Identification of Newly Emerging Influenza Viruses by Detecting the

Subscriber access provided by WEBSTER UNIV

Article

Identification of newly emerging influenza viruses by detecting the virally infected cells based on surface enhanced Raman spectroscopy (SERS) and principal component analysis (PCA) Jae-young Lim, Jung-soo Nam, Hyunku Shin, Jaena Park, Hyein Song, Minsung Kang, Kwang-il Lim, and Yeonho Choi Anal. Chem., Just Accepted Manuscript • DOI: 10.1021/acs.analchem.8b05533 • Publication Date (Web): 04 Mar 2019 Downloaded from http://pubs.acs.org on March 7, 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

Identification of newly emerging influenza viruses by detecting the virally infected cells based on surface enhanced Raman spectroscopy (SERS) and principal component analysis (PCA) Jae-young Lim†, Jung-soo Nam‡, Hyunku Shin†, Jaena Park†, Hye-in Song§, Minsung Kang†, Kwang-il Lim‡,§,*, and Yeonho Choi†,∥,* †Department

of Bio-convergence Engineering, Korea University, Seoul, 02841, South Korea. of Medical & Pharmaceutical Sciences, Sookmyung Women’s University, Seoul, 04310, South Korea. §Department of Chemical and Biological Engineering, Sookmyung Women’s University, Seoul, 04310, South Korea. ∥School of Biomedical Engineering, Korea University, Seoul, 02841, South Korea. ‡Department

E-mail: [email protected], [email protected] Tel.: +82-2-2077-7627, +82-2-3290-5657

Abstract Rapid diagnosis and quarantine of influenza virus mutants-infected people is critical to contain the fatal viral infection spread because effective antiviral drugs are normally not available. Conventional methods, however, cannot be used for the diagnosis because these methods need pre-defined labels, likely also unavailable for just emerging viruses. Here, we propose labelfree identification of cells infected with different influenza viruses based on surface-enhanced Raman spectroscopy (SERS) and principal component analysis (PCA). Viral envelope proteins that are displayed on the surface of cells after infection of influenza viruses were targeted for this identification. Cells that expressed the envelope proteins of A/WSN/33 H1N1 or A/California/04/2009 H1N1 influenza viruses produced distinct SERS signals. Cells that displayed combinations of the envelope proteins from these two viral variants, an indication of emergence of a new virus, generated also characteristic SERS patterns. However, cell’s own surface proteins often hindered the identification of virally infected cells by producing SERS peaks similar to viral ones. PCA of the obtained SERS patterns could effectively capture the virus-specific signal components from the jumbled SERS peaks. Our study demonstrates a

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

potential of combination of SERS and PCA to identify newly emerging influenza viruses through sensing the cells infected with the viruses. Keywords: infected cells, surface-enhanced Raman spectroscopy, principal component analysis, genome-assorted virus mutants, cellular surface proteins

Influenza virus infections occur worldwide every year.1-3 Influenza viruses easily and quickly evolve not only by point mutations but also by shuffling of segmented genomes, and thereby new variants keep emerging. Furthermore, new influenza virus variants can be fatal as exemplified in the recent pandemics.4-6 New virus infections can spread widely in a short time period; thus, rapid diagnosis of influenza virus infection is critical to slow down the viral spread by quarantining infected people since effective antiviral drugs are normally unavailable. However, methods for rapid detection of new influenza viruses are not available either. Two conventional methods for detection of influenza viruses are based on ELISA and RT-qPCR.7-11 These methods require probe molecules specific to viruses, such as antibodies and DNA oligomers. For this reason, the conventional methods are not suitable for rapid detection of new influenza viruses because construction of probe molecules for new viruses requires much time. In addition, these conventional diagnostic methods involve complicated labor-intensive steps and it may take several days to obtain the final results. In this study, we developed a novel label-free method based on surface enhanced Raman spectroscopy (SERS) to detect new influenza virus infections. Many researchers have demonstrated SERS approaches for detection of cells.12-14 Based on its potential to detect cells, the SERS method can be applied to identify influenza virus-infected cells instead of viral particles. We firstly obtained the surface-enhanced Raman spectra from infected cells without use of any probe molecules and performed subsequent principal component analysis (PCA) for


