. and bottom layers serving as the on reference. In addition to top and bottom layers as the In addition to dependence particle size, the nature dependence on particle size, the nature (constructive or destructive) and magnitude of the interference (constructive or destructive) and magnitude of the interference depends on the particle’s refractive depends the particle’s refractive and index, illumination wavelength, the thickness of the index, theonillumination wavelength, the the thickness of the substrate’s topand layer. substrate’s Basedtop on layer. the groundbreaking work from the Ünlü research group [1–4], we implemented an Based on the groundbreaking work from Ünlü research group [1–4],microscope we implemented an interferometric reflectance imaging biosensor as the an epi-illumination brightfield (Figure 2b). interferometric reflectance biosensor as an epi-illumination brightfield (Figure 2b). A multi-LED light sourceimaging (WLS-22-A, Mightex Systems, Pleasanton, CA, microscope USA) is fiber-optically A multi-LED source (WLS-22-A, Mightex Pleasanton, CA, USA) fiber-optically coupled to anlight achromatic relay lens pair whoseSystems, focal plane is aligned with the is back focal plane coupled to an achromatic relay lens pair whose focal plane is aligned with the back focal plane thus of a of a 50×/0.8 NA infinity-corrected microscope objective (LU Plan EPI, Nikon, Tokyo, Japan), 50×/0.8 NA infinity-corrected microscope objective (LU Plan EPI, Nikon, Tokyo, Japan), thus
Sensors 2018, 18, 1670
4 of 12
Sensors 2018, 18, x FOR PEER REVIEW
4 of 12
providing illumination onto onto the thebiosensing biosensingsubstrate. substrate.This Thisobjective objectiveis is designed used providing Köhler Köhler illumination designed to to bebe used in in air without a coverslip at 1 mm working distance. A 50/50 cube beamsplitter reflects light from air without a coverslip at 1 mm working distance. A 50/50 cube beamsplitter reflects light from the the relay lenses to the objective, reflecting the sample returns through the objective relay lenses to the objective, andand lightlight reflecting fromfrom the sample returns through the objective and and passes through the beamsplitter, to an achromatic tube lens which images onto a monochrome passes through the beamsplitter, to an achromatic tube lens which images onto a monochrome CMOS CMOS (PL-B771U, Pixelink, ON, Canada). A rotatable linear polarizer was between inserted cameracamera (PL-B771U, Pixelink, Ottawa,Ottawa, ON, Canada). A rotatable linear polarizer was inserted between the beamsplitter andlens tubefor lens for some of our imaging, allowing isolationofof any any linear the beamsplitter and tube some of our imaging, allowing forforisolation linear polarization state at the camera. The two-layer substrate is a silicon wafer coated with a 100 polarization state at the camera. The two-layer substrate is a silicon wafer coated with a 100 nm-thick nm-thick thermal thermal silicon silicon oxide oxide (SiO (SiO22)) film film (Silicon (Silicon Valley Valley Microelectronics, Microelectronics,Santa SantaClara, Clara,CA, CA,USA). USA). With target objects on the substrate in focus within the microscope’s field With target objects on the substrate in focus within the microscope’s field of of view, view, we we captured captured images sequentially with eight different narrowband LEDs whose nominal central wavelengths images sequentially with eight different narrowband LEDs whose nominal central wavelengths are are 400, 455, 470, 505, 530, 590, 625, 656 nm. We also captured background images (no LED illumination) 400, 455, 470, 505, 530, 590, 625, 656 nm. We also captured background images (no LED illumination) to to account account for for ambient ambient light light and and camera camera digital digital offset offset contributions contributions to to the the images images at at each each of of the the LED LED wavelengths. We then used routines written in MATLAB with the acquired images to generate wavelengths. We then used routines written in MATLAB with the acquired images to generate an an 8-point spectrum (signal versus LED wavelength) for each object on the substrate. 8-point spectrum (signal versus LED wavelength) for each object on the substrate.
(a)
(b)
Figure 2. 2. Interferometric Interferometric reflectance reflectance imaging imaging for for biosensing. biosensing. (a) principle. Figure (a) Illustration Illustration of of the the basic basic principle. The solid green line approximates a sample path of light waves reflected off the nanoparticle, and the the The solid green line approximates a sample path of light waves reflected off the nanoparticle, and dashed red line approximates a reference path reflection from the interface between the two substrate dashed red line approximates a reference path reflection from the interface between the two substrate layers. Dimensions Dimensions are are not not to to scale. scale. (b) (b) Photograph Photograph of of our our laboratory laboratory biosensor biosensor setup. setup. layers.
