Nanogap traps for passive bacteria concentration and single-point confocal Raman spectroscopy

INTRODUCTION

Section:

ChooseTop of pageABSTRACTINTRODUCTION <<MATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONREFERENCESPrevious sectionNext sectionThe identification of bacteria is an essential task in many microbiological workflows. In addition to the need for determining the causative agents involved in bacterial infections, bacteria identification is important for a broader range of applications ranging from epidemiologic studies to food safety and beyond. Bacteria identification is also an important step in the process of screening for antibiotic resistance, which presents a significant and growing public health challenge.11. CDC, “Antibiotic resistance threats in the United States,” Report (U.S. Department of Health and Human Services, CDC, Atlanta, GA, U.S.A., 2019). The classical approach to identifying bacteria involves the use of cell culture followed by phenotypic characterization based on cell morphology, staining, or various biochemical assays.22. L. Váradi, J. L. Luo, D. E. Hibbs, J. D. Perry, R. J. Anderson, S. Orenga, and P. W. Groundwater, “Methods for the detection and identification of pathogenic bacteria: Past, present, and future,” Chem. Soc. Rev. 46, 4818–4832 (2017). https://doi.org/10.1039/C6CS00693K However, not all organisms can be cultured in vitro, and culture-based assays typically require several days to generate actionable results.33. F. P. Maurer, M. Christner, M. Hentschke, and H. Rohde, “Advances in rapid identification and susceptibility testing of bacteria in the clinical microbiology laboratory: Implications for patient care and antimicrobial stewardship programs,” Infect. Dis. Rep. 9, 6839 (2017). https://doi.org/10.4081/idr.2017.6839 Because of the long time scale for culture-based bacteria identification, treatment selection often relies on clinical observation alone, resulting in the unnecessary use of broad spectrum drugs and leading directly to the increasing prevalence of antibiotic resistant pathogens.44. B. A. Lipsky, M. Dryden, F. Gottrup, D. Nathwani, R. A. Seaton, and J. Stryja, “Antimicrobial stewardship in wound care: A position paper from the British society for antimicrobial chemotherapy and European wound management association,” J. Antimicrob. Chemother. 71, 3026–3035 (2016). https://doi.org/10.1093/jac/dkw287 It is widely recognized that new culture-free assays are needed to improve clinical decision-making for bacterial infections.55. E. Avdic and K. C. Carroll, “The role of the microbiology laboratory in antimicrobial stewardship programs,” Infect. Dis. Clin. North Am. 28, 215–235 (2014). https://doi.org/10.1016/j.idc.2014.01.002 To overcome these limitations, rapid culture-independent molecular assays have emerged as a powerful alternative for bacterial identification, with the most common approach based on multiplexed PCR targeting the high-conserved 16S rRNA gene that presents species-specific regions.66. R. Franco-Duarte, L. Černáková, S. Kadam, K. S. Kaushik, B. Salehi, A. Bevilacqua, M. R. Corbo, H. Antolak, K. Dybka-Stępień, M. Leszczewicz et al., “Advances in chemical and biological methods to identify microorganisms-from past to present,” Microorganisms 7, 130 (2019). https://doi.org/10.3390/microorganisms7050130 While genotypic assays can be significantly faster than culture-based characterization, these tests remain cumbersome and expensive and offer limited agility in adapting to new pathogens.Raman spectroscopy is an optical imaging technique that has been successfully harnessed to evaluate bacteria from clinical specimens. Raman spectra contain information from the inelastic scattering of light due to the vibrational and rotational states of the target molecules.77. R. R. Jones, D. C. Hooper, L. Zhang, D. Wolverson, and V. K. Valev, “Raman techniques: Fundamentals and frontiers,” Nanoscale Res. Lett. 14, 231 (2019). https://doi.org/10.1186/s11671-019-3039-2 The technique is a label-free approach that can generate pathogen-specific fingerprints reflecting distinct bacterial phenotypes based on specific molecular attributes.88. W. R. Premasiri, J. C. Lee, A. Sauer-Budge, R. Théberge, C. E. Costello, and L. D. Ziegler, “The biochemical origins of the surface-enhanced Raman spectra of bacteria: A metabolomics profiling by SERS,” Anal. Bioanal. Chem. 408, 4631–4647 (2016). https://doi.org/10.1007/s00216-016-9540-x Raman imaging has been widely demonstrated for strain-level bacteria identification9–119. S. Pahlow, S. Meisel, D. Cialla-May, K. Weber, P. Rösch, and J. Popp, “Isolation and identification of bacteria by means of Raman spectroscopy,” Adv. Drug Delivery Rev. 89, 105–120 (2015). https://doi.org/10.1016/j.addr.2015.04.00610. S. Stöckel, J. Kirchhoff, U. Neugebauer, P. Rösch, and J. Popp, “The application of Raman spectroscopy for the detection and identification of microorganisms,” J. Raman Spectrosc. 