Rapid antibiotic susceptibility test with surface ... - ACS Publications

Subscriber access provided by UNIV AUTONOMA DE COAHUILA UADEC

Article

Rapid antibiotic susceptibility test with surfaceenhanced Raman scattering in a microfluidic system Kai-Wei Chang, Ho-Wen Cheng, Jessie Shiue, Juen-Kai Wang, Yuh-Lin Wang, and Nien-Tsu Huang Anal. Chem., Just Accepted Manuscript • DOI: 10.1021/acs.analchem.9b01027 • Publication Date (Web): 07 Aug 2019 Downloaded from pubs.acs.org on August 8, 2019

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 23 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Analytical Chemistry

Rapid antibiotic susceptibility test with surface-enhanced Raman scattering in a microfluidic system Kai-Wei Chang,1 Ho-Wen Cheng,2 Jessie Shiue,3 Juen-Kai Wang,2,4,5 Yuh-Lin Wang,2,6 Nien-Tsu Huang1,7* 1Graduate

Institute of Biomedical Electronics and Bioinformatics, National Taiwan University, Taipei, Taiwan

2Institute

of Atomic and Molecular Sciences, Academia Sinica, Taipei, Taiwan

3Institute

of Physics, Academia Sinica, Taipei, Taiwan

4Center

for Condensed Matter Sciences, National Taiwan University, Taipei, Taiwan

5Center

for Atomic Initiative for New Materials, National Taiwan University, Taipei, Taiwan

6Department

of Physics, National Taiwan University, Taipei, Taiwan

7Department

of Electrical Engineering, National Taiwan University, Taipei, Taiwan

*Corresponding authors email: [email protected] ABSTRACT Antibiotic susceptibility test (AST) is essential in clinical diagnosis of serious bacterial infection, such as sepsis, while it typically takes 2-5 days for sample culture, antibiotic treatment and reading result. Detecting metabolites secreted from bacteria with surface-enhanced Raman scattering (SERS) enables rapid determination of antibiotic susceptibility, reducing the AST time to 1-2 days. However, it still requires oneday culture time to obtain sufficient quantity of bacteria for sample washing, bacterial extraction, and antibiotic treatment. Additionally, the whole procedure manually performed in open environment often suffers from contamination and human error. To address above problems, a microfluidic system integrating membrane filtration and the SERS-active substrate (MF-SERS) was developed to perform on-chip bacterial enrichment, metabolite collection, and in-situ SERS measurements for antibiotic susceptibility test. Using E. coli as the prototype bacterium, the lowest SERS detection limit of bacterial concentration of the MF-SERS system is 103 CFU/mL, which is four orders of magnitude lower than that using centrifugation-purification procedure, significantly shortening the bacterial culture time. The bacteria and secreted metabolites are enclosed during bacterial trapping, metabolite filtration, and SERS detection, thus minimizing possible contamination and human errors. Finally, the successful demonstration of AST on E. coli with a concentration of 103 CFU/mL is presented. Overall, the MF-SERS system with a miniature size and well-confined microenvironment allows the integration of multiple bacteria processes for bacterial enrichment, culture and determination of AST.

1 ACS Paragon Plus Environment

Analytical Chemistry 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 2 of 23

INTRODUCTION Bacterial identification and characterization reveal meaningful messages in various fields, such as microbial monitoring of water1, 2 or food3 and clinical diagnosis of bacterial pathogen infection. For example, sepsis, a life-threatening disease, would trigger multi-organ failures, extremely low blood pressure and eventually could lead to fatality, is usually caused by bacterial infection in whole blood, lungs, brain, urinary tract, skin, and abdominal organs. Currently, one of the most common treatments for sepsis is the usage of antibiotics.4, 5 To avoid any overuse or misuse of antibiotics, antibiotic susceptibility test (AST) is used to select a proper antibiotic treatment and evaluate its efficacy.6 Although standard AST techniques include disk diffusion, broth dilution, and commercially automated systems,7 are well-developed, due to the low bacteria concentration in human’s clinical samples, e.g. blood, CSF (1-100 CFU/mL), urine (105-106 CFU/mL), a prolonged culture time (1-3 days) is usually required to reach a detectable bacteria level.8,

9

Generally, the whole process

including bacterial culture and antibiotic treatment may take 2-5 days, which are too long for real-time clinical diagnosis and timely treatment. Currently, many researchers have utilized microfluidic techniques to enable rapid bacterial detection and AST. The first example is the imaging-based method. By directly tracking the number or the area of bacterial colonies confined in the microchannel or droplet, the growth rate or morphological change after antibiotic treatment can be characterized.10-14 This method is straightforward without any labeling or surface functionalization. However, the counting may be inaccurate when bacteria aggregated or unevenly distributed. The second example is the antibody-based method, using functionalized pathogen-specific antibodies in the microchannel to capture targeted bacteria followed by the detection of electrochemical sensors15 or microcantilevers.16 The problem of this method is the prolonged and complicated functionalization protocols. The third method is using nucleic acids, including droplet-based single cell platform, digital PCR, digital LAMP methods, to identify bacteria with specific nucleotide sequences.17-21 The problem is the potential false positive and negative result if sequences are not well-designed. In contrast to previous approaches, detection metabolites secretion of bacteria with surface-enhanced Raman scattering (SERS) has demonstrated the ability to determine the susceptibility of antibiotic of both Staphylococcus aureus (S. aureus) and Escherichia coli (E. coli) —dubbed as the SERS-AST method. The 2 ACS Paragon Plus Environment