2.4. EBOV EBOV Virus-Like Particles (VLPs) (VLPs) 2.4. Virus-Like Particles Cell fusion-capable fusion-capable but but replication-incompetent replication-incompetent EBOV EBOV surface surface glycoprotein glycoprotein (GP)-pseudotyped (GP)-pseudotyped Cell VLPs have emerged as a viable alternative to study filoviruses under biosafety level 22 (BSL-2) (BSL-2) VLPs have emerged as a viable alternative to study filoviruses under biosafety level conditions [20,21]. The VLPs used in the current study were produced by co-transfecting 293T cells conditions [20,21]. The VLPs used in the current study were produced by co-transfecting 293T with a plasmid encoding the EBOV GP strain Kikwit (VRC 6001), a plasmid encoding EBOV matrix cells with a plasmid encoding the EBOV GP strain Kikwit (VRC 6001), a plasmid encoding EBOV protein VP40 (pCAGG VP40), and a plasmid encoding VP40 linked to green fluorescent protein matrix protein VP40 (pCAGG VP40), and a plasmid encoding VP40 linked to green fluorescent protein (VP40-GFP). 12–16 h after co-transfection the 293T cells were gently washed and incubated in fresh (VP40-GFP). 12–16 h after co-transfection the 293T cells were gently washed and incubated in fresh Dulbecco’s Modified Eagle’s Medium supplemented with 10% fetal calf serum, L-glutamine and nonDulbecco’s Modified Eagle’s Medium supplemented with 10% fetal calf serum, L-glutamine and essential amino acids. The VLP-containing cell culture medium was collected after 24 and 48 h, non-essential amino acids. The VLP-containing cell culture medium was collected after 24 and 48 h, cleared from cell debris by low speed centrifugation and passed through a 0.45 µm low protein cleared from cell debris by low speed centrifugation and passed through a 0.45 µm low protein binding filter. The VLPs were further concentrated by ultracentrifugation through a 20% sucrose binding filter. The VLPs were further concentrated by ultracentrifugation through a 20% sucrose cushion as previously described [12]. We then removed the sucrose from the VLP suspension via cushion as previously described [12]. We then removed the sucrose from the VLP suspension via microdialysis with an aqueous 0.5% NaCl solution as the dialysate. It was previously demonstrated microdialysis with an aqueous 0.5% NaCl solution as the dialysate. It was previously demonstrated that VLPs generated using VP40 display the characteristic filamentous morphology [11,12], and we that VLPs generated using VP40 display the characteristic filamentous morphology [11,12], and we confirmed this morphology via fluorescence microscopy with a blue excitation/green emission filter set to visualize GFP (Figure 3).
Sensors 2018, 18, 1670
5 of 12
confirmed this morphology via fluorescence microscopy with a blue excitation/green emission filter Sensors 2018, 18, x FOR PEER REVIEW 5 of 12 set to visualize GFP (Figure 3). For For biosensing biosensing experiments, experiments, aa 0.5 0.5 µL µL droplet droplet of of the the dialyzed dialyzed VLP VLP suspension suspension was was pipetted pipetted onto onto aa clean clean biosensing biosensingsubstrate substrateand andplaced placed into a humidity chamber for 1 h to allow VLPs to settle. into a humidity chamber for 1 h to allow VLPs to settle. Upon Upon removal from the humidity chamber, the excess moisture the droplet was carefully absorbed removal from the humidity chamber, the excess moisture of theofdroplet was carefully absorbed with with a laboratory wipe. We identified isolated VLPs bound to the substrate surface with fluorescence a laboratory wipe. We identified isolated VLPs bound to the substrate surface with fluorescence microscopy microscopy prior prior to to biosensing biosensing measurements. measurements.
Figure 3. Fluorescence microscopy of EBOV VLPs, demonstrating filamentous morphology with Figure 3. Fluorescence microscopy of EBOV VLPs, demonstrating filamentous morphology with varying lengths. Scale bar is 5 µm. varying lengths. Scale bar is 5 µm.
3. Results 3. Results 3.1. Nanostructure Dimensions 3.1. Nanostructure Dimensions We fabricated a total of 48 nanostructures, comprised of eight groups of six different lengths. We fabricated a total of 48 nanostructures, comprised of eight groups of six different lengths. The eight groups arise from having duplicate sets of two different widths and two different The eight groups arise from having duplicate sets of two different widths and two different orientations orientations (vertical and horizontal). Thickness was defined primarily by the HSQ resist spin coat (vertical and horizontal). Thickness was defined primarily by the HSQ resist spin coat process process and electron beam dose [22]. Figure 4 shows representative SEM images of the and electron beam dose [22]. Figure 4 shows representative SEM images of the nanostructures. nanostructures. Though designed as rectangles, the final shapes have rounded ends due to the Though designed as rectangles, the final shapes have rounded ends due to the expected blurring effect expected blurring effect of the electron beam’s point spread function. Rough edges in all structures of the electron beam’s point spread function. Rough edges in all structures result from randomness result from randomness in electron and developer interactions with the resist. More gross shape in electron and developer interactions with the resist. More gross shape errors were apparent in errors were apparent in some structures (e.g., 500 × 40 nm and 1000 × 40 nm structures shown in some structures (e.g., 500 × 40 nm and 1000 × 40 nm structures shown in Figure 4) due to occasional Figure 4) due to occasional irregularities in the EBL system operation. All of these deviations from irregularities in the EBL system operation. All of these deviations from the ideal rectangular shape the ideal rectangular shape were considered negligible given their size relative to the overall structure were considered negligible given their size relative to the overall structure size and to the wavelengths size and to the wavelengths of light for biosensing. of light for biosensing. Table 1 summarizes the dimensions of all the fabricated nanostructures. Measured values Table 1 summarizes the dimensions of all the fabricated nanostructures. Measured values represent the mean ± one standard deviation, with the CV shown in parentheses. The SEM image of represent the mean ± one standard deviation, with the CV shown in parentheses. The SEM image one of the 2000 nm-long structures was corrupted, so we took the mean length from only three of one of the 2000 nm-long structures was corrupted, so we took the mean length from only three structures instead of four. Overall, we achieved excellent consistency in length, with CV generally structures instead of four. Overall, we achieved excellent consistency in length, with CV generally shrinking as structure length increased; width and thickness were slightly less consistent. Measured shrinking as structure length increased; width and thickness were slightly less consistent. Measured length was always slightly larger than design length, while measured width was considerably larger length was always slightly larger than design length, while measured width was considerably larger than designed. We attribute these results primarily to the same blurring effect which rounded the than designed. We attribute these results primarily to the same blurring effect which rounded the ends ends of the structures. Uncertainty in the spin coat process and electron beam-HSQ interaction gave of the structures. Uncertainty in the spin coat process and electron beam-HSQ interaction gave rise to rise to the thickness error, which we consider to be reasonably small. the thickness error, which we consider to be reasonably small.
