Development of an Ultrasensitive Dual-Channel Flow Cytometer for

Mar 4, 2009 - Shaobin Zhu , Ling Ma , Shuo Wang , Chaoxiang Chen , Wenqiang Zhang , Lingling Yang , Wei Hang , John P. Nolan , Lina Wu , and Xiaomei ...
2 downloads 0 Views 1MB Size
Anal. Chem. 2009, 81, 2555–2563

Development of an Ultrasensitive Dual-Channel Flow Cytometer for the Individual Analysis of Nanosized Particles and Biomolecules Lingling Yang, Shaobin Zhu, Wei Hang, Lina Wu, and Xiaomei Yan* Department of Chemical Biology, College of Chemistry and Chemical Engineering, The Key Laboratory for Chemical Biology of Fujian Province, The Key Laboratory of Analytical Science of the Ministry of Education, Xiamen University, Xiamen 361005, China A compact, high-sensitivity, dual-channel flow cytometer (HSDCFCM) was developed for the individual analysis of nanosized particles and biomolecules. A hydrodynamic focusing technique was applied to confine the sample stream and enable small probe volume. Fluorescence bursts from single R-phycoerythrin (R-PE) molecules passing through the laser beam were well resolved from the background with signal-to-noise ratio of 17. Excellent size discrimination was demonstrated with a mixture of three sizes of polystyrene nanoparticles. Simultaneous measurement of fluorescence and light scattering signals from individual nanoparticles was demonstrated with the 100 nm fluorescent latex beads. Doxorubicin-loaded ZrO2 nanoparticles and fluorescently stained Escherichia coli ER2738 cells were analyzed successfully with dual-channel detection. Particle counting is demonstrated with the 210 nm fluorescent latex beads, and excellent correlation (R2 > 0.998) between the manufacturer-reported concentrations and those measured by HSDCFCM enumeration was obtained. The measured sample detection efficiency was ∼90% on average for particle concentrations ranging from 1.62 × 105 to 3.93 × 107 particles/mL. Sample mixtures with varying proportions of fluorescently labeled and unlabeled nanoparticles were also analyzed with good ratio correspondence. By providing rapid, quantitative, and multiparameter characterization of nanoparticles, it is believed that the HSDCFCM will find many applications in the fields of bionanotechnology, bioanalytical chemistry, and biomedicine. Detection of nanosized particles is becoming increasingly important due to their wide bioapplications in labeling, delivering, therapeutics, sensing, separation, and purification.1-6 Besides, a wide variety of biological particles such as bacteria, viruses, * To whom correspondence should be addressed. Phone: 86-592-2184519. Fax: 86-592-2189959. E-mail: [email protected]. (1) Davis, M. E.; Chen, Z. G.; Shin, D. M. Nat. Rev. Drug Discovery 2008, 7, 771–782. (2) LaVan, D. A.; McGuire, T.; Langer, R. Nat. Biotechnol. 2003, 21, 1184– 1191. (3) Sun, C.; Lee, J. S.; Zhang, M. Adv. Drug Delivery Rev. 2008, 60, 1252– 1265. (4) Scheerlinck, J. P.; Greenwood, D. L. Drug Discovery Today 2008, 13, 882– 887. 10.1021/ac802464a CCC: $40.75  2009 American Chemical Society Published on Web 03/04/2009