Page 3 of 23 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


protocol was presented previously22 and is briefed here. After overnight culture, bacteria are treated with antibiotics for two hours, followed by repetitive centrifugation-plus-dilution process to remove the culture medium and antibiotic. The sample is then applied to a SERS substrate for SERS measurement of purine derivatives from bacteria.23, 24 Albeit its success, the SERS-AST method still needs further improvement for clinical approaches. First, it still requires prolonged bacteria sub-culture time for significant SERS detection. Second, the whole protocol involves several manual operations—transferring samples, culturing, washing, antibiotic treatment, etc.—that might introduce human errors. Third, SERS signal from the dried bacterial sample might fluctuated due to uneven bacteria seeding and drying process.25-27 In light of the issues faced in the SERS-AST method, we present a microfluidic device integrated membrane filtration and SERS (MF-SERS) for bacterial enrichment, metabolites extraction, and in-situ SERS measurements. A membrane, attached inside a microchannel to form a chamber, filtrates and concentrates bacteria. Inlets connected with syringe pumps can sequentially introduce bacteria and reagents (including culture media, antibiotics, washing buffers, etc.). After the exclusion of the culture medium, bacteria are incubated in the filtration chamber, and accumulated metabolites released from bacteria are guided into the microchannel attaching a SERS substrate for SERS detection. The closed microenvironment can ensure a stable and uniform adsorption of metabolite on the SERS substrate. Though the operations are not fully automated, manual dispensing procedures are eliminated in this integrated system, and a potential automated system is applicable as reported before.28 Overall, the MF-SERS system could resolve three issues in the previous SERS-AST method: (1) entailing prolonged culture time to obtain a large number of bacteria, (2) human errors and contamination owing to manual operations, and (3) large SERS signal variation due to nonuniform and unstable molecular adhesion on SERS substrate.



Page 4 of 23

MATERIALS AND METHODS Bacteria and antibiotic E. coli (ATCC 25922) was purchased from American Type Culture Collection (ATCC) and E. coli (DH5Alpha) transfected with kanamycin-resistant plasmid was used as the resistant strain, while the ATCC 25922 strain served as the susceptible strain. The antibiotic used was kanamycin (MDBio, Inc.). S. aureus was the standard model HG001 strain. All bacterial samples were cultured in Mueller Hinton Broth (MHB, BD BBL). The porous membrane used in this study was hydrophilic polycarbonate membrane filter with a pore size of 0.22 μm (GTTP01300, Merck Millipore). Both E. coli and S. aureus transfected with green fluorescent protein (GFP) were used in the verification of the filtration performance. Bacterial growth and preparation To multiply microbial organisms, single colonies of both the susceptible and resistant strains cultured on agar plate were collected using sterile plastic inoculating loops. Bacteria were then suspended in 5 mL of MHB and incubated in an orbital shaking incubator operated at 200 rpm and 37°C for 16 hours. Then, 0.2 mL of the cultured sample was added into 5 mL of MHB and sub-cultured in the same condition until OD600 reached ~0.5, which represented the concentration of ~108 CFU/mL. Lower bacteria concentration was acquired from serial dilution with MHB and verified with double grid counting slides (Shih-Yung medical instruments). During the antibiotic treatment, the bacteria were co-cultured with the antibiotic during the sub-culture step. All bacteria samples were prepared inside the laminar flow cabinet at room temperature and followed proper biosafety regulations. Centrifugal purification The collection procedure of bacterial metabolite that was used in the performance of SERS detection was reported previously.22 Briefly, 1 mL bacteria sample was centrifuged at 10000g for 2 minutes at room temperature. Then, the 0.9-mL supernatant was replaced by the same amount of deionized (DI) water. The above steps were repeated twice again to dilute the concentration of the culture medium by 1000 times. After a 20-min incubation, the sample was centrifuged under the same condition once again. The supernatant, mainly composed of secreted metabolites, was then collected for SERS measurements. Microfluidic device fabrication and system setup 4 ACS Paragon Plus Environment

