Technical Note pubs.acs.org/ac
Rapid Enumeration of Phage in Monodisperse Emulsions Katrina F. Tjhung,† Sean Burnham,‡ Hany Anany,‡ Mansel W. Griffiths,‡,§ and Ratmir Derda*,† †
Department of Chemistry, University of Alberta, Edmonton, AB T6G 2G2, Canada Department of Food Science, University of Guelph, Guelph, ON N1G 2W1, Canada § Canadian Research Institute for Food Safety, University of Guelph, Guelph, ON N1G 2W1, Canada ‡
S Supporting Information *
ABSTRACT: Phage-based detection assays have been developed for the detection of viable bacteria for applications in clinical diagnosis, monitoring of water quality, and food safety. The majority of these assays deliver a positive readout in the form of newly generated progeny phages by the bacterial host of interest. Progeny phages are often visualized as plaques, or holes, in a lawn of bacteria on an agar-filled Petri dish; however, this ratelimiting step requires up to 12 h of incubation time. We have previously described an amplification of bacteriophages M13 inside droplets of media suspended in perfluorinated oil; a single phage M13 in a droplet yields 107 copies in 3−4 h. Here, we describe that encapsulation of reporter phages, both lytic T4-LacZ and nonlytic M13, in monodisperse droplets can also be used for rapid enumeration of phage. Compartmentalization in droplets accelerated the development of the signal from the reporter enzyme; counting of “positive” droplets yields accurate enumeration of phage particles ranging from 102 to 106 pfu/mL. For enumeration of T4-LacZ phage, the fluorescent signal appeared in as little as 90 min. Unlike bulk assays, quantification in emulsion is robust and insensitive to fluctuations in environmental conditions (e.g., temperature). Power-free emulsification using gravity-driven flow in the absence of syringe pumps and portable fluorescence imaging solutions makes this technology promising for use at the point of care in lowresource environments. This droplet-based phage enumeration method could accelerate and simplify point-of-care detection of the pathogens for which reporter bacteriophages have been developed.
I
cis,10 and the FDA-approved KeyPath MRSA/MSSA Blood Culture Test − BT, for Staphylococcus aureus.11 Most of these detection methods depend on engineered reporter phages that confer properties to their target bacteria in the form of enzyme,4 fluorescent7 or chemiluminescent8 proteins. While amplification of phage is significantly faster than growth of bacteria, detection of phage progeny still require lengthy incubation steps. For example, a classical method for detection of phageformation of plaques in a lawn of reporter bacteria encapsulated in agar (Figure 1)requires an overnight incubation to allow for sufficient amplification of phage. Development of the signal from the reporter protein could also require lengthy time for development of a detectable and quantifiable signal. Other methods for detection of the pathogen-produced progeny phage exist; examples are quantitative real-time PCR, MALDI-TOF, and competitive ELISA (see review2 and references within). These techniques, however, rely on complex instrumentation, which might not be compatible with point-of-care (POC) detection. We have previously demonstrated that libraries of phages can propagate in monodisperse water−oil emulsions and that a single phage in an aqueous droplet can multiply by a factor of
n this Technical Note, we describe an emulsion-based method for rapid detection of bacteriophages that contain enzymatic reporter genes (“reporter bacteriophage”). Traditionally, detection and identification of pathogenic bacteria for clinical and food and water safety applications largely depend on culture-based assays, where collected bacterial samples must be grown to sufficient density for characterization. In the past 50 years, bacteriophage-based assays have been developed to offer a rapid and specific method for the detection of pathogenic bacteria that negates time-consuming primary culture.1,2 Phage-based detection assays capitalize on the hostspecificity and rapid amplification of bacteriophages. As phage amplify faster than the host, detection of phage progeny is faster than detection of bacterial growth: for example, Escherichia coli phage generates a burst of 100 to 1000 progeny every 40 min, whereas E. coli doubles every 20 min. This property is especially useful for detection of slow-growing bacteria, such as Mycobacterium tuberculosis. This bacteria requires 30 days to yield visible colonies, whereas the assays based on detection of amplified mycobacterium phage require only 24−48 h.3 Bacteriophages are able to quickly transfer genetic material to specific bacterial hosts, which then produce and amplify these gene products. Phage-borne transfer of detectable proteins to bacteria permit a rapid and robust detection of specific pathogens, with existing tests for E. coli,4 Salmonella,5 Shigella,6 Mycobacterium tuberculosis,7,8 Yersinia pestis,9 Bacillus anthra© 2014 American Chemical Society
Received: January 19, 2014 Accepted: May 20, 2014 Published: May 20, 2014 5642
dx.doi.org/10.1021/ac500244g | Anal. Chem. 2014, 86, 5642−5648
Analytical Chemistry
Technical Note
Figure 1. Comparison of time courses of phage enumeration using plaque-formation assay in agar overlay and phage enumeration in monodisperse emultions. Signal development in monodisperse emulsion shortens detection time to ∼2.5 h, compared to classic agar overlay assays requiring an overnight incubation.
