Spectroscopic Method for Fast and Accurate Group A Streptococcus

Article pubs.acs.org/ac

Spectroscopic Method for Fast and Accurate Group A Streptococcus Bacteria Detection Dillon Schiff,† Hagit Aviv,† Efraim Rosenbaum,‡ and Yaakov R. Tischler*,† †

Department of Chemistry, Institute for Nanotechnology and Advanced Materials, Bar-Ilan University, Ramat-Gan 5290002, Israel Leumit Health Care Services, Ramat Beit Shemesh, Beit Bhemesh 99851, Israel

‡

S Supporting Information *

ABSTRACT: Rapid and accurate detection of pathogens is paramount to human health. Spectroscopic techniques have been shown to be viable methods for detecting various pathogens. Enhanced methods of Raman spectroscopy can discriminate unique bacterial signatures; however, many of these require precise conditions and do not have in vivo replicability. Common biological detection methods such as rapid antigen detection tests have high specificity but do not have high sensitivity. Here we developed a new method of bacteria detection that is both highly specific and highly sensitive by combining the specificity of antibody staining and the sensitivity of spectroscopic characterization. Bacteria samples, treated with a fluorescent antibody complex specific to Streptococcus pyogenes, were volumetrically normalized according to their Raman bacterial signal intensity and characterized for fluorescence, eliciting a positive result for samples containing Streptococcus pyogenes and a negative result for those without. The normalized fluorescence intensity of the Streptococcus pyogenes gave a signal that is up to 16.4 times higher than that of other bacteria samples for bacteria stained in solution and up to 12.7 times higher in solid state. This method can be very easily replicated for other bacteria species using suitable antibody−dye complexes. In addition, this method shows viability for in vivo detection as it requires minute amounts of bacteria, low laser excitation power, and short integration times in order to achieve high signal.

T

Raman excitation in the ultraviolet range can also increase the Raman signal, but the ultraviolet radiation often degrades samples too quickly without constant rotation in addition to requiring long acquisition times.7,8 Photoluminescence shows the unique wavelength peak in the luminescence spectrum of a dyed bacteria species. The photoluminescence spectrum is acquired by optically exciting a sample, often via laser.9,10 Photoluminescence has been used previously in the detection of pathogens, in methods such as fluorescent nanoparticle detection of colon cancer11 and in more bacterially specific methods by detecting antibodies linked to luminescent nanocrystals.12 The most common method of photoluminescence of pathogen identification is found in the visual antigen detection of RADTs, where antibodies are linked to fluorescent microparticles. When the microparticle concentration is high enough, an agglomeration occurs which allows for a visual confirmation of antigen presence within several minutes. RADTs have a high specificity, of up to 95%, but a lower sensitivity of only 85%.13 A secondary detection method is therefore standard for a negative test,14 most commonly in the form of a BAP culture which has a 95% sensitivity. However, this takes an additional 1−2 days for confirmation.15−17 False-negative lateral flow tests are associated with the difficulty in obtaining specimens either because of the

here is a continuing need for rapid identification of bacterial and viral pathogens, whether for research applications or for the health of the general population. The use of spectroscopy to identify such pathogens is already well established.1 Spectroscopic methods have the advantage of accuracy, as spectral signatures can be unique and necessitate only minute amounts of bacteria as little as a few μm3. In addition, spectroscopic measurements can be taken in a matter of seconds, compared to minutes for common biological tests such as a rapid antigen detection test (RADT), or at least a day for a blood agar plate (BAP) culture if the species is able to be cultured. The two types of spectroscopy this work employs are Raman spectroscopy and photoluminescence. Raman spectroscopy measures the wavelengths of the photons that are inelastically scattered upon laser excitation of a sample. Recent works have shown that Raman spectroscopy can be used to identify bacterial strains.2−4 However, these techniques require enhanced conditions to produce statistically significant signatures. One such method requires the isolation of each species through a filtering process. Without isolation, the Raman signature becomes noisy and undetectable. Another Raman method uses surface-enhanced Raman scattering (SERS) to enhance the Raman signatures via localized surface plasmon resonances, thus increasing clarity in an otherwise noisy spectrum.5,6 SERS allows for easier discrimination but requires precise sample conditions and many parameter refinements for achieving the optimal enhancement. Resonance © XXXX American Chemical Society