Page 5 of 23 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


The microfluidic device, including the filtration and detection zone, was constructed with three poly(methyl methacrylate) (PMMA) components fabricated by computer numerical control (CNC) machining (Figure 1A). In the filtration zone, a porous membrane filter was trimmed to be 3 mm2 and clamped by two PMMA components using UV glue. The structure could sustain a fluidic pressure of ~100 kPa for a sample injection rate of 2 mL/min. A SERS-active substrate with a sensing area of 2.51 mm2 was attached with the doublesided tape on the top side of the detection zone, a microchannel of 1010.4 mm3. The substrate was fabricated by silver evaporation (EB-EVA-613, I Shien) at a rate of 0.03 nm/sec onto an acetone and sodium hydroxide pre-cleaned glass slide (Figure 1B), forming a silver-island film with an average thickness of 12 nm. The filtration and detection zones were bonded by the double-sided tape with the dimension of 211 cm3 (Figure 1C). To prevent any solution leakage, UV glue was used to seal the edges of substrate attached to the microchannel. Inlets and outlets were connected with syringes through Luer lock needles. The needle for the injection of bacteria and DI water was connected to two separated syringes via a three-way value. The collected metabolites were drawn into the detection zone with another syringe. Microfluidic filtration and incubation The operation of the MF-SERS system is illustrated in Figure 2. It can be divided into four steps: (1) filtration, (2) washing, (3) incubation, and (4) detection. First, the bacterial solution was injected into the filtration zone at 2 mL/min and bacteria were trapped on the membrane filter. Then, trapped bacteria were subsequently washed with deionized water to remove the residual culture medium at 2 mL/min for one minute. In the first two steps, the metabolite outlet was blocked by a three-way valve to prevent the liquid flow into the detection zone. In step 3, 200 µL of air was pumped from the metabolite outlet to completely remove the liquid on the bottom side of the membrane. Therefore, a 1µL reaction volume on the top side of the membrane with trapped bacteria can be prepared. At this step, there was no reverse flow out of the sample inlet due to a much higher flow resistance of the membrane filter than the waste outlet. After 20-minute incubation at room temperature, the solution containing secreted metabolites flowed through the membrane filter to the detection zone at 1 μL/min, while the waster outlet was blocked, leaving bacteria on the membrane filter. The total operational time was 30 minutes. Trapping of bacteria on membrane filter 5 ACS Paragon Plus Environment


Page 6 of 23

To optimize the filtration performance, GFP-transfected E. coli trapped on the membrane were counted under a fluorescence microscope (IX-73, Olympus) with a 10 objective lens. A sequence of images, recorded by a dual-color charge-coupled device (CCD) (DP-80, Olympus) with an integration time of 0.5 second, were obtained by raster-scanning the sample with use a motorized stage (MFD-2, Mashausor). The acquired images were then stitched with a microscopic analysis software (CellSens, Olympus), yielding a combined image with area of 1.51.5 mm2 covering the whole membrane filter. The bacteria on the filter was counted by a customized software based on ImageJ. (see Figure S1 in Supporting Information for more details). SERS measurement and spectral processing SERS spectra were measured using a commercial Raman microscope (SuperHead HE 633, Horiba Jobin Yvon) with a 632.8-nm HeNe laser as the excitation source. The laser beam was band-pass filtered to remove residual plasma lines and was focused by a 20 objective lens to the SERS-active substrate of the MF-SERS system. The system was placed on top of a x-y motorized stage (EK 32, Märzhäuser) to allow the irradiation laser spot to move to different spots on the SERS substrate. The scattered radiation was collected using the same objective lens, filtered through a Raman long-pass filter, and sent to an 80-cm spectrograph (1200 gr/mm) with a liquid nitrogen-cooled CCD. The resultant spectral resolution and error were 4 and 0.1 cm-1, respectively. The laser irradiation power was about 5 mW with 0.1-second integration time. Typically, 9-16 SERS spots were measured for the same sample on single substrate. The SERS spectra were processed with a baseline removal program, based on the nonlinear iterative peak clipping algorithm proposed by Miroslav’s group,29 and then averaged to obtain the mean spectrum and standard deviation. To evaluate the antibiotic susceptibility of susceptible and resistant strain, we defined the signal ratio at 740 cm-1 - namely, 𝑟740 = 𝐼740, 𝑡𝑟𝑒𝑎𝑡𝑒𝑑

𝐼740, 𝑛𝑜𝑡 𝑡𝑟𝑒𝑎𝑡𝑒𝑑,

where I740, treated and I740, not treated is the signal of the Raman peak at 740 cm-1 of bacterial sample

with and without antibiotic treatment, respectively. The standard deviations of 𝑟740 is defined as 𝛿(𝑟740) = 𝑟740 × (