Page 2 of 23

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


the obtained spectra data. We especially target viral envelope proteins displayed on the surface of infected cells as key marker molecules for diagnosis. Influenza virus is an enveloped virus of 80–120 nm in diameter with segmented negativesense RNA genomes. This virus has hemagglutinin (HA) and neuraminidase (NA) as main envelope proteins, on its surface. When influenza virus infects cells, these viral envelope proteins are presented on the cell surface, later leading to budding of progeny virus particles.15-17 At the early stage of infection each infected cell can effectively amplify the viral components including viral envelope proteins. As immune systems target the envelope proteins that are mostly displayed on the cell surface to identify the infected cells and attack them, we can also target the viral envelope proteins on cells for diagnosis of virus. Influenza viruses can rapidly evolve via superinfection of host cells, which involves entry of multiple different virus strains into the same cell, to finally generate genome-assorted mutants. Emergence of such mutants can be also detected by monitoring the surface of cells displaying the envelope proteins of the multiple virus strains. We supposed that early detection of influenza viruses is possible if we can detect HA and NA proteins on the infected cell surface through SERS. Our SERS-based method captures enhanced Raman signals from the interface between gold nanoparticles (GNPs) and HA and/or NA proteins on virally infected cells (Figure 1). This method can detect the emergence of new influenza viruses that may have evolved by reassortment of segmented genomes of two different strains by simultaneously sensing HA and NA proteins of the two strains expressed on the infected cells.18-21 Infected cells display not only viral envelope proteins but also their own surface proteins. These cellular proteins may produce SERS signals similar, in part, to those of viral proteins, consequently hindering the identification of the type of virus that infected cells. To overcome this difficulty, we applied a statistical method, PCA, which allows comparison of Raman spectra as a whole, between cells infected by different viruses. With PCA we could systematically extract the key features of Raman spectra corresponding to each type of virus 3 ACS Paragon Plus Environment


and finally successfully distinguish cells infected by different influenza virus strains (Figure 1).22-24

Experimental section Cell culture and transfection. Human embryonic kidney (HEK) 293T cells were grown in Iscove’s modified Dulbecco’s medium (Gibco Life Technologies, Carlsbad, CA, USA) supplemented with 10 % fetal bovine serum (FBS) at 37 °C and 5 % CO2. To mimic virally infected cells, we introduced plasmid DNAs that encode envelope proteins (NA and HA) of influenza viruses into HEK 293T cells via a calcium-phosphate-based transfection method. The cells were seeded in 10-cm dishes or multi-well plates. At 24 hours post-seeding cells were transfected with plasmids each containing the HA or NA gene of the influenza WSN strain or CAL strain under CMV promoter. At 24 hours post-transfection, culture media containing plasmids were replaced with fresh ones. Analysis of viral protein expression in cells. To visualize cellular expression, the WSN influenza virus HA or the CAL HA was fused to a fluorescent protein, mAG (green color), and the WSN NA or the CAL NA was fused to mCherry. To allow cellular expression of these fusion proteins, multiple plasmids were constructed (pcDNA3 IVS CMV-CALHA-mAG, pcDNA3 IVS CMV-CALNA-mCherry, pcDNA3 IVS CMV-WSNHA-mAG, and pcDNA3 IVS CMS-WSNNA-mCherry). Expression of viral proteins in cells was quantified by flow cytometry analysis of fluorescence from transfected cells at 36 hours post-addition of plasmids with the fusion genes. Preparation of SERS substrates. Cover glasses were washed with piranha solution (3:1 of H2SO4 and H2O2), rinsed with deionized water (DIW), and dried with N2 gas. For SERS substrates, 80 nm spherical gold nanoparticle solution (GNPs, BBI Solutions) was mixed with 10 μM CuSO4 solution at the 50:1 v/v ratio and then poured onto the cover glass. After complete drying, we obtained the SERS spectra from a dried cell pellet on our SERS substrate. The pellet 4 ACS Paragon Plus Environment

Page 4 of 23

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


was prepared by centrifugation of transfected cells in conical tubes. To prevent proteins in the culture media from contributing to the SERS signals, the culture media top on the cell pellet in the tubes were completely removed by aspiration. 50 μl of PBS was added to the pellet and the cell pellet was re-dispersed in the PBS. The resulting cell-dispersed solution was dropped onto the SERS substrate and kept in room temperature until the solution was thoroughly vaporized. The final sample was used for measurement of the SERS spectra. SERS measurement. After excitation by 10 mW, 785 nm CW laser, SERS signals were collected with an inverted microscope (Axiovert Carl Zeiss) and sent to a spectrometer (SP2300, PI Acton). Our system was connected to a CCD detector that was cooled to -70 °C. The laser light was coupled through a 50x objective lens (NA = 0.70), which was also used to collect the Raman signals (exposure time = 10 s). All data processing was performed using MATLAB (MathWorks, Inc., Natick, MA). The elimination of baseline and denoising of each spectrum were utilized by BEADS (baseline estimation and denoising w/ sparsity) toolbox in MATLAB. Principal component analysis (PCA). A total of 1,200 Raman shift wavenumbers from 474 to 1988 cm-1 were chosen as the variables for PCA. PCA was conducted using MATLAB codes. The built-in ‘pca’ function was used to get principal component coefficients, principal component scores, and principal component variances. Two principal components (PCs), PC1 and PC2, were chosen to best capture the characteristics of Raman spectra from cells. Loadings were performed with the eigenvectors of covariance matrix for multiple sets of the sample Raman spectra data in comparison. Raman spectra were projected onto score plot proportionally to loadings. Error ellipses of 95% confidence were plotted using the ‘error_ellipse’ function.25