Received: October 6, 2015 Accepted: January 10, 2016

A

DOI: 10.1021/acs.analchem.5b03754 Anal. Chem. XXXX, XXX, XXX−XXX

Article

Analytical Chemistry

Marking Group A Streptococcus by Linking a Dyed Antibody in Solution. Each experiment was conducted on four Gram-positive bacteria: Streptococcus pyogenes, Staphyloccous aureus, Streptococcus pneumoniae, and Staphylococcus epidermidis, which were chosen for their prevalence in oral infections. Bacteria species grown on BAPs were carefully removed using sterile cotton swabs and then transferred to 1.5 mL polymerase chain reaction (PCR) tubes. Then, 1.2 mL of PBS solution were added to each of the PCR tubes and mixed vigorously for several minutes using a vortex mixer until all the bacteria had dispersed. The PCR tubes were then centrifuged for 15 min at 8000 rpm followed by a careful removal of the supernatant, leaving only the “washed” bacteria. This washing process was repeated 3 times. Then, 1.2 mL of the primary antibody, Mouse anti-Streptolysin O/Mouse anti-Streptococcus Group A in PBS solution (0.25 μg/mL), was added. The PCR tubes were left to rotate at 30 rpm for 2 h followed by centrifugation and removal of the supernatant. The precipitant was then washed 5 times according to the above procedure, and 1.2 mL of the secondary dyed antibody, Goat anti-Mouse IgGDyLight, was added in PBS solution (2.5 μg/mL). To ensure maximum binding of the dye, the PCR tubes were then covered with foil to prevent photobleaching and left to rotate for an additional 2 h followed by centrifugation and removal of the supernatant. The precipitant was then washed 5 times according to the above procedure. After the final removal of supernatant, the solid bacteria precipitate was transferred onto clean glass microscope slides for spectral analysis. This process was repeated for all bacteria samples. As the bacteria were washed via dispersion in liquid solutions and then precipitated with centrifugation, we refer to this method as “in solution”. Centrifugation of the different solutions was performed by Thermo Scientific Heraeus Biofuge Stratos centrifuge. Figure 1 presents a scheme of the two steps marking procedure of the Group A Streptococcus in solution and in solid state.

discomfort involved to the patient, the inexperience of the care provider in their technique of obtaining the specimen, or testing the patient too early in the course of the disease. Over the past few years, a number of new products have been developed in order to increase sensitivity over traditional lateral flow immunoassays. These new technologies may add optical detection which pick up “weak” positive readings of the streptococcal assay which otherwise would have been undetected visually. Most traditional lateral flow assays have a limit of detection of 105 to 107 cfu/mL (colony-forming units per mL), but the new technologies have a threshold of 103 to 105 cfu/nL (91% sensitivity and 96% specificity). Other new technologies untilize PCR (polymerase chain reaction) technology, which has a “limit of detection” of just 1−2 streptococcal bacteria. As such, the sensitivity of 95.9% and specficity 94.6% of the PCR test approach that of BAP culture itself. Again, the remaining major challenge with lateral flow and PCR testing is sample acquisition.18 Combining Raman and photoluminescence spectroscopies is a relatively new practice that can provide more discriminative data for analytical chemistry and biological applications. It has seen more widespread usage as in a recent publication which combined the two methods to identify uranium19 and another which combined the use of photoluminescence measurements in order to interpret Raman spectral data of crystalline rubrene.20 Biologically, a research team was recently able to combine fluorescence and Raman in an endoscopic system as a method of tumor characterization.21 This method presents realtime, in vivo, and multiple target successful detection of a specific cancer, based on the fast imaging capability of fluorescence signals and the multiplex capability of simultaneously detected SERS signals using an optical fiber bundle for intraoperative endoscopic system. In our investigation, we combine the two spectroscopic methods by using bacterial Raman signatures to normalize photoluminescence measurements of biologically stained bacteria. Similar normalization methods were previously demonstrated to calibrate fluorescence data of organic material via the Raman shift intensity of the surrounding water.22,23 We show that this unique and efficient combination enables high relative specificity via biological staining and high relative sensitivity via Raman spectroscopy for the detection of Streptococcus pyogenes or Group A streptococcus (GAS). The importance of rapidly detecting GAS is evident in the approximate 7.3 million annual U.S. outpatient physician visits by children due to sore throats of which approximately 25% can be attributed to Group A streptococcus.24 The detection methods employed by this work seek to combine the standard specific methods of biological detection with the highly sensitive spectroscopic methods. This novel approach has many advantages for future pathogen detection, particularly in vivo.