47, 89–109 (2016). https://doi.org/10.1002/jrs.484411. B. Lorenz, C. Wichmann, S. Stöckel, P. Rösch, and J. Popp, “Cultivation-free Raman spectroscopic investigations of bacteria,” Trends Microbiol. 25, 413–424 (2017). https://doi.org/10.1016/j.tim.2017.01.002 and sub-typing,1212. A. Germond, T. Ichimura, T. Horinouchi, H. Fujita, C. Furusawa, and T. M. Watanabe, “Raman spectral signature reflects transcriptomic features of antibiotic resistance in Escherichia coli,” Commun. Biol. 1, 85 (2018). https://doi.org/10.1038/s42003-018-0093-8,1313. K. C. Henderson, A. J. Benitez, A. E. Ratliff, D. M. Crabb, E. S. Sheppard, J. M. Winchell, R. A. Dluhy, K. B. Waites, T. P. Atkinson, and D. C. Krause, “Specificity and strain-typing capabilities of nanorod array-surface enhanced Raman spectroscopy for Mycoplasma pneumoniae detection,” PLoS One 10, e0131831 (2015). https://doi.org/10.1371/journal.pone.0131831 and the use of machine learning techniques for automated feature classification have enabled a wide range of bacterial pathogens to be identified from clinical isolates with exceptional accuracy.1414. C.-S. Ho, N. Jean, C. A. Hogan, L. Blackmon, S. S. Jeffrey, M. Holodniy, N. Banaei, A. A. E. Saleh, S. Ermon, and J. Dionne, “Rapid identification of pathogenic bacteria using Raman spectroscopy and deep learning,” Nat. Commun. 10, 4927 (2019). https://doi.org/10.1038/s41467-019-12898-9 Raman spectroscopy has further been explored as a culture-free approach to the evaluation of antibiotic susceptibility by monitoring spectral changes during the exposure of bacteria to selected antimicrobial agents.12–2112. A. Germond, T. Ichimura, T. Horinouchi, H. Fujita, C. Furusawa, and T. M. Watanabe, “Raman spectral signature reflects transcriptomic features of antibiotic resistance in Escherichia coli,” Commun. Biol. 1, 85 (2018). https://doi.org/10.1038/s42003-018-0093-815. U. Münchberg, P. Rösch, M. Bauer, and J. Popp, “Raman spectroscopic identification of single bacterial cells under antibiotic influence,” Anal. Bioanal. Chem. 406, 3041–3050 (2014). https://doi.org/10.1007/s00216-014-7747-216. U. C. Schröder, C. Beleites, C. Assmann, U. Glaser, U. Hübner, W. Pfister, W. Fritzsche, J. Popp, and U. Neugebauer, “Detection of vancomycin resistances in enterococci within 3 1/2h,” Sci. Rep. 5, 8217 (2015). https://doi.org/10.1038/srep0821717. L. Teng, X. Wang, X. Wang, H. Gou, L. Ren, T. Wang, Y. Wang, Y. Ji, W. E. Huang, and J. Xu, “Label-free, rapid and quantitative phenotyping of stress response in E. Coli via ramanome,” Sci. Rep. 6, 34359 (2016). https://doi.org/10.1038/srep3435918. U. C. Schröder, J. Kirchhoff, U. Hübner, G. Mayer, U. Glaser, T. Henkel, W. Pfister, W. Fritzsche, J. Popp, and U. Neugebauer, “On-chip spectroscopic assessment of microbial susceptibility to antibiotics within 3.5h,” J. Biophotonics 10, 1547–1557 (2017). https://doi.org/10.1002/jbio.20160031619. Y. Tao, Y. Wang, S. Huang, P. Zhu, W. E. Huang, J. Ling, and J. Xu, “Metabolic-activity-based assessment of antimicrobial effects by D2O-labeled single-cell Raman microspectroscopy,” Anal. Chem. 89, 4108–4115 (2017). https://doi.org/10.1021/acs.analchem.6b0505120. C. Assmann, J. Kirchhoff, C. Beleites, J. Hey, S. Kostudis, W. Pfister, P. Schlattmann, J. Popp, and U. Neugebauer, “Identification of vancomycin interaction with Enterococcus faecalis within 30 min of interaction time using Raman spectroscopy,” Anal. Bioanal. Chem. 407, 8343–8352 (2015). https://doi.org/10.1007/s00216-015-8912-y21. A. I. M. Athamneh, R. A. Alajlouni, R. S. Wallace, M. N. Seleem, and R. S. Senger, “Phenotypic profiling of antibiotic response signatures in Escherichia coli using Raman spectroscopy,” Antimicrob. Agents Chemother. 58, 1302–1314 (2014). https://doi.org/10.1128/AAC.02098-13 Raman analysis has also shown significant potential for assessing antibiotic susceptibility without the need for drug exposure. For example, Raman spectroscopy can be used to directly identify transcriptomic features of antibiotic resistance1212. A. Germond, T. Ichimura, T. Horinouchi, H. Fujita, C. Furusawa, and T. M. Watanabe, “Raman spectral signature reflects transcriptomic features of antibiotic resistance in Escherichia coli,” Commun. Biol. 1, 85 (2018). https://doi.org/10.1038/s42003-018-0093-8 and can differentiate susceptible vs resistant strains with high confidence after model training using bacteria with different resistance profiles.1414. C.