Sensors 2018, 18, 1670 Sensors 2018, 18, x FOR PEER REVIEW
6 of 12 6 of 12
Figure 4. Scanning electron microscope (SEM) images of six fabricated nanostructures. Labeled lengths Figure 4. Scanning electron microscope (SEM) images of six fabricated nanostructures. Labeled lengths and widths are design values. Scale bars vary for each image as shown. and widths are design values. Scale bars vary for each image as shown. Table 1. Dimensions of fabricated nanostructures. Measured value is the mean ± one standard deviation, Table 1. Dimensions of fabricated nanostructures. Measured value is the mean ± one standard with the coefficient of variation (CV) shown in parentheses. deviation, with the coefficient of variation (CV) shown in parentheses. % Error Number of Design Nanostructure Measurement Design Value Number of Structures Measured Value (Meas. % Error (Meas. Measured Value Nanostructure Measurement Structures Dimension Method (nm) Value Measured (nm) vs. Design) Dimension Method (nm) vs. (nm) Measured 500 4 505 ± 19 (3.8%) Design)1.1% 800 4 811 ± 11 (1.3%) 1.4% 500 5051026 ± 19±(3.8%) 1000 44 18 (1.8%) 1.1% 2.6% Length SEM 800 8111539 ± 11±(1.3%) 1500 44 20 (1.3%) 1.4% 2.6% 2000 34 24 (1.2%) 2.6% 3.1% 1000 10262062 ± 18±(1.8%) Length SEM 3000 4 32 (1.0%) 1500 4 15393084 ± 20±(1.3%) 2.6% 2.8% 40 12 61 ± 3 (5.2%) 2000 3 2062 ± 24 (1.2%) 3.1% 53% SEM Width 80 12 102 ± 5 (5.3%) 3000 4 3084 ± 32 (1.0%) 2.8% 27% Thickness AFM 80 24 87 ± 5 (5.6%) 40 12 61 ± 3 (5.2%) 53% 8.8% Width SEM 80 12 102 ± 5 (5.3%) 27% 80 24 87 ± 5 (5.6%) 8.8% Thickness AFM 3.2. Unpolarized Signals from Nanostructures
Figure 5 shows a representative biosensor image of our 48 nanostructures with 530 nm 3.2. Unpolarized Signals from Nanostructures illumination. Additional structures with other lengths and widths were included in the EBL pattern Figure 5 orientation shows a representative biosensor image of our 48 nanostructures with 530 to assist with and positioning. Since SEM or AFM could alter a nanostructure and nm the illumination. Additional structures other lengths and widths included in the EBL pattern adjacent surface, as is evident from with the brightness variations of thewere left group of vertical 61 nm-wide to assist with orientation Since SEM or measurements AFM could alter a nanostructure and the nanostructures, one set ofand 24 positioning. was used for SEM/AFM and the other set was left adjacent surface, as is evident the brightness variations of the group of vertical wavelengths 61 nm-wide untouched for optical imaging.from For our biosensor setup with NA of left 0.8 and illumination nanostructures, one set nm, of 24the was used for SEM/AFM measurements and thedisk other set was left ranging from 400 to 656 diffraction-limited optical resolution (i.e., Airy width) ranged untouched imaging. For our biosensor with NA of 0.8but andwell illumination wavelengths from 260 tofor 420optical nm, below the minimum length ofsetup the nanostructures above their maximum ranging from 400 to from 656 nm, the 5, diffraction-limited optical resolution Airy disk width) ranged width. Accordingly, Figure there are visual differences in length(i.e., but all nanostructures appear from 260 to 420 nm, below the minimum length of the nanostructures but well above their maximum to have the same width. The two different widths instead exhibit a clear brightness difference between width. from wavelength. Figure 5, there are visual differences in length but all nanostructures appear them atAccordingly, this illumination to have the same width. The two different widths instead exhibit a clear brightness difference between them at this illumination wavelength.