Figure 1. Scheme describing the two steps procedure for marking the Streptococcus pyogenes bacteria species. A represents the Mouse antiStreptococus Group A (primary antibody), and B represents the Goat anti-Mouse IgG- DyLight 550 that binds any mouse antibody (secondary fluorescent antibody).

Marking Group A Streptococcus via Linking a Dyed Antibody in Solid State. This procedure was adapted from previous solid-phase immunoassay methods.25 Bacteria species grown on BAPs were carefully removed using sterile cotton swabs and then transferred to individual Polyvinylidene Fluoride (PVDF) membrane filters with 0.2 μm pore size. A single filter was placed on a Büchner funnel-flask apparatus. The bacteria species was washed with 5 mL of a PBS solution containing BSA (1% w/v). Then, 2 mL of the primary antibody solution, Mouse anti-Streptococcus Group A antibody in PBS (0.25 μg/mL) containing BSA (1% w/v) were added to the funnel for 10 min at 37 °C, and excess antibody was removed by vacuum filtration. The filter was washed again with 5 mL of a PBS solution containing BSA (1% w/v). Then, 2 mL of the

■

EXPERIMENTAL METHODS Materials. All bacteria cultures were grown at Leumit Health Services Laboratories, Israel. BAPs were purchased from Hylabs, Israel. The antibodies: Mouse anti-Streptolysin O (0.1 mg/mL), Mouse anti-Streptococcus Group A (0.1 mg/mL), and Goat anti-Mouse IgG- DyLight 550 (2 mg/mL) were purchased from Biotest, Israel. Phosphate-buffered saline (PBS) and bovine serum albumin (BSA) (98%) were purchased from Sigma-Aldrich, Israel. B

DOI: 10.1021/acs.analchem.5b03754 Anal. Chem. XXXX, XXX, XXX−XXX

Article

Analytical Chemistry secondary dyed antibody solution, Goat anti-Mouse IgGDyLight in PBS (2.5 μg/mL) containing BSA (1% w/v) was added to the funnel for 10 min at 37 °C, and again excess antibody was removed by vacuum filtration. Finally, the filter was again washed with 5 mL of a PBS solution containing BSA (1% w/v) and subsequently removed from the funnel for spectral analysis of the bacteria. This process was repeated for all bacteria samples. As the bacteria species were never dispersed in liquid solution, rather remained adhered to the filter paper, we refer to this method as in “solid state”. Fluorescence of Dye. Fluorescence spectrum of the dyed antibody was recorded using a spectrofluorometer (Cary Eclipse, Agilent Technologies Inc.). Excitation and emission slits were fixed to excite and collect a bandwidth of 5 nm. The spectrofluorometer was corrected for the wavelength dependence of the detector. Excitation wavelength was set to 532 nm. Figure 2 presents the full fluorescence spectrum of the Goat anti-Mouse IgG- DyLight in a PBS solution; the photoluminescence peak appears around 575 nm.

Figure 3. Decay in the photoluminescence peak intensity of the DyLight 550 that is used to mark the bacteria. A: corresponds to the decay when OD3 was used (60 μW) with 5 s integration. B: corresponds to the decay when OD2 was used (0.6 mW) with 2 s integration.

Raman scattering measurement was taken in the same focus location as the photoluminescence and acquired using the same laser setup. The bacteria species were excited individually by a laser with an excitation wavelength of λex = 784 nm at 100 mW with acquisition time of 30 s and a grating density of 600 g/ mm. Previously, the Raman spectra of the bacteria samples were taken and three clear Raman shifts were observed with significant signal in all bacterial samples. When agarose from the culture plate was present in the measured sample, however, a high background was observed. Therefore, spectra were taken from fresh bacteria sources when possible. Figure 4 presents the bacterial Raman spectrum taken from a fresh GAS sample. Figure 2. Fluorescence spectrum of the Goat anti-Mouse IgGDyLight 550 solution (2.5 μg/mL). The dye solution was excited at 532 nm.