Results and Discussion Construction of models for human cells infected with influenza viruses. To secure safety for experimenters we constructed human cells that transiently express the envelope proteins of influenza viruses as models for cells infected with the potentially pathogenic viruses. These 5 ACS Paragon Plus Environment


model cells were prepared by introducing plasmids that encode the HA and NA of influenza virus A/WSN/33 H1N1, called as the WSN strain here, or influenza virus A/California/04/2009 H1N1, called as the CAL strain, into human embryonic kidney (HEK) 293T cells via calcium phosphate-mediated transfection. The WSN strain is a widely used laboratory-adapted influenza virus strain and the CAL strain is a recently emerging one that caused the 2009 flu pandemics.4,5,6,26 Another model for cells infected with new influenza viruses that can be generated by recombination or reassortment of segmented genomes of two different influenza viruses was also constructed by adding plasmids that encode the HA of the WSN strain and the NA of the CAL strain into cells. In total, three models were made for cells infected with the WSN (CWSN HA+NA), CAL strains (CCAL HA+NA) and a new virus containing mixed genomes (CWSN HA+CAL NA).27 Verification of viral envelope protein expression in cells. DNA transfection-mediated expression of influenza virus envelope proteins in HEK 293T cells was tested by imaging. For this test we newly constructed recombinant genes that encode a fusion protein of the WSN HA and Azami-Green (mAG) fluorescent protein or a fusion protein of the WSN NA and mCherry fluorescent protein (Figure 2a). The recombinant genes were under control of CMV promoter for expression. When plasmids that encode the WSN HA-mAG fusion protein or WSN NAmCherry fusion protein were introduced, cells clearly produced green and cherry colors, respectively, as shown in Figure 2b indicating the cellular expression of the corresponding viral envelope proteins. When both types of plasmids were introduced at the same time, cells expressed both envelope proteins. As the merged one of fluorescent green and cherry images for the case of addition of both HA and NA genes shows yellowish color spots (see the WSN HA NA panel in the Figure 2b), it is apparent that HA and NA proteins were often co-localized in the transfected cells. As more plasmids were added into cells, stronger fluorescence was observed, indicating higher levels of expression of the corresponding viral envelop proteins (not shown). The results of this imaging experiment confirmed that model cells expressing viral 6 ACS Paragon Plus Environment

Page 6 of 23

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


envelope proteins could be well developed via introduction of genes encoding viral envelope proteins into cells. SERS signal from cells expressing influenza virus envelope proteins. SERS substrates are normally prepared by varying the size and shape of nanostructures. These variations of nanostructures can be performed by various methods involving the use of focused-ion beam and photolithography.28, 29 Our SERS substrate was simply and cost-effectively fabricated by salt-induced aggregation of 80 nm gold nanoparticles (GNPs). 30, 31 We obtained Raman signals from cells, on the nanoparticle substrate, expressing influenza virus envelope proteins. Cells were genetically manipulated to express only HA or NA proteins or both for the WSN strain or CAL strain. In addition, some cells were adjusted to express HA and NA proteins of the WSN and CAL strain, respectively. SERS spectra were obtained from multiple different spots of cells and shown in Figure 3 after averaging. Depending on the viral proteins presented on cells, characteristic Raman spectra were generated. However, since viral and cellular proteins are heterogeneously distributed on the surface of cells and we measured spectra from randomly sampled spots of interface between cells and the substrate, the obtained Raman spectra modestly varied over measurements given a viral protein type or combination (Figure S1). This led to a difficulty in clearly identifying the type of the virus that infected cells from the obtained SERS signals. PCA of Raman spectra from cells. To overcome the difficulty from varying Raman spectra for the same sample, we considered systematic computational ways. As a simple statistical method, PCA was applied to capture the key characteristics from the noisy Raman spectra. PCA is a method for data compression or data dimensionality reduction to highlight the core information only. It involves finding of linearly independent principal components (PCs) and projection of the whole data to linear spaces of lower dimension that the PCs generate. PCs are selected eigenvectors of covariance matrix of Raman spectra for samples in comparison. The eigenvector corresponding to the largest eigenvalue is chosen as the first principal component 7 ACS Paragon Plus Environment