Sensors 2018, 18, 1670 Sensors 2018, 18, x FOR PEER REVIEW
7 of 12 7 of 12
Figure 5. 5. Scattered-light Scattered-light image image of of nanostructures nanostructures on on biosensing biosensing substrate substrate with with 530 530 nm nm illumination. illumination. Figure The six six different differentdesign designlengths lengths(in (innm) nm)are arelabeled labeledin inwhite whiteacross acrosstwo twoofofthe theeight eightgroups. groups. The
To generate a biosensor spectrum, each LED-illuminated image was first background To generate a biosensor spectrum, each LED-illuminated image was first background subtracted, subtracted, then regions of interest (ROIs) were manually selected in each image enclosing individual then regions of interest (ROIs) were manually selected in each image enclosing individual objects objects and one blank region. The final signal intensity (B) for an object at a particular wavelength (λ) and one blank region. The final signal intensity (B) for an object at a particular wavelength (λ) was was calculated as: calculated as: Imax (λ) B(λB)(λ =) = I max (λ )−−1,1 , (1) (1) Iblank (λ)
I blank (λ )
where Imax is the mean of the image intensities from the 25 brightest pixels in the object’s ROI and Imax is the mean of the image intensities from the 25 brightest pixels in the object’s ROI and Iblank Iwhere blank is the mean image intensity of the blank ROI. Normalizing by the blank intensity accounts for is the mean image intensity the blank ROI. Normalizing by the blank intensity accounts forwith the the illumination level at eachofwavelength. If the object ROI were also blank, then ideally (i.e., illumination level at each wavelength. If the object ROI were also blank, then ideally (i.e., with zero zero noise and spatially uniform illumination) Imax /Iblank would be unity, so subtracting one sets noise and spatially uniform illumination) /Iblank would so subtracting onefrom sets the the baseline the baseline value of B to zero. Figure 6Imax shows graphsbe ofunity, the biosensor spectra vertical value of B to zero. Figure 6 shows graphs of the biosensor spectra from the vertical nanostructures at nanostructures at three different time points: two weeks after fabrication, then 5 and 9 months after three different time points: two weeks after fabrication, then 5 and 9 months after fabrication. In fabrication. In between imaging sessions, the chip was stored in a Petri dish at room temperature on a between imaging sessions,tothe chip was stored Petri dish at nm room temperature on a laboratory laboratory bench exposed room lights. Except in foravalues at 400 where the structures have very bench exposed to room lights. Except for values at 400 nm where the structures have spectrum, very low low reflectivity, intensities clearly increase with structure length and width across the entire reflectivity, intensities clearly increase with structure length and width across the entire spectrum, as as expected, since a larger object generally tends to scatter more light. Apparent changes in the spectra expected, since larger object generally tendsmaterial to scatter more light. Apparent changes in the spectra over time can beaattributed to HSQ/substrate property changes (e.g., due to hygroscopicity), over time can be attributed to HSQ/substrate material property changes (e.g., due to hygroscopicity), operator focusing variability across imaging sessions, and any long-term drift in our optical imaging operator variability across imaging andisany long-term drift in across our optical imaging setup. Atfocusing any given timepoint, the shape of sessions, the spectra markedly consistent the different setup. At any lengths given timepoint, the shape of the spectra is markedly consistent across the different nanostructure and widths. nanostructure lengths and widths. The relationship between signal at 530 nm and nanostructure size shown in Figure 7 exemplifies The relationship between signal at 530 nm andeffect nanostructure size shown in Figure 7 exemplifies that structure width has a much more profound on the signal than length at filovirus size that structure width has a much more profound effect on the signal than length at filovirus size length scales: scales: a less than two-fold increase in width roughly doubles the signal intensity, but a six-fold a less than two-fold increase inthe width roughly doubles theinsignal intensity, but a six-fold length increase yields less than double signal. The uncertainty 530 nm signal intensity, represented increase yields less than double signal. The uncertainty in 530shows nm signal intensity, represented as as the standard deviation acrossthe 9 months of imaging sessions, no dependence on structure the standard deviation across 9 months of imaging sessions, shows no dependence on structure size. size. We observed similar results at the other seven wavelengths. Intensity CV over 9 months for each We observed similar results ranged at the other seven wavelengths. structure at each wavelength between 0.3% and 1.8%. Intensity CV over 9 months for each structure at each wavelength ranged between 0.3% and 1.8%.
Sensors 2018, 18, 1670 Sensors 2018, 18, x FOR PEER REVIEW Sensors 2018, 18, x FOR PEER REVIEW
8 of 12 8 of 12 8 of 12
Figure spectra measured from the vertical nanostructures at points in 6.Biosensor Biosensor spectra measured from vertical nanostructures at different points in Each time. Figure 6. 6. Biosensor spectra measured from thethe vertical nanostructures at different different points in time. time. Each spectrum consists of only eight measured points; dotted lines are to help guide the eye. Each spectrum consists of eight only eight measured points; dotted to guide help guide the eye. spectrum consists of only measured points; dotted lines lines are toare help the eye.