107 in approximately 4 h.12,13 It is also known that encapsulation of enzymes in micrometer-sized compartments can accelerate the enzymatic assay by increasing the relative concentration of the enzyme and its substrate.14 Enzymatic assays in droplets have been used as early as the 1960s to detect single molecules of β-D-galactosidase.15 Recently, they have been empowered by development of microfluidic tools for rapid generation and analysis of monodisperse emulsions.16−18 In this Technical Note, we use enzymatic assays in droplets to detect the phage amplified within a droplet. This combination allows us to visualize the presence of phage in emulsion culture and yield a rapid method for the quantification of reporter phage. Co-encapsulation of reporter phage, its host bacterium, and a substrate for the enzymatic reporter into isolated, spherical microdroplets (165 μm in diameter) leads to rapid development of a positive signal and yields a robust quantification of concentrations of phage spanning 5 orders of magnitude (102−106 pfu/mL). Reporter detection in emulsion is more rapid than plaque-forming assay in agar and simpler than PCR-based detection or ELISA. We also explored the feasibility of power-free emulsification and portable fluorescence imaging alternatives to further improve accessibility of phage-based detection technologies.
Biotechnologies.13,21 The aqueous phase contained the LB medium supplemented with phage, log-phase bacterial host at a 1:10 dilution, 60 μM FDG, 185 nM Alexa Fluor 546 (Invitrogen, Eugene, OR), and 170 μM isopropyl thiogalactopyranoside (IPTG) (Thermo Fisher, Waltham, MA). We validated the phage concentrations in the aqueous phase using the well-established plaque formation assay in agar overlay (see examples for M1322 and T423). We collected droplets, which contained LB, bacteria, and phage, in a 3 cm Petri dish containing 0.5 mL of HFE-7500 until a continuous layer of droplets was formed. The remaining nonemulsified aqueous phase was distributed onto a 96-well plate at 150 μL per well to compare “bulk” detection of the reporter signal. To image the droplets, we sampled approximately 100 μL of emulsion from a Petri dish and deposited the emulsion onto a glass slide to form a droplet monolayer. The droplets were imaged within 3−5 min using a fluorescence scanner at a 25 μm resolution (Typhoon FLA 9500, GE Healthcare Life Sciences, Uppsala, Sweden). All droplet sampling and imaging was performed in triplicate. We used a 473 nm emission filter to detect fluorescein, which was generated from FDG by the transduced β-galactosidase. Additionally, a 532 nm emission filter was used to visualize Alexa Fluor 546 present in the media in all droplets. The time course of signal development was established by sampling and imaging the emulsions every 30 min. We observed that the Typhoon FLA 9500 produces a halo during scanning (Figure S3 of the Supporting Information). The imaging artifact causes a minor (100 fold. In
Figure 4. Development of fluorescent signal inside individual droplets containing E. coli and (A, B, and C) T4-LacZ phage or (D, E, and F) M13 phage. The signal was measure by averaging 16-bit gray scale intensity f 10 × 10 pixel images inside the positive droplets (n = 20 for A, B, and C; n = 30 for D, E, and F). Droplets that encapsulate different concentrations of phage reach similar fluorescence intensity at saturation; however, due to imaging artifact (fluorescent halo, Figure S3 of the Supporting Information), intensity of individual droplets appeared higher in images that contain higher fraction of fluorescent droplets. All data are shown. The height of the rectangles are equal to 2 × (standard deviation). The line designates sample mean. Raw pixel intensities were used.