(PC1) and this vector shows the direction along which we can find the maximum difference in spectra between samples. The following eigenvector corresponding to the second largest eigenvalue is chosen as the second principal component (PC2), which is orthogonal to PC1. Two to three PCs are often considered during PCA (resulting in data dimensionality reduction from N-dimension (N-D) to 2-D and 3-D, respectively) to ignore vague variations but to more easily find the key variations in spectra between samples. 32, 33 We applied PCA to the Raman spectra from cells uninfected as a control and infected with influenza viruses. Raman spectra were obtained multiple times for the two cell cases. To eliminate intensity fluctuations of Raman signals which would arise from the different level of surface-enhancement over measuring SERS hot-spots, the signal intensities were adjusted by min-max normalization. The whole sets of the adjusted spectra were then projected onto 2-D PCA score space. This PCA process finally yielded multiple points with different coordinate values on PC1 and PC2 (Figure 4a and 4c). The multiple projected points for each cell case could be enclosed by a distinct 95% confidence ellipse. In comparison between cells infected with the WSN strain and uninfected cells, the two 95% confidence ellipses corresponding to each case did not overlap (Figure 4a), indicating that the Raman spectra for cells infected with the WSN strain were statistically significantly different from those for the uninfected cell case (P-value < 0.05). In similar, projected Raman spectra for the cells infected with the CAL strain were significantly different from those for the control case (Figure 4c). Key Raman shifts for comparison between infected and uninfected cells. After we distinguished Raman signals of influenza virus-infected cells from those of uninfected cells via projection of the whole spectra onto PCA score space, we further tried to capture the key features of spectra specific for each cell case. PCs were actually constructed from the multiple sets of spectra for samples in comparison, by applying different weights on each Raman shift wavelength, to find the direction of maximum variance between samples. In other word, the Raman shift wavelengths that contribute to PCs in positively and negatively high levels, those 8 ACS Paragon Plus Environment

Page 8 of 23

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


with high and low loading values in PC loading plots (shown as shift wavelengths at multiple peaks in Figure 4b and 4d), are the keys to distinguish infected cells from uninfected cells. The PCA result showing that the SERS for CWSN HA+NA were uniquely projected to the third (PC1 < 0 and PC2 < 0) and fourth quadrants (PC1 > 0 and PC2 < 0) of the score plot unlike those for uninfected cells (projected mostly to the first quadrant, Figure 4a) indicates that the Raman shifts contributing much to negative values for PC2 would be the keys to distinguish the virally infected cells (CWSN HA+NA) from uninfected ones. As highlighted with purple color in Figure 4b, the Raman shifts at 737, 1331, 1467, and 1625 cm-1 significantly contributed to negative values for PC2. The score plotting result denotes that these Raman shifts represent the unique characteristics of SERS for the cells corresponding to CWSN HA+NA. Similarly, the SERS for CCAL HA+NA were uniquely projected to the first (PC1 > 0 and PC2 > 0) and second quadrants (PC1 < 0 and PC2 > 0) of the score plot unlike those for uninfected cells (projected mostly to the fourth quadrant, Figure 4b). Considering the key Raman shifts that contributed much to positive values for PC2, it could be concluded that the Raman shifts at 609, 720, and 1625 cm1

are the most important to distinguish cells infected with the CAL strain from uninfected cells.

In addition, the Raman shifts at 985 cm-1 and 1008 cm-1 were thought to be representative indicators for uninfected cells based on that the SERS for uninfected cells were mostly projected to the first (Figure 4a) and fourth quadrants of the scoring plots (Figure 4c) and that the Raman shifts at the two wavelengths significantly contributed to positive values for PC1 and PC2 (when compared with CWSN HA+NA) and to positive values for PC1 and negative values for PC2 (when compared with CCAL HA+NA). Comparing Raman signals produced from different envelope proteins. The plausible target molecules for our SERS analysis were the surface proteins expressed by cells because they can interact with the GNP substrates. Therefore, we thought that the key Raman shift features were associated with the surface protein types. As a proof of concept, we tried to compare Raman signals from each type of viral envelope protein (HA or NA). PCA result from cells expressing 9 ACS Paragon Plus Environment


either CAL strain HA or NA is shown in Figure S2. Since the confidence ellipses for the HA, NA, and the control cases were not coincided (Figure S2a), we can say that our SERS analysis can successfully distinguish the existence of HA or NA and the lack of either protein for the CAL strain. The Raman spectra of for the control case were projected on the first quadrant of the score plot indicating that their PC1 and PC2 values were both positive. On the other hand, the spectra for CCAL HA and CCAL NA were projected to the domain of negative PC1 and positive PC2 values and domain of negative PC2 values, respectively (Figure S2a). Subsequent loading plot analysis (Figure S2a and S2b), as described above with the data in Figure 4, indicated that the key characteristic spectra peaks for CCAL HA and CCAL NA are the one at 1206 cm-1 and the ones at 740 and 859 cm-1, respectively (Figure S2b and S2c). This result highlights that PCA analysis is helpful not only to distinguish which type of influenza virus variant infects cells, but also to identify which viral envelope protein is specifically expressed on the infected cells. By the same way, we found the characteristic Raman peaks for CWSN HA at 740 cm-1 and for CWSN NA

at 1107 cm-1 (Figure S3). These Raman shift peaks were also previously detected from the