0.7 0.7 0.6 0.6
102 102 nm nm width width
0.5 0.5 0.4 0.4
61 61 nm nm width width
0.3 0.3 0.2 0.2 500 500
1000 1000
1500 2000 2500 1500 2000 2500 Nanostructure length (nm) Nanostructure length (nm)
3000 3000
Figure 7. intensity at 530 nm for the six different lengths and two different widths of Figure 7. Signal Signal intensity nm for different lengths and two different widths of Figure 7. Signal intensity at at 530 530bars nmrepresent for the the six six different lengths and of two different widths of nanostructures. Horizontal error one standard deviation the measured length; nanostructures. Horizontal error bars represent one standard deviation of the measured length; nanostructures. Horizontal error bars represent one standard deviation of the measured length; vertical error bars represent one standard deviation of the across 99 months of vertical errorbars barsrepresent represent standard deviation ofintensity the intensity intensity of imaging imaging vertical error oneone standard deviation of the acrossacross 9 monthsmonths of imaging sessions. sessions. Dotted lines are to help guide the eye. sessions. Dotted lines are to help guide the eye. Dotted lines are to help guide the eye.
3.3. 3.3. Nanostructures Nanostructures versus versus VLPs VLPs 3.3. Nanostructures versus VLPs Our Our simple simple protocol protocol to to deposit deposit VLPs VLPs on on the the biosensor biosensor substrate substrate was was successful successful but but inefficient. inefficient. Our simple protocol to deposit VLPs ontothe biosensor substrate was successful but inefficient. An isolated and unobstructed VLP bound the substrate surface and positively An isolated and unobstructed VLP bound to the substrate surface and positively confirmed confirmed by by An isolated and unobstructed VLP boundFigure to the substrate surface and positively confirmed by fluorescence fluorescence microscopy microscopy was was aa rare rare event. event. Figure 88 shows shows the the representative representative signals signals from from isolated isolated fluorescence microscopy was a rare event.fluorescence Figure 8 shows the representative signals from isolated VLPs VLPs of of different different lengths lengths (measured (measured with with fluorescence microscopy) microscopy) resulting resulting from from three three deposition deposition VLPs of different lengths (measured with fluorescence microscopy) resulting from three deposition sessions. sessions. Their Their overall overall intensities intensities vary vary substantially substantially without without any any clear clear dependence dependence on on length, length, though though all four spectra share a similar shape with a slightly blue-shifted peak compared to the all four spectra share a similar shape with a slightly blue-shifted peak compared to the nanostructure nanostructure spectra. spectra. Since Since we we observed observed spectral spectral shape shape and and peak peak location location to to be be consistent consistent with with nanostructure nanostructure
Sensors 2018, 18, 1670
9 of 12
sessions. Their overall intensities vary substantially without any clear dependence on length, though all four2018, spectra share a similar Sensors 18, x FOR PEER REVIEWshape with a slightly blue-shifted peak compared to the nanostructure 9 of 12 spectra. Since we observed spectral shape and peak location to be consistent with nanostructure length and and width, width, the the only only two two remaining remaining variables variables which which could could explain explain the the spectral spectral differences differences length between nanostructure and thickness and refractive index, whose product is theiswellbetween andVLP VLPsignals signalsare are thickness and refractive index, whose product the known physical parameter of optical thickness. This explanation is supported by thin film well-known physical parameter of optical thickness. This explanation is supported byreflectance thin film calculated calculated with the Fresnel equations [23], which predict spectral shift with changing optical reflectance with the Fresnel equations [23], whichapredict a spectral shift with changing thickness of a thin to theto SiO 2-Si substrate because of changing levels of constructive and optical thickness of film a thinapplied film applied the SiO because of changing levels of constructive 2 -Si substrate destructive interference (inset(inset of Figure 8). Though the Fresnel-calculated reflectance spectraspectra strictly and destructive interference of Figure 8). Though the Fresnel-calculated reflectance apply only to only filmsto with effectively infinite length width, qualitatively corroborate the VLPstrictly apply films with effectively infiniteand length andthey width, they qualitatively corroborate nanostructure spectral spectral shift when similar to optical are applied in the Fresnel the VLP-nanostructure shift when changes similar changes to thickness optical thickness are applied in the equations. A refractive index difference of 1.42of(virus) − 1.39−(HSQ) = 0.03= is expected, but but this Fresnel equations. A refractive index difference 1.42 (virus) 1.39 (HSQ) 0.03 is expected, difference could be be greater since wewe measured water this difference could greater since measureddried driedVLPs, VLPs,devoid devoidof of genetic genetic material and water presentin inthe theviruses virusescharacterized characterizedby byPang Panget etal. al.[15]. [15]. Instead Instead of of falling falling in in the the intensity intensity range range of of our our present wider(102 (102nm) nm)nanostructures nanostructureswhose whose width close that of the EBOV virion, intensities wider width is is close to to that of the EBOV virion, the the VLPVLP intensities are are more similar to those fromnarrower our narrower (61 nm) nanostructures, likely because dried VLPs more similar to those from our (61 nm) nanostructures, likely because dried VLPs shrunk shrunk from theirsize native size in aqueous suspension. Therefore, the VLP thickness could also be less from their native in aqueous suspension. Therefore, the VLP thickness could also be less than the than 87 nm nanostructure 87 nmthe nanostructure thickness.thickness. As a structural wild-type EBOV, thethe VLPs can can provide onlyonly a general sense As structuralapproximation approximationofofthe the wild-type EBOV, VLPs provide a general of theofsimilarity in signal from nanostructures versus actual VLP sense the similarity in signal from nanostructures versus actualfilovirus filovirusspecimens. specimens.Even Even ifif a VLP produces the identical signal as a filovirus virion, we do not consider the observed spectral shift to produces the identical signal as a filovirus virion, we do not consider the observed spectral shift to be be large enough to disqualify nanostructures as reference materials, which meant provide large enough to disqualify thethe nanostructures as reference materials, which areare meant to to provide a a controlled, stable source of signal for a biosensor resembling a real specimen’s signal but not controlled, stable source of signal for a biosensor resembling a real specimen’s signal but not necessarily necessarily identical to it. identical to it.