𝛿(𝐼740, 𝑡𝑟𝑒𝑎𝑡𝑒𝑑) 2 𝐼740, 𝑡𝑟𝑒𝑎𝑡𝑒𝑑

) +(

𝛿(𝐼740, 𝑛𝑜𝑡 𝑡𝑟𝑒𝑎𝑡𝑒𝑑) 2 𝐼740, 𝑛𝑜𝑡 𝑡𝑟𝑒𝑎𝑡𝑒𝑑

) . Above parameters are defined based on our previous study.22


Page 7 of 23 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


RESULTS AND DISCUSSION Filtration efficiency The trapping efficiency of the membrane in the MF-SERS system was first determined by counting the GFPtransfected E. coli in obtained fluorescence images. The efficiencies of total 18 combinations of sample volume (VS) and initial concentration (Cin) of the E. coli samples were obtained. The tests of the same bacterial concentration with different volumes determine the dependence of the filtration on the process time, while the test of the same volume with different bacterial concentrations clarify the sensitivity of the filtration to the accumulated cell amount. Figure 3A exemplifies 5 combinations: (1) 10 mL, 6.7103 CFU/mL; (2) 5 mL, 8.1103 CFU/mL; (3) 1 mL, 6.7103 CFU/mL; (4) 1 mL, 6.7104 CFU/mL; (5) 1 mL, 1.3106 CFU/mL. The trapped bacteria numbers (Ntrap) were 5.0104, 3.4104, 5.9103, 5.0104 and 1.0106, respectively, indicating that trapped bacteria show positive correlation with both initial concentration and sample volume. The correlation plot between Ntrap and Cin of 1, 5 and 10 mL is shown in Figure 3B. It is worthwhile to mention that data from three sample volumes plotted in the log-log plot all follow approximately linear lines, indicating that the trapping performance is independent of Cin and seemingly insensitive to the sample volume. Accordingly, the trapping efficiencies (𝜂trap = 𝑁trap (𝐶in × 𝑉S)) calculated are shown in Figure 3C. Both data of 1 and 5 mL are located roughly between 75% to 100%, while those of 10 mL ostensibly decrease with Cin. This result indicates that 𝜂trap of both the sample volumes of 1 and 5 mL is about the same while the trapping performance would degrade or varied in time slightly. The possible reason may be due to bacteria aggregation or membrane deformation, which make the bacteria counting less accurate. Besides successful trapping of the rod-shaped E. coli with the dimension of 2 μm long and 1 μm in diameter, we further check if the membrane filtration can be applied in other pathogens. Here, we choose the cocci-shaped S. aureus with the dimension of 500 nm ~ 1 μm in diameter, which is also the most representative gram-positive bacteria strain. The same filtration experiment was repeated using GFP-transfected S. aureus. As shown in Figure S2C, the filtration efficiencies of S. aureus are between 70% to 100%, indicating the similar trapping efficiency. The results demonstrate both gram-positive and gram-negative bacteria strains can be efficiently trapped using the membrane filtration. Removing residual culture medium 7 ACS Paragon Plus Environment


Page 8 of 23

Both adenine and hypoxanthine are main constituents released from most pathogens (including E. coli, E. faecium, E. faecalis, S. pneumonia, P. aeruginosa, and S. aureus, etc.).20 The corresponded Raman signal strength at 740 cm-1 reflects the amount of two molecules released from bacteria. For other rare pathogens, such as A. baumannii, the major constituents are xanthine and guanine, which are at 660 cm-1. Since a large quantity of purines and derivatives in MHB may cause signal interference (Figure S3), it is necessary to remove the residual culture medium before the incubation step during the SERS detection.20,21 The removal efficiency of the remaining culture medium during the washing step, shown in Figure 2, therefore needs to be verified. Here, we continuously injected deionized water to flush the filtration chamber. Every 0.5 mL washout sample was collected for analysis. As shown in Figure S4, after 1.5 mL of deionized water deployed, the signal strength at 740 cm-1, I740, decreases to the same signal level of deionized water and 1000 diluted culture medium (the dilution requirement of our previous SERS-AST method). The detailed signal comparison of each washing volume and 1000 dilution is shown in Supporting Information. Spatial uniformity of SERS detection Since the variation is intrinsic in the SERS detection, owing to the inhomogeneity of SERS substrate in sample coverage and in electromagnetic enhancement, multiple measurements on different spots on single SERS substrate are required to obtain statistically meaningful results. Figure S5 compared the signal strengths of I740 from the SERS spectra of adenine with concentrations of 10-6 and 10-7 M, respectively, inside microchannel or just dried on the SERS substrate. The result shows a less than 10% signal variance in microchannel case, compared to over 30% signal variance in dried case (see Supporting Information for more details). Such variance is mainly resultant from the “coffee ring effect”—the ring-like pattern caused by the replenishment of the liquid at the fast-evaporating edge by the liquid from the interior. The uniform signal strength in the microchannel is due to the equilibrated adsorption-desorption process of the dissolved molecules on the surface of the SERS substrate. Accordingly, the coverage of the adsorbed molecules would be stable without the concern of unsure drying condition in dried droplets. Furthermore, a large uniform sensing area can potentially enable multi-parallel or multiplex SERS measurements. SERS signal vs. bacterial concentration