-S. Ho, N. Jean, C. A. Hogan, L. Blackmon, S. S. Jeffrey, M. Holodniy, N. Banaei, A. A. E. Saleh, S. Ermon, and J. Dionne, “Rapid identification of pathogenic bacteria using Raman spectroscopy and deep learning,” Nat. Commun. 10, 4927 (2019). https://doi.org/10.1038/s41467-019-12898-9These latter techniques make it possible to select an optimal treatment based on optical interrogation of a clinical sample without the need for cell culture or labeling reagents. However, due to the low scattering efficiency associated with Raman spectroscopy, high signal-to-noise ratios are necessary to extract sufficient signal for effective bacteria characterization. As a result, long data acquisition times and high bacteria concentrations are required, constraining the utility of conventional Raman detection for many clinical applications. To address this limitation, Raman spectroscopy using confocal optics can significantly improve noise rejection and enable fingerprinting of small numbers of bacteria with short imaging times.2222. D. Wang, P. He, Z. Wang, G. Li, N. Majed, and A. Z. Gu, “Advances in single cell Raman spectroscopy technologies for biological and environmental applications,” Curr. Opin. Biotechnol. 64, 218–229 (2020). https://doi.org/10.1016/j.copbio.2020.06.011 However, the femtoliter-scale detection volume associated with confocal detection imposes the need for highly time-consuming sample scanning to acquire target cells, making this approach unsuitable for high throughput diagnostics involving samples with low bacteria counts. A promising approach for overcoming the confocal Raman throughput constraint is to trap bacteria at a known location on a target substrate, allowing efficient spectra acquisition without the need for extensive sample scanning. To this end, a number of techniques for bacteria trapping prior to Raman-based characterization have been explored. For example, optical tweezing has been widely used to capture bacteria down to the single cell level for confocal Raman detection.23–2923. C. Xie, J. Mace, M. A. Dinno, Y. Q. Li, W. Tang, R. J. Newton, and P. J. Gemperline, “Identification of single bacterial cells in aqueous solution using confocal laser tweezers Raman spectroscopy,” Anal. Chem. 77, 4390–4397 (2005). https://doi.org/10.1021/ac050497124. D. V. Petrov, “Raman spectroscopy of optically trapped particles,” J. Opt. A Pure Appl. Opt. 9, S139–S156 (2007). https://doi.org/10.1088/1464-4258/9/8/S0625. C. Xie, C. Goodman, M. A. Dinno, and Y.-Q. Li, “Real-time Raman spectroscopy of optically trapped living cells and organelles,” Opt. Express 12, 6208 (2004). https://doi.org/10.1364/OPEX.12.00620826. J. W. Chan, A. P. Esposito, C. E. Talley, C. W. Hollars, S. M. Lane, and T. Huser, “Reagentless identification of single bacterial spores in aqueous solution by confocal laser tweezers Raman spectroscopy,” Anal. Chem. 76, 599–603 (2004). https://doi.org/10.1021/ac035015527. C. Xie and Y. Q. Li, “Confocal micro-Raman spectroscopy of single biological cells using optical trapping and shifted excitation difference techniques,” J. Appl. Phys. 93, 2982–2986 (2003). https://doi.org/10.1063/1.154265428. T. Dahlberg, D. Malyshev, P. O. Andersson, and M. Andersson, “Biophysical fingerprinting of single bacterial spores using laser Raman optical tweezers,” Proc. SPIE 11416, 114160I (2020). https://doi.org/10.1117/12.255810229. F. Sinjab, D. Awuah, G. Gibson, M. Padgett, A. M. Ghaemmaghami, and I. Notingher, “Holographic optical trapping Raman micro-spectroscopy for non-invasive measurement and manipulation of live cells,” Opt. Express 26, 25211–25225 (2018). https://doi.org/10.1364/OE.26.025211 Although optical tweezers can serve to immobilize and concentrate bacteria in bulk solution while positioning the detection volume away from the device substrate to reduce background signal, the technology requires complex optics and is not readily automated for use in clinical environments. Similarly, a variety of microfluidic-enabled trapping methods based on acoustic,3030. V. O. Baron, M. Chen, B. Hammarstrom, R. J. H. Hammond, P. Glynne-Jones, S. H. Gillespie, and K. Dholakia, “Real-time monitoring of live mycobacteria with a microfluidic acoustic-Raman platform,” Commun. Biol. 3, 236 (2020). https://doi.org/10.1038/s42003-020-0915-3 electrophoretic,3131. B. Krafft, A. Tycova, R. D. Urban, C. Dusny, and D. Belder, “Microfluidic device for concentration and SERS-based detection of bacteria in drinking water,” Electrophoresis 42, 86–94 (2021). https://doi.org/10.1002/elps.202000048 or dielectrophoretic16–3516. U. C. Schröder, C. Beleites, C. Assmann, U. Glaser, U. Hübner, W. Pfister, W. Fritzsche, J. Popp, and U. Neugebauer, “Detection of vancomycin resistances in enterococci within 3 1/2h,” Sci. Rep. 5, 8217 (2015). https://doi.org/10.1038/srep0821732. H.