Photoluminescence of Bacteria. Photoluminescence spectra were measured using a micro-PL setup (HORIBA Scientific LabRAM HR) in air at room temperature. The bacteria species were individually excited by a laser with an excitation wavelength of λex = 532 nm at 60 μW with acquisition time of 5 s and a grating density of 600 g/mm. In solid state, decay in the fluorescence of DyLight 550 occurs. Figure 3 presents two different decay plots that refer to different intensities of the exciting laser. Line A in Figure 3 presents the decay when OD3 was used (60 μW) with 5 s integration, these were the conditions in the bacteria photoluminescence measurements. Line B in Figure 3 presents the decay when OD2 was used (0.6 mW) with 2 s integration. Even though the slope of the decay can be decreased by lowering the laser intensity, the decay cannot be completely avoided. Thus, in order to get reliable photoluminescence intensities for the normalization process, only the relevant window was measured, and the bacteria fluorescence was taken for the peak range alone (560−600 nm). Normalization via Raman. In order to compare the photoluminescence spectra produced by the varying volumes of samples, normalization was required. For this procedure, a

Figure 4. General bacterial Raman spectrum taken from a fresh Group A streptococcus sample. Labeled peaks correspond to general bacteria markers.

The shift appearing at 1455 cm−1 correlates to the CH2 vibrations of the lipid chains in the bacterial membrane and to some of the proteins, and the shift at 1665 cm−1 correlates to the C−N vibrations of the amide bond in the proteins.26 The strongest Raman shift was observed at 2930 cm−1 and correlates to C−H vibrations of the lipid chains in the GramC

DOI: 10.1021/acs.analchem.5b03754 Anal. Chem. XXXX, XXX, XXX−XXX

Article

Analytical Chemistry positive bacterial membrane. This peak intensity was used to infer relative volumes of measured bacteria enabling normalization of the photoluminescence signal of the marked bacteria according to the equation: In =

Ipl IR

where Ipl is the raw photoluminescence intensity, IR is the corresponding Raman signal intensity at 2930 cm−1, and In is the resulting volumetrically normalized photoluminescence intensity. This procedure is shown in practice in Figure 5 and a full example that includes the Raman shift intensity at 2930 cm−1 appears in the Supporting Information. It is important to mention that absorption spectra of the bacteria showed no absorption in the wavelength of interest indicating a nonresonant Raman shift at 2930 cm−1. This enables a reliable correlation between the Raman shift intensity and the measured amount of material. Absorption spectra of the bacteria are presented in the Supporting Information. The normalized fluorescence intensity of specifically linked bacteria species was compared to the normalized fluorescence intensity of the nonspecifically linked bacteria species. This comparison was performed by dividing the peak intensity of the normalized fluorescence spectra and the produced ratio was used to determine positive and negative results.

■

RESULTS AND DISCUSSION In order to distinguish streptococcus samples, the initial antibody, Mouse anti-Streptolysin O, was linked to the bacteria species followed by the secondary dyed antibody. The linkage procedure in solution was performed according to the Experimental Methods section. Figure 5 shows the broad initial and normalized photoluminescence spectra of the bacteria species with the linked anti-Streptolysin O + dyed antibody complex. Initial results clearly demonstrate the need for normalization as the significances of spectra are unclear making bacteria species incomparable. The volumetrically normalized results show the relative specificity of the antistreptolysin O antibody to the bacteria species in the Streptococcus genus as well as Staphylococcus aureus given by the significant signal intensities of species A, B, and C. When the normalized fluorescence intensities of A, B, and C were divided by the normalized fluorescence intensity of D which represents only nonspecific binding, the calculated ratios were 38.4, 33.4, and 43.4, respectively. This antibody shows only minimal specificity as it links to three of the four species tested in which all three produce similar exotoxins in the pyrogenic toxin superantigens (PTSAgs) family.27,28 In order to obtain higher specificity for this method, the primary antibody, Mouse anti-Streptolysin O, was replaced in favor of Mouse anti-Streptococcus Group A antibody. Figure 6 shows the normalized photoluminescence results of bacteria species with the linked Mouse anti-Streptococcus Group A + dyed secondary antibody complex in solution. The in-solution linkage procedure was performed according to the Experimental Methods section. The presented localized spectra show much greater photoluminescence from the Streptococcus pyogenes sample with a signal that is up to 16.4 times higher than other three bacteria, indicating specificity of only Group A streptococcus. Its peak photoluminescence intensity of 4800 arbitrary units shows significance versus peak photolumines-

Figure 5. Photoluminescence results before (I) and after (II) the Raman normalization procedure of the marked bacteria after primary (antistreptolysin O) and secondary antibody linkage in solution. A: Streptococcus pyogenes, B: Streptococcus pneumoniae, C: Staphylococcus aureus, D: Staphylococcus epidermidis.