Page 9 of 23 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


After examining filtration performance, washing protocol, and the uniformity of SERS detection, the dependence of SERS signal of released metabolites on the bacterial concentration was evaluated. Namely, 1 and 10 mL of sub-cultured E. coli (ATCC 25922) in MHB at different concentrations were loaded into the filtration chamber so that they were trapped on the membrane filter. 1.5 mL of deionized water was then injected to the chamber to force the culture medium to pass through the filter. The trapped bacteria were then release metabolites in the chamber filled with deionized water for 20 minutes. The metabolite solution in the chamber was subsequently compelled through the filter and guided into the SERS detection zone for SERS measurements. Figure 4A and B show the SERS spectra of two sample volumes. As a comparison, Figure 4C shows the SERS spectra of the dried droplets of the supernatant extracted after centrifugal purification process of bacterial concentrations ranging from 103 to 106 CFU/mL. The signal strengths of I740 of all measured SERS spectra were extracted and plotted as a function of bacterial concentration, shown in Figure 4D. In each bacterial concentration condition, the standard deviations were obtained from three separate samples in three devices. Here, three qualitative findings can be summarized. First, at one initial bacterial concentration, both concentrated-bacterium cases give significantly higher I740 values than the centrifugation-purification case. Furthermore, the limit of detection (LOD) of 1- and 10-mL concentrated-bacterium cases are around 104 and 103 CFU/mL, respectively, while that of the centrifugation-purification case is about 107 CFU/mL. Second, at the same I740, the corresponding concentrations of 1- and 10-mL concentrated-bacterium cases differ approximately by a factor of ten. This feature agrees with the one shown in the trapping performance of Figure 3B—namely, the trapped bacterial number is proportional to the product of the initial bacterial concentration, Cin, and the sample volume, VS. Third, I740 increases sub-linearly with the bacterial concentration in three cases in their respective concentration ranges, indicating that the Langmuir-type isotherm behavior30,

31

governs the adsorption-desorption characteristics of the metabolites on the SERS substrate. Overall, above results indicate that the bacterial enrichment scheme (bacteria trapping and confinement in the small volume chamber) in the MF-SERS system can greatly reduce the required bacterial concentration for SERS detection by three to four orders of magnitude. Antibiotic susceptibility test



Page 10 of 23

Next, we perform a rapid antibiotic susceptibility test using the MF-SERS system. As proof of concept, susceptible and resistant E. coli of high concentration (107 CFU/mL) in MHB was prepared using centrifugal purification and then treated with kanamycin of 16 μg/mL for 0.5, 1, 2 and 4 hours separately. The antibiotic concentration (16 μg/mL) was chosen according to CLSI M100-S27. Based on the CLSI report, the minimum inhibitory concentration (MIC) of the susceptible strain must be lower than 16 μg/mL. The same E. coli was cultured for the same time without kanamycin as the comparison. The AST experiments performed in the MFSERS system were conducted in a staggered manner for different samples. Therefore, this concentration would ensure effective inhibition on the susceptible strain. As shown in Figure S6A, the signal strength with the antibiotic treatment (red curves) did not show a distinct change when time increased, indicating the bacteria growth rate was inhibited, whereas the strength of susceptible strain without antibiotic treatment (blue curves) slightly increased along the treatment time. For the resistant strain (Figure S6B), the signal strengths with and without antibiotic treatment both showed a distinct signal increase along the treatment time, indicating the antibiotic had almost no effect on the resistant strain. The ratios between the signal strengths of I740 with and without antibiotic treatment, r740, are shown in Figure S6C. In each treatment time condition, the standard deviations were obtained from three separate measurements of single bacteria sample in one device. The ratio of the susceptible strain dropped significantly after 2-hour treatment; therefore, at least 2 hours are required for antibiotic treatment. The detailed signal comparison of two strains is shown in Supporting Information. To sum up, above off-chip AST confirmed that SERS measurements could successfully distinguish the resistant and susceptible strain under 2-hour antibiotic treatment. Finally, we conduct multiple on-chip AST experiments using the MF-SERS system. In each set of experiments, 10 mL of E. coli samples were injected with a much lower bacteria concentration (103 mL-1) compared to the off-chip case (107 mL-1) and treated with antibiotic at different concentrations (0, 16, 32, 64 μg/mL). As shown in Figure 5, I740 of the susceptible strain with the antibiotic treatment is considerably smaller than that without the treatment (Figure 5A) , while I740 of the resistant strain with the antibiotic treatment is comparable to that without the treatment (Figure 5B). Figure 5C shows the comparison of r740 between the susceptible and resistant strain. The r740 of the susceptible strain treated for different antibiotic concentrations are 0.24, 0.26, 0.25, respectively, which is significantly smaller those of the resistant strain (0.95, 1.04, 1.02). 10 ACS Paragon Plus Environment

Page 11 of 23 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


In each antibiotic concentration treatment condition, the standard deviations were obtained from three separate samples in three devices. Overall, this result confirms that the system can perform AST of E. coli with concentration of four-order magnitude lower than previous SERS-AST method, which can greatly shorten the culture time to only 5-10 hours (assume 1-103 CFU/mL initial bacterial concentration).