SERS analysis of the virus particles of the WSN strain21, consistent scientific observations for the same type of influenza virus. We have also attempted to compare the Raman spectra between cells expressing a single type of viral envelope protein (HA or NA) and cells expressing both types (HA + NA) (Figure S4 and S5). In both WSN and CAL strain infection cases, the Raman spectra projected onto PCA score plot for the cells expressing both types of envelope proteins (HA + NA) enclosed the projected spectra for the cells expressing either type (HA or NA) (Figure S4b, WSN strain case) or were between the projected spectra (Figure S5b, CAL strain case). This result indicates that the Raman spectra from cells expressing both HA and NA had some characters of the spectra from cells expressing HA or NA, as rationally expected. PCA of Raman spectra from the cells infected with mutant viruses. For our attempted goal, we finally applied our method to identify the mutant influenza virus that infected human cells. 10 ACS Paragon Plus Environment

Page 10 of 23

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


First, the Raman spectra projected to PCA score plot for the mutant-infected cells were clearly distinguished from those for uninfected cells (Figure 5a). More importantly, the projected Raman spectra for the mutant case could be distinguished from those of the WSN (Figure 5b) and CAL strain (Figure 5c) infection cases, more clearly for the latter case. However, the 95% confidence ellipse for the mutant case was partially overlapped with those for the parental virus infection cases (infection with the WSN and CAL strains, Figure 5b and 5c). It should be expected because the surface components of the parental viruses (WSN HA and NA, and CAL HA and NA) would contribute not only to the generation of Raman spectra from the cells infected with the parental viruses but also to the generation of the spectra from the mutant that carried the same surface components (WSN HA and CAL NA). Based on the greater overlapping of the projected spectra for the mutant (CWSN HA + CAL NA) and WSN strain case (CWSN HA + WSN NA, Figure 5b) than for the mutant and CAL strain case (CCAL HA + WSN NA, Figure 5c), we could suppose that HA contributes more than NA to the obtained Raman spectra. Following the steps that were applied to comparison of Raman spectra between different cases (Figure 4), we finally found that the Raman shifts at 696, 737 cm-1 were the characteristic ones for the cells infected with the mutant virus (colored with blue in Figure 5d and Table S1). Uninfected cells also had some characteristic Raman shifts at 985, 1008, and 1578 cm-1. Sensitivity of SERS-based detection of viral protein expression. SERS analysis experiments were performed using cells transfected with three different amounts (100- fold variations, from 0.1 to 10 μg) of DNA that encodes the CAL strain HA to determine the sensitivity of the SERSbased detection system (Figure S6). Introduction of 0.1, 1.0, and 10.0 μg DNA into cells led to different levels of cellular expression of the corresponding HA-GFP fusion proteins; With 0.1 μg DNA addition 0.03% of cells showed GFP expression at the averaged intensity of 214.7. With 1.0 μg and 10.0 μg DNA additions, 0.19% and 15.7% of cells showed GFP expression at the averaged intensities of 210 and 1034, respectively. With this variation of amount of introduced DNA, we could adjust the level of expression of the viral envelope protein (CAL 11 ACS Paragon Plus Environment


HA) in cells. Interestingly, even with a great reduction in the percentage of cells expressing the viral proteins (from 15.7% to 0.03%, greater than 500-fold variation), the SERS system generated equivalent levels of SERS peaks at a representative Raman shift wavelength for the CAL HA, 1206 cm-1, highlighting the high sensitivity and consistent performance of the developed detection system.

Conclusions Early detection and identification of newly emerging viruses is currently highly needed in the era of rapid viral evolution. In this study, we newly targeted the virally infected cells for early identification of influenza viruses. We applied a label-free SERS-based detection of cells to identify the type of influenza viruses. We observed SERS signals from cells infected with influenza virus strains and a mutant virus that can be generated via genome re-assortment in the future (WSN + CAL strain). To facilitate the early detection of influenza viruses by systematically capturing the key characteristic Raman shift information from the jumbled Raman spectra data, we employed a statistical method based on PCA. The combination of SERS and PCA enabled clear identification of cells infected with the different types of influenza viruses. Moreover, our method detected the mutant virus-infected cell with reasonable reliability. Further studies utilizing this approach may be required a novel methodology including effective SERS substrate and signal analysis, this will provide more precise and sensitive diagnosis method of newly emerging influenza viruses. Our ideas and approaches can be easily extended to detect and identify other enveloped viruses.