number of phage, we compared the number of fluorescent droplets detected in emulsion with the total plaque forming units (PFU) of phage detected by conventional agar overlay assay. We assumed that the distribution of phage to droplets 5645
dx.doi.org/10.1021/ac500244g | Anal. Chem. 2014, 86, 5642−5648
Analytical Chemistry
Technical Note
not be detected due to averaging of the infection events in a large population of bacteria and diffusion of the turned-over substrate. Emulsification of phage accelerates enumeration and makes this process more robust to fluctuations in the environment (e.g., temperature). Separation of phage to droplets of the emulsion replaces classical encapsulation in agar (agar overlay assay). Droplet generation, in turn, can be performed by a variety of methods and instruments, some of which have been developed into commercial instruments and some have been designed to be compatible with low-resource environments.31,32 We demonstrate that droplet encapsulation simplifies the enumeration of phage and relaxes the detection constraints. An analogous increase in robustness has been recently reported for “digital” assays that involve isothermal amplification of DNA: loop-mediated isothermal amplification (LAMP) in monodisperse compartments is significantly more robust than analogous amplification in bulk.32 Detection time in droplets is suitable for point-of-care (POC) detection of phage or phage-producing pathogens. The current infrastructure for droplet generation and imaging, however, is not designed for POC assays. For example, we used syringe pumps for emulsification and a fluorescence imager for detection. To allow for POC detection, both droplet production and imaging can be performed with simpler techniques, which are amenable for low-resource environments. For example, droplets can be produced without any syringe pumps using vacuum-syringe microfluidics31 or gravity-driven flow. By placing the aqueous and perfluorinated compartments at ∼60 cm height from the channel, we produced droplets of the desired size, albeit at a 10fold slower rate (Figure 6, panels A−C). Portable solutions for imaging of fluorescent droplets have been developed,33 and they can be readily integrated into this assay. In our version of a “portable imaging solution”, we imaged the fluorescent signal by illuminating the sample from ∼10 cm distance at a 45degree angle with a 3 W, 365 nm UV LED (Shenzhen, China) and imaged the sample at a 90-degree angle with an iPhone-4 camera (model A1332). Sample and LED were located inside a cardboard box; the phone was attached on the outside. 45degree illumination minimized the backscatter of the excitation light and allowed fluorescence imaging without any excitation or emission filters. The resulting image can be transferred to a centralized location for quantification. Alternatively, it can be quantified by custom image analysis software that can quantify the number of green pixels in a sample and a control sample (6D); if necessary, a built-in color strip can be used for adjustment of the image to different lighting conditions.34 Such analysis could be simple to implement on any smartphone with a compatible digital camera and any operational system (OS, Android, etc.). Adding objectives and emission filters to the smartphone can further increase the resolution and signal-tonoise of fluorescence imaging.32 We envision that optimization of the imaging and droplet generation solution can be readily performed, and the final assay can be produced as an all-in-one integrated emulsification/incubation/imaging device. While this method can be performed by minimally trained personnel (undergraduate students), its direct implementation by the nonspecialist still could be difficult because droplet generation, incubation, and imaging are performed as three disconnected steps. Combining these steps into one platform can make this assay more user-friendly. Examples of microemulsion-based integrated platforms for bacterial culture in droplets have been reported.35
Figure 5. Fraction of fluorescent droplets as a function of phage concentration for (A) T4-LacZ phage and (B) M13KE phage. The fraction was defined as the ratio of fluorescent vs total droplets in randomly selected rectangular frames with ∼100−400 total droplets. Phage concentration was determined by plaque-forming assay (see Supporting Information for details). Each data point represents an independent emulsification experiment. Vertical error bars are 2 × (st dev) of ratio (st dev = standard deviation from mean value measured in 4−5 independent areas); horizontal bars are 2 × (st dev) of titer. Relative standard error of the tittering procedure in agar overlay (horizontal error bar) is ∼10−20% (Figure S1 of the Supporting Information). The dashed line represents theoretical fractions (f) of positive droplets at phage concentration Cp calculated from the Poisson distribution as f = (1 − exp(−Cp/CN), where CN = 4.5 × 105 droplets/mL for 165 μm droplets.