Page 12 of 23

CONCLUSIONS A microfluidic system integrating membrane filtration and SERS-active substrate (MF-SERS) was developed to perform on-chip bacterial enrichment, washing, and in-situ SERS measurements of the released metabolites. The system holds three important features. First, the lowest SERS detection limit of the initial concentration of E. coli is demonstrated to be 103 CFU/mL, thus enabling a shorter culture time. Second, the fluidic unit integration of MF-SERS system can reduce multiple manual operations and prevent any potential human error and contamination. Third, the system encloses the sample liquid to enable a stable and uniform adsorptiondesorption equilibrium on the SERS substrate while performing SERS measurement, therefore decreasing SERS signal variation. In the future, multiple MF-SERS systems can potentially be assembled in parallel for determination of AST and MIC of different antibiotics. Another possible improvement is to integrate microfluidic-based bacteria isolation techniques utilizing osmotic shock32, inertial flow33, dielectrophoresis,34 and acoustic trapping,35 with the MF-SERS system to perform AST in clinical samples (e.g. whole blood or urine) applications. ASSOCIATED CONTENT Supporting Information The Supporting Information is available free of charge on the ACS Publications website. Additional experimental details including, image processing algorithm for cell counting on the membrane, bacterial trapping performance of the membrane using GFP-transfected S. aureus, SERS spectra of E. coli metabolites, MHB, Adenine, Hypoxanthine, Kanamycin and DI water, washing efficiency of the MF-SERS system for the culture medium, the comparison of spatial uniformity of SERS detection between molecules inside the microchannel and dried droplet on the SERS substrate, the antibiotic susceptibility test of E. coli for kanamycin using centrifugal purification. AUTHOR INFORMATION Corresponding Authors: *E-mail: [email protected] ORCID: Nien-Tsu Huang: 0000-0002-2569-805X Notes: The authors declare no competing financial interest.


Page 13 of 23 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


ACKNOWLEDGEMENT This work was supported by the Ministry of Science and Technology, Taiwan under grant “MOST 106-2221E-002-058-MY3” and “MOST 106-2745-M-001-004-ASP”. We are thankful to Dr. Li-Kwan Chang’s lab for support of GFP-transfected E. coli, Dr. Mei-Hui Lin’s lab for support of GFP-transfected S. aureus, Mr. KuoKai Hsu for support of kanamycin-resistance transfected E. coli, Dr. Chi-Hung Lin’s lab for providing kanamycin and Dr. Yuh-Lin Wang’s lab for facility support.



TABLE OF CONTENT


Page 14 of 23

Page 15 of 23 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Figure 1 The MF-SERS system. (A) The schematic diagram of the microfluidic system and device; (B) SEM image of SERS substrate. (scale bar: 200 nm); (C) The photo of the microfluidic device.



Page 16 of 23

Figure 2 Operational procedure of the microfluidic device: (1) Filtration—bacteria (orange ovals) is introduced from the inlet and blocked on the membrane, (2) Washing— the culture medium (yellow) is washed away by DI water (blue); (3) Incubation—the bacteria in the chamber are incubated and their metabolites (red) are released, and (4) Detection—the secreted metabolites are guided to the SERS-active substrate (orange, rectangle) for detection. To prevent the liquid flowed into undesired channel, the outlet was connected to the three-way valve to blocked the flow.


Page 17 of 23 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Figure 3 Bacterial trapping performance of membrane filter. (A) Five initial concentration (Cin) -sample volume (VS) combinations and their resultant fluorescence images and trapped cell numbers (Ntrap) of GFPtransfected E. coli; (B) Ntrap vs. Cin and (C) trapping efficiency (𝜂 = 𝑁trap (𝐶 × 𝑉 )) vs. Cin for all 18 Cintrap

in

S

VS combinations. The scale bar in (A) is 200 μm. The vertical yellow highlight in (B) indicates Cin in the range of 6-8103 CFU/mL. The data of the five exemplified cases according to the numbers indicated in (A) are marked in (B). The data of the sample volumes of 1, 5 and 10 mL in (B) and (C) are presented as red squares, green triangles and blue circles, respectively.



Page 18 of 23

Figure 4 SERS spectra of secreted metabolites of E. coli of different cell concentrations with sample filtration volumes of (A) 1 mL and (B) 10 mL acquired with the MF-SERS system; (C) those acquired with centrifugal purification procedure with a bacterial sample volume of 1 mL; (D) signal strengths at 740 cm-1 (highlighted with yellow in the SERS spectra) plotted as a function of bacterial concentration of the above (A), (B) and (C) situations (represented as red squares, blue circles and violet triangles, respectively).


Page 19 of 23 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Figure 5 Antibiotic susceptibility test results of E. coli treated with kanamycin for 2 hours in the MF-SERS system: (A) SERS spectra of susceptible strain treated with the antibiotic of 0, 16, 32, 64 µg/mL; (B) SERS spectra of resistant strain treated with the antibiotic of 0, 16, 32, 64 µg/mL; Red, orange, green and blue curves in (A) and (B) represent SERS spectra, while light curves represent to their corresponding standard deviations. (C)(D) The signal ratio (𝑟740) of the signal strength at 740 cm-1 of (C) susceptible strain and (D) resistant strain.