Acknowledgements This research was supported by a grant of the Korea Health Technology R&D Project through the Korea Health Industry Development Institute (KHIDI), funded by the Ministry of Health & Welfare, Republic of Korea (grant number: HR14C0007-060018). This research was also 12 ACS Paragon Plus Environment

Page 12 of 23

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


supported by a grant from the National Research Foundation of Korea (NRF2016M3A9B6947831). This work was supported by Korea Institute of Planning and Evaluation for Technology in Food, Agriculture and Forestry (IPET) through Animal Disease Management Technology Development Program, funded by Ministry of Agriculture, Food and Rural Affairs(MAFRA) (Grant number: 118094-03). There is no conflict of interest to declare. Jaeyoung Lim, Jung-soo Nam, and Hyunku Shin contributed equally to this work.

Supporting information Supplementary data associated with this article can be found in the online version: Representative data of the SERS spectra, Raman signal distinction between viral envelop proteins of the CAL strain and the WSN strain, comparison of Raman spectra between cells expressing viral envelope proteins, Raman signal intensity variation depending on the expression level of proteins, and characteristic Raman shifts for cells infected with influenza viruses.



References (1) Park, M., Wu, P., Goldstein, E., Kim, W. J., & Cowling, B. J. Influenza-associated Excess Mortality in South Korea. Am. J. Prev. Med., 2016, 50(4), e111-e119. (2) Koh, M. T., Eg, K. P., & Loh, S. S. Hospitalised Malaysian Children with Pandemic (H1N1) 2009 Influenza: Clinical Characteristics, Risk Factors for Severe Disease and Comparison with the 2002–2007 Seasonal Influenza. Singapore Med. J., 2016, 57(2), 81. (3) Thompson, W. W., Shay, D. K., Weintraub, E., Brammer, L., Cox, N., Anderson, L. J., & Fukuda, K. Mortality Associated with Influenza and Respiratory Syncytial Virus in the United States. Jama, 2003, 289(2), 179-186. (4) Tumpey, T. M., & Belser, J. A. Resurrected Pandemic Influenza Viruses. Annu. Rev. Microbiol., 2009, 63, 79-98. (5) Neumann, G., Noda, T., & Kawaoka, Y. Emergence and Pandemic Potential of Swine-origin H1N1 Influenza Virus. Nature, 2009, 459(7249), 931. (6) Cohen, J. Past Pandemics Provide Mixed Clues to H1N1's Next mMves. Science, 2009, 324(5930), 996-997 (7) De Boer, G. F., Back, W., & Osterhaus, A. D. M. E. An ELISA for Detection of Antibodies Against Influenza A Nucleoprotein In Humans and Various Animal Species. Arch. Virol., 1990, 115(1-2), 47-61. (8) Yang, J. R., Lo, J., Liu, J. L., Lin, C. H., Ho, Y. L., Chen, C. J., Wu, H. S., & Liu, M. T. Rapid SYBR Green I and Modified Probe Real-Time Reverse Transcription-PCR Assays Identify Influenza H1N1 Viruses and Distinguish between Pandemic and Seasonal Strains. J. Clin. Microbiol., 2009, 47(11), 3714-3716. (9) Yang, J. R., Lo, J., Liu, J. L., Lin, C. H., Ho, Y. L., Chen, C. J., Wu, H. S., & Liu, M. T. Rapid SYBR Green I and Modified Probe Real-Time Reverse Transcription-PCR Assays Identify Influenza H1N1 Viruses and Distinguish between Pandemic and Seasonal Strains. J. Clin. Microbiol., 2009, 47(11), 3714-3716. 14 ACS Paragon Plus Environment