Figure8.8.Biosensor Biosensor spectra spectrafrom fromvirus-like virus-likeparticles particles(VLPs) (VLPs)and andsimilarly-sized similarly-sizednanostructure. nanostructure. VLP VLP Figure lengths and and nanostructure nanostructure length length and and width width are are indicated indicated in in the the legend. legend. Vertical Vertical error error bars bars on on the the lengths nanostructure plot represent one standard deviation of the intensity across 9 months of imaging nanostructure plot represent one standard deviation of the intensity across 9 months of imaging sessions. Dotted Dotted lines lines are are to to help help guide guide the theeye. eye. Inset: Inset: Reflectance Reflectance spectra spectra calculated calculated with with Fresnel Fresnel sessions. 2-Si substrate. equations for thin films of different optical thickness on SiO equations for thin films of different optical thickness on SiO -Si substrate. 2
3.4. Polarization Dependence of Signal We investigated polarization effects on the signal by capturing images of our nanostructures with the linear polarizer inserted in the light collection path and rotated to image two polarization states: vertical (i.e., aligned parallel with our vertical nanostructures) and horizontal. Figures 9a,b show images at 530 nm from the two orthogonal polarizations, revealing a distinct intensity dependence with polarization state. The structures aligned parallel to the polarization are substantially brighter
Sensors 2018, 18, 1670
10 of 12
3.4. Polarization Dependence of Signal We investigated polarization effects on the signal by capturing images of our nanostructures with the linear polarizer inserted in the light collection path and rotated to image two polarization states: vertical (i.e., aligned parallel with our vertical nanostructures) and horizontal. Figure 9a,b show images at 530 nm from the two orthogonal polarizations, revealing a distinct intensity Sensors 2018, 18, x FOR PEER REVIEW 10 of 12 dependence with polarization state. The structures aligned parallel to the polarization are substantially brighter their perpendicular counterparts, a relationship which holdsvertical for both vertical and than theirthan perpendicular counterparts, a relationship which holds for both and horizontal horizontal polarizations. Figure 9c plots the polarization ratio (vertically-polarized B(λ) divided by polarizations. Figure 9c plots the polarization ratio (vertically-polarized B(λ) divided by horizontallyhorizontally-polarized B(λ)) for the vertical nanostructures at 530 nm. Unlike the original signal, polarized B(λ)) for the vertical nanostructures at 530 nm. Unlike the original signal, the polarization the polarization is larger structures for the narrower structures andmonotonic there is nolength clear dependence. monotonic length ratio is larger forratio the narrower and there is no clear Even dependence. Even the shortest structures, 1000 nm and below, whose filamentous structure the shortest structures, 1000 nm and below, whose filamentous structure becomes less becomes visually less visually apparent, exhibit strong polarization 1.45 orTherefore, greater. Therefore, polarization apparent, exhibit strong polarization sensitivity sensitivity of 1.45 or of greater. polarization sensing sensing provides complementary information to the unpolarized signal intensity with respect to the provides complementary information to the unpolarized signal intensity with respect to the width of width of filovirus-scale objects, and combining these measurements can enhance identification and filovirus-scale objects, and combining these measurements can enhance identification and sizing of sizingstrongly of such asymmetric strongly asymmetric pathogens. In the practice, the pathogen’s orientation need not be such pathogens. In practice, pathogen’s orientation need not be known a ◦ range could determine known a priori; interrogating multiple linear polarization states over any 90 priori; interrogating multiple linear polarization states over any 90° range could determine the the object’s polarization extent ofasymmetric its asymmetric character. object’s polarization ratioratio andand thusthus the the extent of its character. 1.8
61 nm width 1.7
1.6
102 nm width 1.5
500
(a)
(b)
1000
1500 2000 2500 Nanostructure length (nm)
3000
(c)
Figure at at 530530 nm. (a) Vertically-polarized image. (b) Figure 9. 9. Polarization-induced Polarization-inducedchanges changesin insignal signal nm. (a) Vertically-polarized image. Horizontally-polarized image. Scale bar is 2 µm in both (a,b). Note alternation of brightness between (b) Horizontally-polarized image. Scale bar is 2 µm in both (a,b). Note alternation of brightness between vertical and and horizontal horizontal structures structures between between the the two two images. images. (c) vertical (c) Polarization Polarization ratio ratio (vertically-polarized (vertically-polarized B(λ) divided divided by by horizontally-polarized horizontally-polarized B(λ)) widths of B(λ) B(λ)) for for the the six six different different lengths lengths and and two two different different widths of vertical nanostructures. Horizontal error bars represent one standard deviation of the measured vertical nanostructures. Horizontal error bars represent one standard deviation of the measured length. length. Lines are guide to helpthe guide Lines are to help eye.the eye.