-Y. Lin, C.-H. Huang, W.-H. Hsieh, L.-H. Liu, Y.-C. Lin, C.-C. Chu, S.-T. Wang, I.-T. Kuo, L.-K. Chau, and C.-Y. Yang, “On-line SERS detection of single bacterium using novel SERS nanoprobes and a microfluidic dielectrophoresis device,” Small 10, 4700–4710 (2014). https://doi.org/10.1002/smll.20140152633. I.-F. I-Fang Cheng, H.-C. Chang, T.-Y. Chen, C. Hu, and F.-L. Yang, “Rapid (<5 min) identification of pathogen in human blood by electrokinetic concentration and surface-enhanced Raman spectroscopy,” Sci. Rep. 3, 2365 (2013). https://doi.org/10.1038/srep0236534. X. Chen, Z. Liang, D. Li, Y. Xiong, P. Xiong, Y. Guan, S. Hou, Y. Hu, S. Chen, G. Liu et al., “Microfluidic dielectrophoresis device for trapping, counting and detecting Shewanella oneidensis at the cell level,” Biosens. Bioelectron. 99, 416–423 (2018). https://doi.org/10.1016/j.bios.2017.08.01735. D. Hou, S. Maheshwari, and H.-C. Chang, “Rapid bioparticle concentration and detection by combining a discharge driven vortex with surface enhanced Raman scattering,” Biomicrofluidics 1, 014106 (2007). https://doi.org/10.1063/1.2710191 actuation have been demonstrated for bacteria localization prior to Raman analysis. The use of microfluidic flow cells in these platforms provides the advantage of allowing larger sample volumes to be introduced during the bacteria concentration process to enable sensitive detection from dilute samples. However, these microfluidic techniques require the use of an active trapping mechanism based on either piezoelectric or electrokinetic transduction for bacteria concentration, adding cost and complexity to device fabrication and operation. Furthermore, the high electric fields required for electrophoretic mobilization may damage bacteria cells during the concentration process,3636. G. Beretta, A. F. Mastorgio, L. Pedrali, S. Saponaro, and E. Sezenna, “The effects of electric, magnetic and electromagnetic fields on microorganisms in the perspective of bioremediation,” Rev. Environ. Sci. Bio/Technol. 18, 29–75 (2019). https://doi.org/10.1007/s11157-018-09491-9 particularly in the case of Gram-negative cells that lack a protective peptidoglycan layer,3737. M. Szumski, E. Kłodzińska, E. Dziubakiewicz, K. Hrynkiewicz, and B. Buszewski, “Effect of applied voltage on viability of bacteria during separation under electrophoretic conditions,” J. Liq. Chromatogr. Relat. Technol. 34, 2689–2698 (2011). https://doi.org/10.1080/10826076.2011.593223 thereby modifying their Raman signatures.Here, we describe a simple and fully passive bacteria trapping technique capable of capturing bacteria at a fixed location within a microfluidic chamber for confocal Raman analysis. As depicted in Fig. 1, the devices employ an angled nanogap structure for size-selective bacteria localization, with isolation and concentration of cells at the sharp tip of trap occurring by cell rolling during sample perfusion. A large nanogap perimeter allows sample to be delivered at high perfusion rates, with mobilized cells rolling along the gap to reach the detection zone at the apex. Following bacteria trapping, the system is rinsed with fresh buffer to purify captured cells and reduce optical interference. In addition, the trap contents are dried without disrupting the focused bacteria, eliminating background signal from Raman scattering of water. The volume of the trapping zone is well matched to the confocal Raman optics, allowing high quality spectral data to be rapidly collected without sample scanning. Because the microfluidic flow cell operates by passive perfusion of sample, rinse buffer, and air, it provides robust and reliable operation with minimal infrastructure. Here, we evaluate the performance of the nanogap technology for bacteria trapping and confocal Raman detection. Selections of both Gram-positive and Gram-negative pathogens are evaluated with an average of 100 cells used in each experiment. The resulting Raman measurements yield distinct spectral fingerprints that are found to match favorably with off-chip reference samples requiring manual scanning for spectrum acquisition, with additional features observed in the on-chip spectra offering improved discrimination between the captured bacteria despite significantly smaller cell counts used in the microfluidic experiments.