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


(10) Spackman, E., Senne, D. A., Myers, T. J., Bulaga, L. L., Garber, L. P., Perdue, M. L., Lohman, K., & Suarez, D. L. (2002). Development of a Real-Time Reverse Transcriptase PCR Assay for Type A Influenza Virus and the Avian H5 and H7 Hemagglutinin Subtypes. J. Clin. Microbiol., 2002, 40(9), 3256-3260. (11) Templeton, K. E., Scheltinga, S. A., Beersma, M. F., Kroes, A. C., & Claas, E. C. Rapid and sensitive method using multiplex real-time PCR for diagnosis of infections by influenza A and influenza B viruses, respiratory syncytial virus, and parainfluenza viruses 1, 2, 3, and 4. J. Clin. Microbiol., 2004, 42(4), 1564-1569. (12) Kuku, G., Altunbek, M., & Culha, M. Surface-Enhanced Raman Scattering for Label-Free Living Single Cell Analysis. Anal. Chem., 2017, 89(21), 11160-11166. (13) Kuku, G., Saricam, M., Akhatova, F., Danilushkina, A., Fakhrullin, R., & Culha, M. Surface-Eenhanced Raman Scattering to Evaluate Nanomaterial Cytotoxicity on Living Cells. Anal. Chem., 2016, 88(19), 9813-9820. (14) Navas-Moreno, M., Mehrpouyan, M., Chernenko, T., Candas, D., Fan, M., Li, J. J., Yan, M., & Chan, J. W. Nanoparticles for Live Cell Microscopy: A Surface-Enhanced Raman Scattering Perspective. Sci. Rep., 2017, 7(1), 4471. (15) Covalciuc, K. A., Webb, K. H., & Carlson, C. A. Comparison of Four Clinical Specimen Types for Detection of Influenza A and B Viruses by Optical Immunoassay (FLU OIA test) and Cell Culture Methods. J. Clin. Microbiol., 1999, 37(12), 3971-3974. (16) Rowe, T., Abernathy, R. A., Hu-Primmer, J., Thompson, W. W., Lu, X., Lim, W., Fukuda, K., & Katz, J. M. Detection of Antibody to Avian Influenza A (H5N1) Virus in Human Serum by Using a Combination of Serologic Assays. J. Clin. Microbiol., 1999, 37(4), 937-943. (17) Von Itzstein, M. The War Against Influenza: Discovery and Development of Sialidase Inhibitors. Nat. Rev. Drug Discovery, 2007, 6(12), 967.



Page 16 of 23

(18) Dasary, S. S., Singh, A. K., Senapati, D., Yu, H., & Ray, P. C. Gold Nanoparticle Based Label-Free SERS Probe for Ultrasensitive and Selective Detection of Trinitrotoluene. J. Am. Chem. Soc., 2009, 131(38), 13806-13812. (19) Sha, M. Y., Xu, H., Penn, S. G., & Cromer, R. SERS Nanoparticles: a New Optical Detection Modality for Cancer Diagnosis. Nanomedicine (Lond), 2007, 2, 725-734. (20) Abell, J. L., Driskell, J. D., Dluhy, R. A., Tripp, R. A., & Zhao, Y. P. Fabrication and Characterization of a Multiwell Array SERS Chip With Biological Applications. Biosens. Bioelectron., 2009, 24(12), 3663-3670. (21) Lim, J. Y., Nam, J. S., Yang, S. E., Shin, H., Jang, Y. H., Bae, G. U., Kang, T.; Lim, K. I., & Choi, Y. Identification of Newly Emerging Influenza Viruses by Surface-Enhanced Raman Spectroscopy. Anal. Chem., 2015, 87(23), 11652-11659. (22) Wold, S., Esbensen, K., & Geladi, P. Principal Component Analysis. Chemom. Intell. Lab. Syst., 1987, 2 37-52 (23) Castells, F., Laguna, P., Sörnmo, L., Bollmann, A., & Roig, J. M. Principal Component Analysis in ECG Signal Processing. EURASIP J. Adv. Signal Process., 2007, (1), 074580. (24) Bartlett, M. S., Movellan, J. R., & Sejnowski, T. J. Face Recognition by Independent Component Analysis. IEEE Trans. Neural Netw., 2002, 13(6), 1450-1464. (25)

Johnson,

A.

J.,

(https://www.mathworks.com/matlabcentral/fileexchange/4705-error_ellipse),

error_ellipse MATLAB

Central File Exchange. Retrieved Jul 23, 2015. (26) Zhang, J., Zhang, Z., Fan, X., Liu, Y., Wang, J., Zheng, Z., Chen, R., Wang, P., Song, W., Chen, H., & Guan, Y. Pandemic H1N1 Influenza Virus Replicates in Human Lung Tissues. J. Infect. Dis., 2010, 201(10), 1522-1526. (27) Chen, C. A. Calcium Ohosphate-Mediated Gene Transfer: a Highly Efficient Transfection System for Stably Transforming Cells with Plasmid DNA. Biotechniques, 1988, 6, 632-638.