contrast, the droplet assay could reliably differentiate between closely related phage concentrations (Figures 3B and 5B). The variability of the fluorescence signal signal in the reporter phage assay in droplets (Figure 4, panels A−C for phage T4-LacZ and panels D−F for M13) was significantly higher than in bulk solution (Figure 2, panels A−D, respectively) due to stochastic variability in the rate of infection of individual phage. In T4 phage infection, the variability was the greatest at the point in the infection process where fluorescence intensity was between 20 and 80% of the maximum (ca. 100−140 min in Figure 4), and it decreased as signal reached saturation and all bacteria were lysed (Figure 4). The variability of M13 phage infection was greater than that of T4-LacZ phage (Figure 3); it might originate from the variability in the diffusion rate of FDG into and out of individual bacteria. It is also known that M13 has much higher variability in the number of phage progeny produced, when compared to T4-LacZ phage.30 In bulk, this variability could 5646
dx.doi.org/10.1021/ac500244g | Anal. Chem. 2014, 86, 5642−5648
Analytical Chemistry
Technical Note
For detection of phage originating from pathogens other than E. coli, the components of the assay can be interchanged based on known phage-based detection schemes. Emulsion-based enumeration can be applied to two types of phage (lytic and nonlytic) and two types of reporter systems (mature enzyme expression and protein complementation). If multiple reporter phage types are available, it is beneficial to use the lytic phage to accelerate the assay by alleviating the need for intracellular delivery of the enzyme substrate (here, FDG). Furthermore, phage that transduce the full-length β-galactosidase (β-gal) could accelerate the assay by eliminating the time required for assembly of the enzyme from LacZα and LacZω components. The droplet assay confirmed the existence of stochastic variability in the phage-reporter assay. Although variability in isogenic populations has been known since the 1940s,36 these effects were difficult to study in bulk solutions because they require isolation of individual phage and/or bacteria. Droplets provide a convenient means for large-scale compartmentalization and simplify investigation of phage-host interactions at a single phage level.
■
ASSOCIATED CONTENT
S Supporting Information *
T4-LacZ propagation, detailed emulsification protocol, highresolution images of droplets containing T4-LacZ phage and E. coli, and additional information as noted in text. This material is available free of charge via the Internet at http://pubs.acs.org.
■
AUTHOR INFORMATION
Corresponding Author
*E-mail:
[email protected]. Notes
The authors declare no competing financial interest.
■
ACKNOWLEDGMENTS This work was supported by SENTINEL Bioactive Paper Network and the Alberta Glycomics Centre. Infrastructure support has been provided by the Canada Foundation for Innovation (CFI). K.F.T. acknowledges Alberta InnovatesHealth Solutions (AIHS) and the Natural Sciences and Engineering Research Council of Canada (NSERC) for fellowship support.
Figure 6. Power-free generation of droplets in microfluidics by gravitydriven flow. (A) Scheme of the device. We placed two syringes filled with aqueous and perfluorinated media at different heights (see Experimental procedures for more details) and measured the droplet sizes and rates of formation as a function of heights (Hf and Ha denote the height of the perfluoro phase and the aqueous phase, respectively; see Table S1 of the Supporting Information for numeric values). (B) Representative images of droplets. (C) The droplet sizes (micrometer, size-coded) and rates of formation (Hz, color-coded) as a function of heights, roman numerals correspond to images in (B). (D) Imaging of droplets by blue LED and iPhone camera. (E) Droplets formed from 103 and 104 pfu/mL of M13KE phage can be readily distinguished by iPhone-LED imaging. Abbreviations: HFE, hydrofluoroether; LB, Lysogeny Broth; LED, light emitting diode.