4. Discussion and Conclusions 4. Discussion and Conclusions We have demonstrated demonstrated successful successful fabrication fabrication and and utilization utilization of of shelf-stable, shelf-stable, biohazard-free biohazard-free We have nanostructures reference materials materials particularly particularly suitable label-free optical optical biosensing biosensing which which nanostructures as as reference suitable for for label-free relies on scattered light. We implemented our nanostructures to resemble a target pathogen with relies on scattered light. We implemented our nanostructures to resemble a target pathogen with substantial public health significance—the Ebola virus—for detection with a cutting-edge biosensor substantial public health significance—the Ebola virus—for detection with a cutting-edge biosensor platform under under development. development.To Tothe thebest bestofofour our knowledge, this is the first experimental report of platform knowledge, this is the first experimental report of the the relationships between signal intensity and object dimensions for this biosensor-pathogen relationships between signal intensity and object dimensions for this biosensor-pathogen combination. combination. As long as a thin film of resist can be coated on the biosensing substrate, EBL can create As long as a thin film of resist can be coated on the biosensing substrate, EBL can create nanostructures nanostructures on practically any surface relevant to the myriad of emerging optical biosensing on practically any surface relevant to the myriad of emerging optical biosensing techniques, including techniques, including those based on other physical phenomena like surface plasmon resonance. The those based on other physical phenomena like surface plasmon resonance. The sizes and shapes of sizes and shapes of nanoscale pathogens that can be mimicked are limited only by the EBL system’s nanoscale pathogens that can be mimicked are limited only by the EBL system’s spatial resolution, spatial resolution, on the order of 10 nm. Once an EBL protocol has been properly refined, which invariably requires multiple trial and error cycles, replicating modest quantities (tens to hundreds) of biosensing chips with customized nanostructures would be tenable. The EBL protocol could also potentially be modified to liberate nanostructures from a substrate surface into an aqueous suspension, thereby adding even greater versatility to their use as biosensing reference materials.
Sensors 2018, 18, 1670
11 of 12
on the order of 10 nm. Once an EBL protocol has been properly refined, which invariably requires multiple trial and error cycles, replicating modest quantities (tens to hundreds) of biosensing chips with customized nanostructures would be tenable. The EBL protocol could also potentially be modified to liberate nanostructures from a substrate surface into an aqueous suspension, thereby adding even greater versatility to their use as biosensing reference materials. Author Contributions: A.A. conceived and designed the study, assembled the biosensor apparatus, performed biosensing experiments, analyzed the data, and wrote the paper; J.M. designed, fabricated, and characterized the nanostructures, performed biosensing experiments, analyzed the data, and edited the paper; T.S. and K.A.C. fabricated the virus-like particles and edited the paper. Funding: This work was supported by funding from the FDA Medical Countermeasures Initiative. J.M. was supported by an appointment to the Research Participation Program at the FDA administered by the Oak Ridge Institute for Science and Education. Acknowledgments: We thank the University of Maryland NanoCenter for EBL and SEM equipment and assistance, and the FDA Advanced Characterization Facility for AFM equipment and assistance. We thank Gary Nabel for providing the EBOV GP-encoding plasmids, and Yoshihiro Kawaoka and Paul Bieniasz for providing the EBOV VP40- and VP40-GFP-encoding plasmids, respectively. The mention of commercial products, their sources, or their use in connection with material reported herein is not to be construed as either an actual or implied endorsement of such products by the US Department of Health and Human Services. This publication reflects the views of the authors and should not be construed to represent FDA’s views or policies. Conflicts of Interest: The authors declare no conflict of interest.
References 1.
2.
3.
4.
5.
6. 7. 8.
9.
10.
11.