MATERIALS AND METHODS

Section:

ChooseTop of pageABSTRACTINTRODUCTIONMATERIALS AND METHODS <<RESULTS AND DISCUSSIONCONCLUSIONREFERENCESPrevious sectionNext section

Nanogap chip fabrication

The nanogap chip design consists of a 1 mm diameter and 20 μm deep circular chamber connected to 200 μm wide inlet and outlet channels patterned in a silicon wafer. The circular chamber is occluded by a V-shaped barrier etched to a depth of 500 nm to form a nanogap between the top of the barrier and the upper sealing layer. The barrier possesses a 30° half-angle and has a total length of 1645 μm. As depicted in Fig. 2(a), the channel and nanogap structures were fabricated in a heavily doped 100 mm diameter p-type Si wafer. The nanogap was first patterned by etching the circular chamber to a depth of 500 nm by reactive ion etching, followed by masking the barrier and etching the deeper regions of the chamber and the fluidic channels by silicon deep reactive ion etching (DRIE). The resulting nanogap was designed with a width of 10 μm. The wafer was diced into 2 cm square chips [Fig. 2(b)], and each chip was individually sealed using a 160 μm thick borosilicate glass coverslip (Borofloat 33, Schott, Jena, Germany) by anodic bonding. Before bonding, the coverslip was first patterned to form fluidic access ports using a dry film photoresist (RapidMask High-Tack, Ikonics Imaging, Duluth, MN) as a masking layer during abrasive glass etching using alumina microparticles by micro powder blasting (Accuflo MicroBlaster, Comco Inc., Burbank, CA). The patterned silicon and glass substrates were cleaned by piranha solution and manually aligned before sealing the channels via anodic bonding. Bonding was performed on a hot plate at 525 °C while applying a 500 V bias through a conductive pin pressed into the center of the glass/silicon stack with the silicon substrate held at 0 V. After bonding, fluid reservoirs were formed by punching 2 mm diameter access ports in two pieces of fully cured polydimethylsiloxane (PDMS) and temporarily bonding the PDMS pieces to the glass cover plate with holes aligned to the inlet and outlet ports. Magnified images of the nanogap chamber and angled trap tip are provided in Figs. 2(c) and 2(d), respectively.

Numerical modeling

The nanogap design was studied using the computational fluid dynamics (CFD) and particle tracing modules in COMSOL Multiphysics (COMSOL Inc., Burlington, MA). Simulations were performed through a two-step process. After first determining the steady-state fluid flow profile for time-independent laminar flow, the particle tracing module was applied to simulate time-dependent particle movements via coupled particle tracing within the laminar flow profile. Inlet flow velocity was set to 20 μm/s as the average linear velocity of a fully developed flow. A no-slip boundary condition was applied to all surfaces except the inlet and outlet. The model geometry for the focusing chamber and nanogap structure was identical to the fabricated devices, but with analysis limited to a 250 μm axial chamber length to ensure numerical stability. For particle tracing, 2 μm diameter spherical particles were released at the model inlet. Each particle was simulated as a point mass with a body force vector calculated by considering the particle diameter and density together with the fluid streamlines determined in the first step of the simulation. Particle–particle interactions were not considered, and the particles were prohibited from modifying the flow characteristics, ensuring that nanogap clogging and similar effects that could occur at the high particle counts used in the simulations did not impact the numerical results. Size selection in the nanogap was approximated by assigning a conditional boundary at the nanogap entrance to diffusively scatter particles larger than 500 nm while preserving the kinetic energy of the particle. Wall lift forces were applied to all remaining boundaries, and Stokes drag force was applied to all released particles.

Bacteria trapping

Before introducing bacteria solution into a nanogap chip, the device was first primed by pipetting 10 μl DI water directly into the PDMS outlet port, followed by withdraw of fluid through the inlet port using a manual syringe connected via Tygon tubing to a pipet tip inserted into the PDMS reservoir. The reversed flow used during priming served to minimize the potential for particulates to be introduced into the trap chamber. After verifying that no bubbles were trapped in the internal flow path by the priming solution, 100 μl of bacteria solution was added to the open inlet reservoir, and the pipet tip inserted into the outlet port was connected to a syringe pump operating in withdraw mode for sample perfusion. Experiments were performed using Enterobacter cloacae, Klebsiella pneumoniae, and Corynebacterium diphtheriae as model pathogens. All three bacteria are rod-shaped with minor and major axis dimensions of approximately 0.5 and 2 μm, respectively. To prepare bacteria samples, lyophilized powders of C. cloacae, K. pneumoniae, and C. diphtheriae (Carolina Biological Supply, Burlington, NC) were separately cultured following the manufacturer's protocol in sterile glass tubes with nutrient broth and tryptic soy broth (Carolina Biological Supply). Bacteria stock solutions were prepared by replacing the growth medium of the as-cultured bacteria with 1× phosphate buffered saline (PBS) via centrifugation to a concentration of 103 CFU/ml. For each solution, a 100 μl sample volume containing approximately 100 CFU was perfused through the nanogap device at 10 μl/min using a syringe pump for approximately 10 min until the entire sample was injected through the chip. After bacteria trapping, the device was rinsed with 30 μl 1× PBS followed by 100 μl DI water using the same process employed for sample perfusion. Thorough DI rinsing was necessary to remove salts and other contaminants which can generate significant Raman background. Finally, ambient air was drawn through the device by applying weak vacuum to the outlet until all residual water was depleted, leaving a cluster of dried bacteria within the trap tip for Raman analysis. The time required to complete the entire sample introduction process including rinsing and air drying was approximately 30 min.

Reference sample preparation

Reference samples used for scanning confocal Raman imaging without nanogap processing were prepared by pipetting 20 μl of high-concentration bacteria solution onto a clean glass slide and allowing the slide to dry under a biosafety cabinet for 1 h at room temperature before confocal Raman imaging. The reference sample solutions were formed by replacing the growth medium of harvested bacteria with DI water to a final concentration of 107 CFU/ml. Dilutions were performed using DI water instead of PBS buffer to avoid Raman background signal associated with crystallized buffer salts in the dried sample spots.