Page 20 of 23

Reference 1. Noble, R. T.; Weisberg, S. B., A review of technologies for rapid detection of bacteria in recreational waters. Journal of Water and Health 2005, 3 (4), 381-392. 2. McCoy, W. F.; Rosenblatt, A. A., HACCP-Based Programs for Preventing Disease and Injury from Premise Plumbing: A Building Consensus. Pathogens 2015, 4 (3), 513-28. 3. Gracias, K. S.; McKillip, J. L., A review of conventional detection and enumeration methods for pathogenic bacteria in food. Can J Microbiol 2004, 50 (11), 883-90. 4. Abubakar, I.; Irvine, L.; Aldus, C. F.; Wyatt, G. M.; Fordham, R.; Schelenz, S.; Shepstone, L.; Howe, A.; Peck, M.; Hunter, P. R., A systematic review of the clinical, public health and cost-effectiveness of rapid diagnostic tests for the detection and identification of bacterial intestinal pathogens in faeces and food. Health Technol Assess 2007, 11 (36), 1-216. 5. Ahmed, A.; Rushworth, J. V.; Hirst, N. A.; Millner, P. A., Biosensors for whole-cell bacterial detection. Clinical microbiology reviews 2014, 27 (3), 631-646. 6. Ventola, C. L., The Antibiotic Resistance Crisis: Part 1: Causes and Threats. Pharmacy and Therapeutics 2015, 40 (4), 277-83. 7. Pulido, M. R.; García-Quintanilla, M.; Martín-Peña, R.; Cisneros, J. M.; McConnell, M. J., Progress on the development of rapid methods for antimicrobial susceptibility testing. Journal of Antimicrobial Chemotherapy 2013, 68 (12), 2710-2717. 8. Wang, S.; Inci, F.; L Chaunzwa, T.; Ramanujam, A.; Vasudevan, A.; Subramanian, S.; Chi Fai Ip, A.; Sridharan, B.; Gurkan, U.; Demirci, U., Portable microfluidic chip for detection of Escherichia coli in produce and blood. 2012; Vol. 7, p 2591-600. 9. Opota, O.; Croxatto, A.; Prod'hom, G.; Greub, G., Blood culture-based diagnosis of bacteraemia: state of the art. Clin Microbiol Infect 2015, 21 (4), 313-22. 10. Choi, J.; Jung, Y.-G.; Kim, J.; Kim, S.; Jung, Y.; Na, H.; Kwon, S., Rapid antibiotic susceptibility testing by tracking single cell growth in a microfluidic agarose channel system. Lab Chip 2013, 13 (2), 280287. 11. Kim, S. C.; Cestellos-Blanco, S.; Inoue, K.; Zare, R. N., Miniaturized Antimicrobial Susceptibility Test by Combining Concentration Gradient Generation and Rapid Cell Culturing. Antibiotics (Basel) 2015, 4 (4), 455-66. 12. Matsumoto, Y.; Sakakihara, S.; Grushnikov, A.; Kikuchi, K.; Noji, H.; Yamaguchi, A.; Iino, R.; Yagi, Y.; Nishino, K., A Microfluidic Channel Method for Rapid Drug-Susceptibility Testing of Pseudomonas aeruginosa. PLoS One 2016, 11 (2), e0148797. 13. Choi, J.; Jeong, H. Y.; Lee, G. Y.; Han, S.; Han, S.; Jin, B.; Lim, T.; Kim, S.; Kim, D. Y.; Kim, H. C.; Kim, E. C.; Song, S. H.; Kim, T. S.; Kwon, S., Direct, rapid antimicrobial susceptibility test from positive blood cultures based on microscopic imaging analysis. Sci Rep 2017, 7 (1), 1148. 14. Sabhachandani, P.; Sarkar, S.; C. Zucchi, P.; A. Whitfield, B.; E. Kirby, J.; Hirsch, E.; Konry, T., Integrated microfluidic platform for rapid antimicrobial susceptibility testing and bacterial growth analysis using bead-based biosensor via fluorescence imaging. 2017. 15. Yang, L.; Bashir, R., Electrical/electrochemical impedance for rapid detection of foodborne pathogenic bacteria. Biotechnol Adv 2008, 26 (2), 135-50. 20 ACS Paragon Plus Environment