■
REFERENCES
(1) Schofield, D. A.; Sharp, N. J.; Westwater, C. Bacteriophage 2012, 2, 105−283. (2) Smartt, A. E.; Xu, T.; Jegier, P.; Carswell, J. J.; Blount, S. A.; Sayler, G. S.; Ripp, S. Anal. Bioanal. Chem. 2012, 402, 3127−3146. (3) Minion, J.; Pai, M. The International Journal of Tuberculosis and Lung Disease 2010, 14, 941−951. (4) Derda, R.; Lockett, M. R.; Tang, S. K.; Fuller, R. C.; Maxwell, E. J.; Breiten, B.; Cuddemi, C. A.; Ozdogan, A.; Whitesides, G. M. Anal. Chem. 2013, 85, 7213−7220. (5) Fernandes, E.; Martins, V. C.; Nobrega, C.; Carvalho, C. M.; Cardoso, F. A.; Cardoso, S.; Dias, J.; Deng, D.; Kluskens, L. D.; Freitas, P. P.; Azeredo, J. Biosens. Bioelectron. 2013, 52c, 239−246. (6) Goodridge, L. D. Bacteriophage 2013, 3, e25098. (7) Jain, P.; Hartman, T. E.; Eisenberg, N.; O’Donnell, M. R.; Kriakov, J.; Govender, K.; Makume, M.; Thaler, D. S.; Hatfull, G. F.; Sturm, A. W.; Larsen, M. H.; Moodley, P.; Jacobs, W. R., Jr. J. Clin. Microbiol. 2012, 50, 1362−1369. (8) Riska, P. F.; Su, Y.; Bardarov, S.; Freundlich, L.; Sarkis, G.; Hatfull, G.; Carriere, C.; Kumar, V.; Chan, J.; Jacobs, W. R., Jr. J. Clin. Microbiol. 1999, 37, 1144−1149.
In conclusion, we developed a droplet-based assay for the enumeration of bacteriophage. Conventional phage-based detection of pathogens requires four steps: (i) sample processing that enriches bacteria, (ii) infection of the bacteria and encapsulation with the secondary host in agar, (iii) incubation, and (iv) detection. Here, we show that replacing encapsulation in agar [step (ii)] by encapsulation in monodisperse emulsions decreases the detection time when compared to agar overlay and simplifies phage enumeration when compared to a bulk reporter assay. This method of phage enumeration could be applied to any existing phage-based detection assay to accelerate the final signal development step. 5647
dx.doi.org/10.1021/ac500244g | Anal. Chem. 2014, 86, 5642−5648
Analytical Chemistry
Technical Note
(9) Schofield, D. A.; Molineux, I. J.; Westwater, C. J. Clin. Microbiol. 2009, 47, 3887−3894. (10) Schofield, D. A.; Sharp, N. J.; Vandamm, J.; Molineux, I. J.; Spreng, K. A.; Rajanna, C.; Westwater, C.; Stewart, G. C. J. Microbiol. Methods 2013, 95, 156−161. (11) Bhowmick, T.; Mirrett, S.; Reller, L. B.; Price, C.; Qi, C.; Weinstein, M. P.; Kirn, T. J. J. Clin. Microbiol. 2013, 51, 1226−1230. (12) Derda, R.; Tang, S. K.; Whitesides, G. M. Angew. Chem., Int. Ed. 2010, 49, 5301−5304. (13) Matochko, W. L.; Ng, S.; Jafari, M. R.; Romaniuk, J.; Tang, S. K.; Derda, R. Methods (San Diego, Calif.) 