Raman detection and peak analysis

Raman spectroscopy was performed using a Horiba LabRAM confocal Raman instrument equipped with a 50× NA 0.75 objective and 532 nm excitation laser. The selected lens provided a working distance of 330 μm, sufficient for imaging through the 160 μm thick glass chip lid. Following bacteria trapping, rinsing, and air drying, the nanogap chip was placed in the Raman microscope with the confocal detection volume positioned 3–4 μm behind the apex of the nanogap trap. Raman signal was acquired by averaging over six sequential measurements with a 10 s scan time per measurement. Following data collection, background estimation and correction were performed for each spectrum using the Statistics-sensitive Non-linear Iterative Peak-clipping (SNIP) algorithm implemented in the R programming language.3838. C. G. Ryan, E. Clayton, W. L. Griffin, S. H. Sie, and D. R. Cousens, “SNIP, a statistics-sensitive background treatment for the quantitative analysis of PIXE spectra in geoscience applications,” Nucl. Instrum. Methods Phys. Res. B 34, 396–402 (1988). https://doi.org/10.1016/0168-583X(88)90063-8,3939. M. Morháč and V. Matoušek, “Peak clipping algorithms for background estimation in spectroscopic data,” Appl. Spectrosc. 62, 91–106 (2008). https://doi.org/10.1366/000370208783412762 Automated peak identification from the spectroscopic data was performed using a deconvolution method integrated into the SNIP package.4040. M. Morháč, J. Kliman, V. Matoušek, M. Veselský, and I. Turzo, “Identification of peaks in multidimensional coincidence γ-ray spectra,” Nucl. Instrum. Methods Phys. Res. A 443, 108–125 (2000). https://doi.org/10.1016/S0168-9002(99)01005-0 Identical background estimation and peak identification parameters were employed for all samples including reference bacteria.

RESULTS AND DISCUSSION

Section:

ChooseTop of pageABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSION <<CONCLUSIONREFERENCESPrevious sectionNext sectionDevices were fabricated using 160 μm thick borosilicate glass coverslips to seal the microfluidic channels and form the upper surface of the nanogaps. In addition to supporting anodic bonding for reliable sealing to the silicon substrate, borosilicate glass was employed due to its relatively low background fluorescence during Raman analysis. The coverslip thickness was selected to support optical access to the trapping zone during confocal Raman imaging. One disadvantage to the low thickness is that the millimeter-scale trap chamber can act as a deformable membrane during sample perfusion, potentially altering the nanogap height and impacting the bacteria trapping process. For the 10 μl/min flow rate used in our studies, the approximate pressure across the glass layer was determined to be 0.16 atm based on plane Poiseuille flow through the gap, resulting in an estimated peak deflection of 120 nm based on an analytic model of deflection for a circular membrane.4141. S. Timoshenko and S. Woinoskly-Krieger, Theory of Plates and Shells (McGraw-Hill, 1959).,4242. F. Bitsie, W. P. Eaton, D. W. Plummer, and J. H. Smith, “A new analytical solution for diaphragm deflection and its application to a surface-micromachined pressure sensor,” in Proceedings: International Conference on Modeling and Simulation of Microsystems (NANOTECH, 1999), pp. 640–643. Because the devices are operated by negative pressure, the estimated membrane deflection results in a moderate reduction in nanogap height and is not expected to have a significant impact on the bacteria trapping process.During sample perfusion within the nanogap devices, cells initially follow streamlines that traverse the nanogap, forcing bacteria to impact the gap barrier. The rigid cell structure limits bacteria deformation and prevents significant intrusion of the cells into the gap, thereby accumulating bacteria at the gap opening while allowing smaller and more compliant contaminants such as cell lysis debris to pass through the gap unimpeded. This process is conceptually similar to prior work on compliance-based cell separation and concentration using an array of gap structures.4343. G. Wang, W. Mao, R. Byler, K. Patel, C. Henegar, A. Alexeev, and T. Sulchek, “Stiffness dependent separation of cells in a microfluidic device,” PLoS One 8, e75901 (2013). https://doi.org/10.1371/journal.pone.0075901 The angled V-shaped barrier was designed to increase the gap perimeter within a fixed device length while maintaining a component of fluid momentum across the nanogap aligned to the longitudinal axis of the chamber, thereby forcing bacteria cells to roll along the gap opening toward the tip of the trap and eventually accumulate at the apex. While further studies are needed to assess the impact of gap geometry on bacteria focusing performance, we anticipate that increasing the chamber length and reducing the gap angle can improve the process by allowing higher flow rates and increased longitudinal mobility for more rapid bacteria concentration at the chamber tip. A numerical simulation depicting the focusing process for the current chip design is presented in Fig. 3. In this model, rigid spherical particles released upstream of the chamber tip advect to the nanogap opening under the influence of bulk fluid flow. Upon reaching the gap, particles are prevented from following the streamlines through the gap by size exclusion within the sub-micrometer opening. The constrained particles roll along the gap wall toward the chamber tip due to the presence of a flow velocity component parallel to the wall. The average particle velocity along the trap wall remains constant during this process with a value approximately 20% of the bulk flow velocity through the nanogap itself. As expected, particles migrate to the tip of the chamber where they are immobilized.We note that this simple rigid sphere model provides an idealized view of the bacteria trapping process that ignores interactions between the bacteria and nanogap wall that can impact both the transport rate and fate of bacteria during perfusion. In particular, several factors that may impede transport are not considered. For example, adhesive wall interactions are not included in the model nor are forces associated with flow near the nanogap wall, where bounded shear flow can reduce transport parallel to the wall surface.4444. C. Jayageeth, V. I. Sharma, and A. Singh, “Dynamics of short fiber suspensions in bounded shear flow,” Int. J. Multiphase Flow 35, 261–269 (2009). https://doi.org/10.1016/j.ijmultiphaseflow.2008.11.002,4545. E. Gavze and M. Shapiro, “Particles in a shear flow near a solid wall: Effect of nonsphericity on forces and velocities,” Int. J. Multiphase Flow 23, 155–182 (1997). https://doi.org/10.1016/S0301-9322(96)00054-7 At the same time, several interactions that serve to counteract these retarding factors are also not considered, including wall-induced lift forces and repulsive electrostatic interactions between the negative charges associated with the SiO2 channel surfaces and bacterial lipoproteins.46–4846. A. M. Elmahdy, S. E. El-Mofty, N. A. Abdel-Khalek, and A. A. El-Midany, “Impact of the adsorption of Corynebacterium diphtheriae intermedius bacteria on enhancing the separation selectivity of dolomite and apatite,” Adsorpt. Sci. Technol. 29, 47–58 (2011). https://doi.org/10.1260/0263-6174.29.1.4747. X. Hua, Z. Wu, H. Zhang, D. Lu, M. Wang, Y. Liu, and Z. Liu, “Degradation of hexadecane by Enterobacter cloacae strain TU that secretes an exopolysaccharide as a bioemulsifier,” Chemosphere 80, 951–956 (2010). https://doi.org/10.1016/j.chemosphere.2010.05.00248. T. Velkov, Z. Z. Deris, J. X. Huang, M. A. K. Azad, M. Butler, S. Sivanesan, L. M. Kaminskas, Y.-D. Dong, B. Boyd, M. A. Baker et al., “Surface changes and polymyxin interactions with a resistant strain of Klebsiella pneumoniae,” Innate Immunity 20, 350–363 (2014). https://doi.org/10.1177/1753425913493337 A more significant factor impacting bacteria transport is elastic deformation of the due cells to viscous drag force normal to the nanogap wall. The plasma membrane of both Gram-positive and Gram-negative bacteria is encased by a rigid peptidoglycan layer that confers high stiffness, allowing bacterial cells to withstand at least several atmospheres of intracellular pressure without rupture.4949. J. V. Höltje, “Growth of the stress-bearing and shape-maintaining murein sacculus of Escherichia coli,” Microbiol. Mol. Biol. Rev. 62, 181–203 (1998). https://doi.org/10.1128/MMBR.62.1.181-203.1998,5050. H. H. Tuson, G. K. Auer, L. D. Renner, M. Hasebe, C. Tropini, M. Salick, W. C. Crone, A. Gopinathan, K. C. Huang, and D. B. Weibel, “Measuring the stiffness of bacterial cells from growth rates in hydrogels of tunable elasticity,” Mol. Microbiol. 