Figure 6. Normalized photoluminescence of the marked bacteria after primary (Mouse anti-Streptococcus Group A antibody) and secondary antibody linkage in solution. A: Streptococcus pyogenes, B: Streptococcus pneumoniae, C: Staphylococcus aureus, D: Staphylococcus epidermidis.

cence intensities of less than 400 arbitrary units for the other three measured bacteria samples. In order to conduct an in vivo measurement in the future, the previous procedure that includes 2 h of vigorous mixing and D

DOI: 10.1021/acs.analchem.5b03754 Anal. Chem. XXXX, XXX, XXX−XXX

Article

Analytical Chemistry

rescence, eliciting a positive result for samples containing Streptococcus pyogenes and a negative result for those without. When using the Mouse anti-Streptococcus Group A antibody as the primary antibody, results prove significant signal ratios that are both very low for other tested bacteria while remaining very high for Streptococcus pyogenes both in solution and the solid state. The detection is achieved by combining the high specificity of the standard practice biologic detection methods with the high sensitivity of the spectroscopic methods. This increase in sensitivity efficiently outperforms many standard practice detection methods today by requiring less material and less overall time. This approach has many advantages for future pathogen detection for many bacteria species, particularly in vivo, because the spectroscopy measurement requires very low power and short integration time in addition to biologically compatible antibodies.

multiple washings in order to generate a high ratio between specific and nonspecific binding is not practical. Therefore, the solid-state method in which BSA was added to the PBS solution and to the antibody solutions in order to decrease a nonspecific linkage was examined. Many of the detection methods that involve antigen binding include bovine serum albumin (BSA) 1% (w/v) in the antigen solution. BSA has the ability, because it does not affect biochemical reactions, to significantly increase the signal in assays.29,30 In addition to being biocompatible, it serves as a binding blocker for nonspecific sites, which is useful in noise reduction, particularly in coordination with spectroscopy methods. The solid-state linkage procedure was performed according to the Experimental Methods. In this method, complexes are limited to form only on the surface of the bacterial sample as the bacteria are fixed to the filter similar to in vivo conditions. Figure 7 presents the normalized photoluminescence results of bacteria species with the linked Mouse anti-Streptococcus Group A + dyed secondary antibody complex in solid state.

■

ASSOCIATED CONTENT

* Supporting Information S

The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acs.analchem.5b03754. Full example of the fluorescence normalization process using general bacterial Raman shift at 2930 cm−1 and absorption spectra of all bacteria (PDF)

■

AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected]. Phone: +972 50 4168008 Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS We gratefully acknowledge the professional support of Dr. Ziv Oren. This research was financially supported by Dr. Efraim Rosenbaum, grant number: 245290.

Figure 7. Normalized photoluminescence of the marked bacteria after primary (Mouse anti-Streptococcus Group A antibody) and secondary antibody linkage in solid state. A: Streptococcus pyogenes, B: Streptococcus pneumoniae, C: Staphylococcus aureus, D: Staphylococcus epidermidis.

■

REFERENCES

(1) Kononenko, A. P.; Shustova, K. S. J. Appl. Spectrosc. 1968, 9, 959−960. (2) Pahlow, S.; Meisel, S.; Cialla-May, D.; Weber, K.; Rosch, P.; Popp, J. Adv. Drug Delivery Rev. 2015, 89, 105−120. (3) Lu, X.; Al-Qadiri, H.; Lin, M.; Rasco, B. Food Bioprocess Technol. 2011, 4, 919−935. (4) Culha, M.; Kahraman, M.; Ç am, D.; Sayın, I.; Keseroǧlu, K. Surf. Interface Anal. 2010, 42, 462−465. (5) Jarvis, R. M.; Goodacre, R. Anal. Chem. 2004, 76, 40−47. (6) Jarvis, R. M.; Brooker, A.; Goodacre, R. Anal. Chem. 2004, 76, 5198−5202. (7) Gaus, K.; Rösch, P.; Petry, R.; Peschke, K. D.; Ronneberger, O.; Burkhardt, H.; Baumann, K.; Popp, J. Biopolymers 2006, 82, 286−290. (8) López-Díez, E. C.; Goodacre, R. Anal. Chem. 2004, 76, 585−591. (9) Birks, J.; King, T.; Munro, I. Proc. Phys. Soc., London 1962, 80, 355. (10) Mooradian, A. Phys. Rev. Lett. 1969, 22, 185. (11) Cohen, S.; Margel, S. J. Nanobiotechnol. 2012, 10, 36−36. (12) Liu, Y.; Brandon, R.; Cate, M.; Peng, X.; Stony, R.; Johnson, M. Anal. Chem. 2007, 79, 8796−8802. (13) Gerber, M. A.; Shulman, S. T. Clin. Microbiol. Rev. 2004, 17, 571−580. (14) Schlager, T. A.; Hayden, G. A.; Woods, W. A.; Dudley, S. M.; Hendley, J. Arch. Pediatr. Adol. Med. 1996, 150, 245−248.