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


(28) Brown, L. V., Yang, X., Zhao, K., Zheng, B. Y., Nordlander, P., & Halas, N. J. Fan-Shaped Gold Nanoantennas Above Reflective Substrates for Surface-Enhanced Infrared Absorption (SEIRA). Nano Lett., 2015, 15(2), 1272-1280. (29) Yu, Q., & Golden, G. Probing the Protein Orientation on Charged Self-Assembled Monolayers on Gold Nanohole Arrays by SERS. Langmuir, 2007, 23(17), 8659-8662. (30) Bell, S. E., & Sirimuthu, N. M. Surface-Enhanced Raman Spectroscopy (SERS) for SubMicromolar Detection of DNA/RNA Mononucleotides. J. Am. Chem. Soc., 2006, 128(49), 15580-15581. (31) Choi, D., Kang, T., Cho, H., Choi, Y., & Lee, L. P. Additional Amplifications of SERS via an Optofluidic CD-Based Platform. Lab. Chip, 2009, 9(2), 239-243. (32) Guicheteau, J., Argue, L., Emge, D., Hyre, A., Jacobson, M., & Christesen, S. Bacillus Spore Classification via Surface-Enhanced Raman Spectroscopy and Principal Component Analysis. Appl. Spectrosc., 2008, 62(3), 267-272. (33) Ryder, A. G. Classification of Narcotics in Solid Mixtures Using Principal Component Analysis and Raman Spectroscopy. J. Forensic Sci., 2002, 47(2), 275-284.



Figure 1. A schematic diagram of identification of cells infected with newly emerging influenza viruses. Surface enhanced Raman spectra were obtained from the interface between gold nanoparticles and viral envelope proteins (HA and NA) on virally infected cells. The key features of SERS patterns to identify the type of viruses that infected cells were extracted via PCA of Raman signals.


Page 18 of 23

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 2. Expression of influenza virus envelope proteins in cells. (a) DNA constructs to express the viral envelope proteins of the influenza WSN strain. (b) Expression of the envelope proteins of the influenza WSN strain in HEK 293T cells. Scale bar: 100 µm.



Figure 3. Raman signals from cells. The averaged spectra of (a) cells infected with the WSN strain, (b) cells infected with the CAL strain, and (c) cells expression the WSN HA or the CAL NA or both (for a newly emerging virus case). Control indicates the signal from cells not expressing viral proteins. Number of measuring spots: 9 (WSN HA), 9 (WSN NA), 12 (WSN HA+NA), 11 (CAL HA), 10 (CAL NA), 18 (CAL HA+NA), 20 (WSN HA + CAL NA), and 18 (Control).


Page 20 of 23

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


Figure 4. PCA plots of Raman spectra that were obtained from cells. (a) Score plot; Projection of Raman spectra onto 2-D PCA score space for the cases of cells infected with the WSN strain (CWSN HA NA) and uninfected cells as a control. Purple stars and green squares represent the projected Raman spectra for the infected and uninfected cells, respectively. 95% confidence ellipses enclosing the projected spectra as dots on the PCA score space are also shown. (b) Loading plot; Loading of the PC1 and PC2 for the spectra from cells infected with the WSN strain and uninfected cells. The black line represents the loading of PC1 and the red line represents the loading of PC2. The green and purple color peaks indicate the characteristic Raman shifts for the uninfected cells and cells infected with the WSN strain (CWSN HA+NA), respectively. c) Score plot; Projection of Raman spectra onto 2-D PCA score space for the cases of cells infected with the CAL strain (CCAL HA+NA) and uninfected cells as a control. Pink triangles and green squares represent the projected Raman spectra for the infected and uninfected cells, respectively. 95% confidence ellipses enclosing the projected spectra as dots on the PCA score space are also shown. (d) Loading plot; Loading of the PC1 and PC2 for the spectra from cells infected with the CAL strain and uninfected cells. The black line represents the loading of PC1 and the red line represents the loading of PC2. The green and pink color peaks indicate the characteristic Raman shifts for the uninfected cells and cells infected with the CAL strain (CCAL HA+NA).



Figure 5. Comparison of Raman spectra between cells infected with genome-assorted mutant influenza virus (virus with the WSN HA and CAL NA genomes) and cells infected with its parental virus (WSN or CAL strain). (a) PCA score plot for the spectra from the mutant infected (CWSN HA+CAL NA) and uninfected cells. Blue and green squares represent the projected Raman spectra for CWSN HA+CAL NA and uninfected cells as a control, respectively. Ellipses enclosing the projected dots for each case are 95% confidence ellipses. (b) PCA score plot for the spectra from CWSN HA+NA (purple triangles) and CWSN HA+CAL NA (blue squares). (c) PCA score plot for the spectra from CCAL HA+NA (pink squares) and CWSN HA+CAL NA (blue squares). (d) Characteristic Raman peaks found from PCA for virally infected and uninfected cells.


Page 22 of 23

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


For TOC only


Identification of Newly Emerging Influenza Viruses by Detecting the

Recommend Documents