Page 21 of 23 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


16. Etayash, H.; Khan, M. F.; Kaur, K.; Thundat, T., Microfluidic cantilever detects bacteria and measures their susceptibility to antibiotics in small confined volumes. Nat Commun 2016, 7, 12947. 17. Safavieh, M.; Ahmed, M. U.; Ng, A.; Zourob, M., High-throughput real-time electrochemical monitoring of LAMP for pathogenic bacteria detection. Biosens Bioelectron 2014, 58, 101-6. 18. Mezger, A.; Gullberg, E.; Goransson, J.; Zorzet, A.; Herthnek, D.; Tano, E.; Nilsson, M.; Andersson, D. I., A general method for rapid determination of antibiotic susceptibility and species in bacterial infections. J Clin Microbiol 2015, 53 (2), 425-32. 19. Oh, S. J.; Park, B. H.; Jung, J. H.; Choi, G.; Lee, D. C.; Kim, D. H.; Seo, T. S., Centrifugal loopmediated isothermal amplification microdevice for rapid, multiplex and colorimetric foodborne pathogen detection. Biosens Bioelectron 2016, 75, 293-300. 20. Schoepp, N. G.; Khorosheva, E. M.; Schlappi, T. S.; Curtis, M. S.; Humphries, R. M.; Hindler, J. A.; Ismagilov, R. F., Digital Quantification of DNA Replication and Chromosome Segregation Enables Determination of Antimicrobial Susceptibility after only 15 Minutes of Antibiotic Exposure. Angewandte Chemie International Edition 2016, 55 (33), 9557-9561. 21. Schoepp, N. G.; Schlappi, T. S.; Curtis, M. S.; Butkovich, S. S.; Miller, S.; Humphries, R. M.; Ismagilov, R. F., Rapid pathogen-specific phenotypic antibiotic susceptibility testing using digital LAMP quantification in clinical samples. Science Translational Medicine 2017, 9 (410), eaal3693. 22. Liu, C.-Y.; Han, Y.-Y.; Shih, P.-H.; Lian, W.-N.; Wang, H.-H.; Lin, C.-H.; Hsueh, P.-R.; Wang, J.-K.; Wang, Y.-L., Rapid bacterial antibiotic susceptibility test based on simple surface-enhanced Raman spectroscopic biomarkers. Scientific Reports 2016, 6, 23375. 23. Premasiri, W. R.; Lee, J. C.; Sauer-Budge, A.; Théberge, R.; Costello, C. E.; Ziegler, L. D., The biochemical origins of the surface-enhanced Raman spectra of bacteria: a metabolomics profiling by SERS. Analytical and Bioanalytical Chemistry 2016, 408 (17), 4631-4647. 24. Chiu, S. W.; Cheng, H. W.; Chen, Z. X.; Wang, H. H.; Lai, M. Y.; Wang, J. K.; Wang, Y. L., Quantification of biomolecules responsible for biomarkers in the surface-enhanced Raman spectra of bacteria using liquid chromatography-mass spectrometry. Phys Chem Chem Phys 2018, 20 (12), 8032-8041. 25. Yunker, P. J.; Still, T.; Lohr, M. A.; Yodh, A. G., Suppression of the coffee-ring effect by shapedependent capillary interactions. Nature 2011, 476 (7360), 308-11. 26. Sempels, W.; De Dier, R.; Mizuno, H.; Hofkens, J.; Vermant, J., Auto-production of biosurfactants reverses the coffee ring effect in a bacterial system. Nature Communications 2013, 4, 1757. 27. Sempels, W.; De Dier, R.; Mizuno, H.; Hofkens, J.; Vermant, J., Auto-production of biosurfactants reverses the coffee ring effect in a bacterial system. Nat Commun 2013, 4, 1757. 28. Huang, S. H.; Chang, Y. S.; Juang, J. J.; Chang, K. W.; Tsai, M. H.; Lu, T. P.; Lai, L. C.; Chuang, E. Y.; Huang, N. T., An automated microfluidic DNA microarray platform for genetic variant detection in inherited arrhythmic diseases. Analyst 2018, 143 (6), 1367-1377. 29. Morháč, M.; Matoušek, V., Peak Clipping Algorithms for Background Estimation in Spectroscopic Data. Applied Spectroscopy 2008, 62 (1), 91-106. 30. de Oliveira, K. V.; Rubim, J. C., Surface-enhanced Raman spectroscopy of molecules adsorbed on silver nanoparticles dispersed an agarose gel and their adsorption isotherms. Vibrational Spectroscopy 2016, 86, 290-301. 21 ACS Paragon Plus Environment


Page 22 of 23

31. Altun, A. O.; Bond, T.; Pronk, W.; Park, H. G., Sensitive Detection of Competitive Molecular Adsorption by Surface-Enhanced Raman Spectroscopy. Langmuir 2017, 33 (28), 6999-7006. 32. Zelenin, S.; Hansson, J.; Ardabili, S.; Ramachandraiah, H.; Brismar, H.; Russom, A., Microfluidicbased isolation of bacteria from whole blood for sepsis diagnostics. Biotechnol Lett 2015, 37 (4), 825-30. 33. Hou, H. W.; Bhattacharyya, R. P.; Hung, D. T.; Han, J., Direct detection and drug-resistance profiling of bacteremias using inertial microfluidics. Lab Chip 2015, 15 (10), 2297-307. 34. D'Amico, L.; Ajami, N. J.; Adachi, J. A.; Gascoyne, P. R.; Petrosino, J. F., Isolation and concentration of bacteria from blood using microfluidic membraneless dialysis and dielectrophoresis. Lab Chip 2017, 17 (7), 1340-1348. 35. Ohlsson, P.; Evander, M.; Petersson, K.; Mellhammar, L.; Lehmusvuori, A.; Karhunen, U.; Soikkeli, M.; Seppa, T.; Tuunainen, E.; Spangar, A.; von Lode, P.; Rantakokko-Jalava, K.; Otto, G.; Scheding, S.; Soukka, T.; Wittfooth, S.; Laurell, T., Integrated Acoustic Separation, Enrichment, and Microchip Polymerase Chain Reaction Detection of Bacteria from Blood for Rapid Sepsis Diagnostics. Anal Chem 2016, 88 (19), 9403-9411.


Page 23 of 23 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


TOC

ACS Paragon Plus Environment

Rapid antibiotic susceptibility test with surface ... - ACS Publications

Recommend Documents