84, 874–891 (2012). https://doi.org/10.1111/j.1365-2958.2012.08063.x While the elastic moduli of many bacteria can be in the range of 50–200 MPa,5050. H. H. Tuson, G. K. Auer, L. D. Renner, M. Hasebe, C. Tropini, M. Salick, W. C. Crone, A. Gopinathan, K. C. Huang, and D. B. Weibel, “Measuring the stiffness of bacterial cells from growth rates in hydrogels of tunable elasticity,” Mol. Microbiol. 84, 874–891 (2012). https://doi.org/10.1111/j.1365-2958.2012.08063.x small cell deformations during perfusion may still lead to cell immobilization within the nanogaps. For a rod-shaped bacterium in viscous shear flow, shear stresses arising from the velocity gradient of the flow field will induce the cell to align its major axis to the local streamlines.5151. R. Rusconi, J. S. Guasto, and R. Stocker, “Bacterial transport suppressed by fluid shear,” Nat. Phys. 10, 212–217 (2014). https://doi.org/10.1038/nphys2883,5252. G. B. Jeffery, “The motion of ellipsoidal particles immersed in a viscous fluid,” Proc. R. Soc. London. Ser. A 102, 161–179 (1922). https://doi.org/10.1098/rspa.1922.0078 Because the minor axis dimensions of the bacteria used in this study are similar to the designed gap height, penetration of flow-aligned bacteria into the gap can occur at relatively low forces. This behavior was evaluated as a function of flow rate within the nanogap chips. While higher perfusion rates were observed to result in increased bacteria loss within the nanogaps, limiting the maximum flow rate to 10 μl/min was found to ensure cell loss below 5% for the specific device dimensions studied here.Bacteria trapping and confocal Raman analysis experiments were performed using pathogenic E. cloacae, K. pneumoniae, and C. diphtheriae bacteria. Multi-drug resistant strains of each pathogen have emerged in recent years that present an increased challenge for clinical treatment.53–5553. M. K. Annavajhala, A. Gomez-Simmonds, and A.-C. Uhlemann, “Multidrug-resistant Enterobacter cloacae complex emerging as a global, diversifying threat,” Front. Microbiol. 10, 44 (2019). https://doi.org/10.3389/fmicb.2019.0004454. M. Bassetti, E. Righi, A. Carnelutti, E. Graziano, and A. Russo, “Multidrug-resistant Klebsiella pneumoniae: Challenges for treatment, prevention and infection control,” Expert Rev. Anti-Infect. Ther. 16, 749–761 (2018). https://doi.org/10.1080/14787210.2018.152224955. T. M. Walker, M. Merker, A. M. Knoblauch, P. Helbling, O. D. Schoch, M. J. van der Werf, K. Kranzer, L. Fiebig, S. Kröger, W. Haas et al., “A cluster of multidrug-resistant Mycobacterium tuberculosis among patients arriving in Europe from the Horn of Africa: A molecular epidemiological study,” Lancet Infect. Dis. 18, 431–440 (2018). https://doi.org/10.1016/S1473-3099(18)30004-5E. cloacae is a Gram-negative bacterium that is a routine cause of infection in healthcare settings.5656. R. Keller, M. Z. Pedroso, R. Ritchmann, and R. M. Silva, “Occurrence of virulence-associated properties in Enterobacter cloacae,” Infect. Immun. 66, 645–649 (1998). https://doi.org/10.1128/IAI.66.2.645-649.1998K. pneumoniae is a Gram-negative bacterium that is also associated with multiple nosocomial infections, including pneumonia, meningitis, wound infections, and bloodstream infections.5757. Y. Keynan and E. Rubinstein, “The changing face of Klebsiella pneumoniae infections in the community,” Int. J. Antimicrob. Agents 30, 385–389 (2007). https://doi.org/10.1016/j.ijantimicag.2007.06.019C. diphtheriae is a Gram-positive, pathogenic bacterium that causes diphtheria.5858. P. A. Hoskisson, “Microbe profile: Corynebacterium diphtheriae—An old foe always ready to seize opportunity,” Microbiology 164, 865–867 (2018). https://doi.org/10.1099/mic.0.000627 Following the perfusion and trapping of approximately 100 CFU of each bacterium through a nanogap device, single-point confocal Raman spectroscopy was performed at the defined trapping point. Detection was performed with the Raman microscope focus positioned axially along the centerline of the perfusion chamber and approximately 3–4 μm in front of the chamber apex [Fig. 4(a)] and vertically 3–4 μm below the bottom surface of the glass lid. This position was selected to maximize bacteria signal while avoiding interference from the silicon substrate. The Raman background of the silicon surface was also measured from the tip of the nanogap trap, as shown in Fig. 4(b).The resulting Raman spectra collected after trapping each bacteria sample are presented in Fig. 5. Reference spectra obtained using high-concentration cell suspensions deposited onto bare glass cover slips are also provided in this figure. The reference spectra were collected by manual scanning of the deposited cells performed to maximize signal, providing a direct comparison to conventional Raman analysis of high-concentration bacteria samples. The SNIP algorithm3838. C. G. Ryan, E. Clayton, W. L. Griffin, S. H. Sie, and D. R. Cousens, “SNIP, a statistics-sensitive background treatment for the quantitative analysis of PIXE spectra in geoscience applications,” Nucl. Instrum. Methods Phys. Res. B 34, 396–402 (1988). https://doi.org/10.1016/0168-583X(88)90063-8,3939. M. Morháč and V. Matoušek, “Peak clipping algorithms for background estimation in spectroscopic data,” Appl. Spectrosc. 62, 91–106 (2008). https://doi.org/10.1366/000370208783412762 was used to subtract the silicon background signal, and automated peak identification was performed through deconvolution of the processed spectra.4040. M. Morháč, J. Kliman, V. Matoušek, M. Veselský, and I. Turzo, “Identification of peaks in multidimensional coincidence γ-ray spectra,” Nucl. Instrum. Methods Phys. Res. A 443, 108–125 (2000). https://doi.org/10.1016/S0168-9002(99)01005-0 As can be seen in Fig. 5, there is excellent agreement between individual peaks within the on-chip and reference spectra for each organism. With the exception of a single reference peak for C. diphtheriae that is not found in the corresponding nanogap spectrum, all peaks detected from the reference samples are also identified from the measurements performed using the nanogap chip. Significantly, on-chip Raman analysis of the focused bacteria resulted in the identification of multiple for C. diphtheriae and K. pneumoniae that were not observed from the high-concentration reference samples. In particular, the K. pneumoniae spectrum reveals five additional peaks that are only weakly apparent in the reference specimen at signal-to-noise levels too low to be identified by the deconvolution algorithm. The additional peaks observed from the on-chip samples were likely not detected in the reference specimens due to the low numbers of cells present in the confocal detection volumes, despite extensive scanning to maximize Raman signal from the deposited samples.The minimum number of perfused cells required for effective detection in the nanogap chips is a function of bacteria trapping efficiency, which is largely dependent on cell loss during the concentration process. The performance enhancement observed for the nanogap measurements is believed to result from dense packing of immobilized cells within the Raman detection volume. Initial cell concentrations in thes

留言 (0)

沒有登入
gif