These results clearly show that despite the solid-state surfacearea limitation, a very high photoluminescence from the Group A streptococcus complex is achieved in the solid state with a signal that is up to 12.7 times higher than the photoluminescence signal represented from a nonspecific binding, nearly as high as in solution (3800 arbitrary units in solid state versus 4800 in solution). The other three bacteria species showed an expectedly slightly higher photoluminescence in the solid state versus in solution due to a larger volume of unbound dyed antibody present. The solid-state method demonstrates a very high specificity of only Group A streptococcus in a shorter time, with fixed bacteria and minimal washings.

■

CONCLUSION This work presents a new and accurate method for the detection of Streptococcus pyogenes, which affects millions every year. In this novel approach, bacteria samples, treated with a fluorescent antibody complex specific to Streptococcus pyogenes, were volumetrically normalized according to their Raman bacterial signal intensity and then characterized for fluoE

DOI: 10.1021/acs.analchem.5b03754 Anal. Chem. XXXX, XXX, XXX−XXX

Article

Analytical Chemistry (15) Bisno, A. L.; Gerber, M. A.; Gwaltney, J. M., Jr.; Kaplan, E. L.; Schwartz, R. H. Clin. Infect. Dis. 2002, 35, 113−125. (16) Bisno, A. L.; Gerber, M. A.; Kaplan, E. L.; Schwartz, R. H. Clin. Infect. Dis. 1997, 25, 574−583. (17) Gerber, M. A. Pediatr. Infect. Dis. J. 1989, 8, 820−824. (18) Schuman A. J. Contemporary Pediatrics, Peds v2, September 2015. Available at the following: http://contemporarypediatrics. modernmedicine.com/contemporary-pediatrics/news/office-rapidstrep-tests-state-art. (19) Faulques, E.; Massuyeau, F.; Kalashnyk, N.; Perry, D. L. RSC Adv. 2015, 5, 71219−71227. (20) Sinwani, M.; Tischler, Y. R. J. Phys. Chem. C 2014, 118, 14528− 14533. (21) Jeong, S.; Kim, Y. I.; Kang, H.; Kim, G.; Cha, M. G.; Chang, H.; Jung, K. O.; Kim, Y. H.; Jun, B. H.; Hwang, D. W.; Lee, Y. S.; Youn, H.; Lee, Y. S.; Kang, K. W.; Lee, D. S.; Jeong, D. H. Sci. Rep. 2015, 5, 1−9. (22) Lawaetz, A. J.; Stedmon, C. A. Appl. Spectrosc. 2009, 63, 936− 940. (23) Stedmon, C. A.; Markager, S.; Bro, R. Mar. Chem. 2003, 82, 239−254. (24) Linder, J. A.; Bates, D. W.; Lee, G. M.; Finkelstein, J. A. J. Am. Med. Assoc. 2005, 294, 2315−2322. (25) Butler, J. E. Methods 2000, 22, 4−23. (26) Ferraro, J. R.; Nakamoto, K.; Brown, C. W. In Introductory Raman Spectroscopy, 2nd ed.; Ferraro, J. R., Nakamoto, K., Brown, C. W., Eds.; Academic Press: San Diego, 2003; pp 295−324. (27) Dinges, M. M.; Orwin, P. M.; Schlievert, P. M. Clin. Microbiol. Rev. 2000, 13, 16−34. (28) Bhakdi, S.; Bayley, H.; Valeva, A.; Walev, I.; Walker, B.; Weller, U.; Kehoe, M.; Palmer, M. Arch. Microbiol. 1996, 165, 73−79. (29) Kreader, C. A. Environ. Microbiol. 1996, 62, 1102−1106. (30) Hirayama, K.; Akashi, S.; Furuya, M.; Fukuhara, K. I. Biochem. Biophys. Res. Commun. 1990, 173, 639−646.

F

DOI: 10.1021/acs.analchem.5b03754 Anal. Chem. XXXX, XXX, XXX−XXX