2012, 58, 18−27. (14) Vincent, M. E.; Liu, W.; Haney, E. B.; Ismagilov, R. F. Chem. Soc. Rev. 2010, 39, 974−984. (15) Rotman, B. Proc. Natl. Acad. Sci. U.S.A. 1961, 47, 1981−1991. (16) Anna, S. L.; Bontoux, N.; Stone, H. A. Appl. Phys. Lett. 2003, 82, 364−366. (17) Garstecki, P.; Fuerstman, M. J.; Whitesides, G. M. Phys. Rev. Lett. 2005, 94, 234502. (18) Boedicker, J. Q.; Li, L.; Kline, T. R.; Ismagilov, R. F. Lab Chip 2008, 8, 1265−1272. (19) Goodridge, L. D. A Reporter Bacteriophage: Beta-Galactosidase Assay for Detection of Generic Escherichia coli from Beef Carcasses. Ph.D. Thesis, University of Guelph, 2002. (20) Noren, K. A.; Noren, C. J. Methods (San Diego, Calif.) 2001, 23, 169−178. (21) Clausell-Tormos, J.; Lieber, D.; Baret, J. C.; El-Harrak, A.; Miller, O. J.; Frenz, L.; Blouwolff, J.; Humphry, K. J.; Koster, S.; Duan, H.; Holtze, C.; Weitz, D. A.; Griffiths, A. D.; Merten, C. A. Chem. Biol. 2008, 15, 427−437. (22) Barbas, C. F.; Burton, D. R.; Scott, J. K.; Silverman, G. J. Phage Display: A Laboratory Manual; Cold Spring Harbor Laboratory Press: Cold Spring Harbor, NY, 2004. (23) Sambrook, J.; Fritsch, E. F.; Maniatis, T. Molecular Cloning: A Laboratory Manual; Cold Spring Harbor Laboratory Press: Cold Spring Harbor, NY, 1989. (24) Plovins, A.; Alvarez, A. M.; Ibanez, M.; Molina, M.; Nombela, C. Appl. Environ. Microbiol. 1994, 60, 4638−4641. (25) Stent, G. S. Molecular Biology of Bacterial Viruses; Freeman: Stuttgart, Germany, 1963. (26) Chen, Y.; Wijaya Gani, A.; Tang, S. K. Lab Chip 2012, 12, 5093−5103. (27) Woronoff, G.; El Harrak, A.; Mayot, E.; Schicke, O.; Miller, O. J.; Soumillion, P.; Griffiths, A. D.; Ryckelynck, M. Anal. Chem. 2011, 83, 2852−2857. (28) Skhiri, Y.; Gruner, P.; Semin, B.; Brosseau, Q.; Pekin, D.; Mazutis, L.; Goust, V.; Kleinschmidt, F.; El Harrak, A.; Hutchison, J. B.; Mayot, E.; Bartolo, J.-F.; Griffiths, A. D.; Taly, V.; Baret, J.-C. Soft Matter 2012, 8, 10618−10627. (29) Nichtl, A.; Buchner, J.; Jaenicke, R.; Rudolph, R.; Scheibel, T. J. Mol. Biol. 1998, 282, 1083−1091. (30) De Paepe, M.; De Monte, S.; Robert, L.; Lindner, A. B.; Taddei, F. PLoS One 2010, 5, e11823. (31) Abate, A. R.; Weitz, D. A. Biomicrofluidics 2011, 5, 14107. (32) Selck, D. A.; Karymov, M. A.; Sun, B.; Ismagilov, R. F. Anal. Chem. 2013, 85, 11129−11136. (33) Hatch, A. C.; Fisher, J. S.; Tovar, A. R.; Hsieh, A. T.; Lin, R.; Pentoney, S. L.; Yang, D. L.; Lee, A. P. Lab Chip 2011, 11, 3838− 3845. (34) Funes-Huacca, M.; Wu, A.; Szepesvari, E.; Rajendran, P.; KwanWong, N.; Razgulin, A.; Shen, Y.; Kagira, J.; Campbell, R.; Derda, R. Lab Chip 2012, 12, 4269−4278. (35) Park, J.; Kerner, A.; Burns, M. A.; Lin, X. N. PLoS One 2011, 6, e17019. (36) Delbruck, M. J. Bacteriol. 1945, 50, 131−135.
5648
dx.doi.org/10.1021/ac500244g | Anal. Chem. 2014, 86, 5642−5648