Daaboul, G.G.; Yurt, A.; Zhang, X.; Hwang, G.M.; Goldberg, B.B.; Ünlü, M.S. High-throughput detection and sizing of individual low-index nanoparticles and viruses for pathogen identification. Nano Lett. 2010, 10, 4727–4731. [CrossRef] [PubMed] Daaboul, G.G.; Lopez, C.A.; Chinnala, J.; Goldberg, B.B.; Connor, J.H.; Ünlü, M.S. Digital sensing and sizing of vesicular stomatitis virus pseudotypes in complex media: A model for Ebola and Marburg detection. ACS Nano 2014, 8, 6047–6055. [CrossRef] [PubMed] Scherr, S.M.; Daaboul, G.G.; Trueb, J.; Sevenler, D.; Fawcett, H.; Goldberg, B.; Connor, J.H.; Ünlü, M.S. Real-time capture and visualization of individual viruses in complex media. ACS Nano 2016, 10, 2827–2833. [CrossRef] [PubMed] Daaboul, G.G.; Freedman, D.S.; Scherr, S.M.; Carter, E.; Rosca, A.; Bernstein, D.; Mire, C.E.; Agans, K.N.; Hoenen, T.; Geisbert, T.W.; et al. Enhanced light microscopy visualization of virus particles from Zika virus to filamentous ebolaviruses. PLoS ONE 2017, 12, e0179728. [CrossRef] [PubMed] Ewers, H.; Jacobsen, V.; Klotzsch, E.; Smith, A.E.; Helenius, A.; Sandoghdar, V. Label-free optical detection and tracking of single virions bound to their receptors in supported membrane bilayers. Nano Lett. 2007, 7, 2263–2266. [CrossRef] [PubMed] Mitra, A.; Deutsch, B.; Ignatovich, F.; Dykes, C.; Novotny, L. Nano-optofluidic detection of single viruses and nanoparticles. ACS Nano 2010, 4, 1305–1312. [CrossRef] [PubMed] Mitra, A.; Ignatovich, F.; Novotny, L. Real-time optical detection of single human and bacterial viruses based on dark-field interferometry. Biosens. Bioelectron. 2012, 31, 499–504. [CrossRef] [PubMed] Block, O.; Mitra, A.; Novotny, L.; Dykes, C. A rapid label-free method for quantitation of human immunodeficiency virus type-1 particles by nanospectroscopy. J. Virol. Methods 2012, 182, 70–75. [CrossRef] [PubMed] Enoki, S.; Iino, R.; Morone, N.; Kaihatsu, K.; Sakakihara, S.; Kato, N.; Noji, H. Label-free single-particle imaging of the influenza virus by objective-type total internal reflection dark-field microscopy. PLoS ONE 2012, 7, e49208. [CrossRef] [PubMed] Ma, L.; Zhu, S.; Tian, Y.; Zhang, W.; Wang, S.; Chen, C.; Wu, L.; Yan, X. Label-free analysis of single viruses with a resolution comparable to that of electron microscopy and the throughput of flow cytometry. Angew. Chem. 2016, 128, 10395–10399. [CrossRef] Noda, T.; Sagara, H.; Suzuki, E.; Takada, A.; Kida, H.; Kawaoka, Y. Ebola virus VP40 drives the formation of virus-like filamentous particles along with GP. J. Virol. 2002, 76, 4855–4865. [CrossRef] [PubMed]
Sensors 2018, 18, 1670
12.
13.
14. 15. 16.
17. 18. 19. 20.
21. 22.
23.
12 of 12
Dube, D.; Schornberg, K.L.; Shoemaker, C.J.; Delos, S.E.; Stantchev, T.S.; Clouse, K.A.; Broder, C.C.; White, J.M. Cell adhesion-dependent membrane trafficking of a binding partner for the ebolavirus glycoprotein is a determinant of viral entry. Proc. Natl. Acad. Sci. USA 2010, 107, 16637–16642. [CrossRef] [PubMed] Beniac, D.R.; Melito, P.L.; Hiebert, S.L.; Rabb, M.J.; Lamboo, L.L.; Jones, S.M.; Booth, T.F. The organisation of Ebola virus reveals a capacity for extensive, modular polyploidy. PLoS ONE 2012, 7, e29608. [CrossRef] [PubMed] Golding, C.G.; Lamboo, L.L.; Beniac, D.R.; Booth, T.F. The scanning electron microscope in microbiology and diagnosis of infectious disease. Sci. Rep. 2016, 6, 26516. [CrossRef] [PubMed] Pang, Y.; Song, H.; Cheng, W. Using optical trap to measure the refractive index of a single animal virus in culture fluid with high precision. Biomed. Opt. Express 2016, 7, 1672–1689. [CrossRef] [PubMed] McCord, M.A.; Rooks, M.J. Electron beam lithography. In Handbook of Microlithography, Micromachining, and Microfabrication; Rai-Choudhury, P., Ed.; SPIE Press: Bellingham, WA, USA, 1997; Volume 1, pp. 139–249. ISBN 978-0852969069. Lis, S.; Dylewicz, R.; My´sliwiec, J.; Miniewicz, A.; Patela, S. Application of flowable oxides in photonics. Mater. Sci. Pol. 2008, 26, 189–194. Schneider, C.A.; Rasband, W.S.; Eliceiri, K.W. NIH Image to ImageJ: 25 years of image analysis. Nat. Methods 2012, 9, 671–675. [CrossRef] [PubMed] Bohren, C.F.; Huffman, D.R. Absorption and Scattering of Light by Small Particles; John Wiley & Sons: New York, NY, USA, 1983; ISBN 978-0-471-29340-8. Simmons, G.; Rennekamp, A.J.; Chai, N.; Vandenberghe, L.H.; Riley, J.L.; Bates, P. Folate receptor alpha and caveolae are not required for Ebola virus glycoprotein-mediated viral infection. J. Virol. 2003, 77, 13433–13438. [CrossRef] [PubMed] Manicassamy, B.; Rong, L. Expression of ebolavirus glycoprotein on the target cells enhances viral entry. Virol. J. 2009, 6, 75. [CrossRef] [PubMed] Yang, J.K.; Cord, B.; Duan, H.; Berggren, K.K.; Klingfus, J.; Nam, S.W.; Kim, K.B.; Rooks, M.J. Understanding of hydrogen silsesquioxane electron resist for sub-5-nm-half-pitch lithography. J. Vac. Sci. Technol. B 2009, 27, 2622–2627. [CrossRef] Heavens, O.S. Optical Properties of Thin Solid Films; Dover: New York, NY, USA, 1991; ISBN 0-486-66924-6. © 2018 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).