subcellular organelles, molecular assemblies, and molecules all fall into the nanometer to submicrometer scale. The rapid monitoring of the nanoparticle size and concentration during all preparative stages, the examination of the density and functionality of ligands immobilized onto the nanoparticle surface, and the evaluation of biochemical attributes (nucleic acid content, enzymatic activity, or antigenic determinants) of biological nanoparticles are essential for most biomedical applications. Although electron microscopy (transmission electron microscopy (TEM) and scanning electron microscopy (SEM)) is a reliable method for the visual observation and sizing of nanoparticles, it requires the sample to be fixed and stained with contrast agents. Atomic force microscopy (AFM) is a scanning technique and requires long data acquisition times. Dynamic light scattering (DLS) is able to provide rapid and accurate size distribution for particles over the size ranging from a few nanometers to a few micrometers by measuring ensemble averages.7 However, the above tools commonly used for nanoparticle sizing cannot provide convenient concentration measurements. The extinction spectra of colloidal gold can be used for a simple and fast determination of the size and concentration of nanoparticles,8 yet the biochemical characteristic of nanoparticles cannot be displayed simultaneously. Recently, fluorescence correlation spectroscopy has been used for the size and concentration measurement of nanoparticles, as well as surface coating characterization.9-11 Clearly, there is a great necessity for the development of advanced approaches enabling simultaneous determination of the size, concentration, and biochemical properties of nanosized particles. Optical flow cytometry is a common and well-established method for counting and sorting particles on a large scale by providing rapid, quantitative, and sensitive multiparameter measurements.12,13 Recently, flow cytometry on microfabricated fluidic devices has been developed for cell sorting and quantifica(5) Sperling, R. A.; Rivera Gil, P.; Zhang, F.; Zanella, M.; Parak, W. J. Chem. Soc. Rev. 2008, 37, 1896–1908. (6) Wilson, R. Chem. Soc. Rev. 2008, 37, 2028–2045. (7) Berne, B. J.; Pecora, R. Dynamic Light Scattering: With Application to Chemistry, Biology, and Physics; Dover Publications, 2000. (8) Khlebtsov, N. G. Anal. Chem. 2008, 80, 6620–6625. (9) Kuyper, C. L.; Fujimoto, B. S.; Zhao, Y.; Schiro, P. G.; Chiu, D. T. J. Phys. Chem. B 2006, 110, 24433–24441. (10) Tcherniak, A.; Prakash, A.; Mayo, J. T.; Colvin, V. L.; Link, S. J. Phys. Chem. C 2009, 113, 844–848. (11) Jurkiewicz, P.; Konak, C.; Subr, V.; Hof, M.; Stepanek, P.; Ulbrich, K. Macromol. Chem. Phys. 2008, 209, 1447–1453. (12) Shapiro, H. M. Practical Flow Cytometry, 4th ed.; Wiley-Liss: NJ, 2003.

Analytical Chemistry, Vol. 81, No. 7, April 1, 2009

2555

tion of bacterial cells using fluorescent staining.14-17 Fluorescently labeled and unlabeled latex particles with diameters of 1 and 2 µm were detected, counted, and distinguished from each other based on their light scattering intensity and fluorescence on a microfabricated device employing electrokinetic focusing.18 Online measurement of the mean size and dispersity of cadmium selenide nanoparticles during the formation process has been reported using fluorescence detection in a continuous flow chipbased microreactor.19 However, one limitation of conventional flow cytometry is the inability to measure small particles less than 0.5 µm or dim particles having less than several hundred fluorescent molecules. In recent years, single-molecule detection (SMD) techniques have been rapidly developed with significant advances in optics, fluidics, and optical design for improved detection in a flowing system.20-24 Especially, researchers at Los Alamos National Laboratory have done a good job in detecting and identifying a single fluorescent molecule in fluid solution.25-33 SMD makes it possible to detect single nanoparticle by fluorescent labeling of the biological species.34,35 Individual analysis of particles offers the advantage of statistical treatment by revealing the intrinsic heterogeneity that is not disclosed by ensemble analysis. The objective of the present work was to develop a compact and high-sensitivity dual-channel flow cytometer (HSDCFCM) specifically for the individual analysis of nanosized particles and biomolecules in a rapid and quantitative manner. A hydrodynamic focusing technique enables the reduction of the probe volume and the background signal from the solvent for SMD detection. By measuring fluorescence and light scattering simultaneously, (13) Gilman-Sachs, A. Anal. Chem. 1994, 66, 700A–707A. (14) Dittrich, P. S.; Schwille, P. Anal. Chem. 2003, 75, 5767–5774. (15) Fu, A. Y.; Chou, H. P.; Spence, C.; Arnold, F. H.; Quake, S. R. Anal. Chem. 2002, 74, 2451–2457. (16) Sakamoto, C.; Yamaguchi, N.; Yamada, M.; Nagase, H.; Seki, M.; Nasu, M. J. Microbiol. Methods 2007, 68, 643–647. (17) McClain, M. A.; Culbertson, C. T.; Jacobson, S. C.; Ramsey, J. M. Anal. Chem. 2001, 73, 5334–5338. (18) Schrum, J. P.; Culbertson, C. T.; Jacobson, S. C.; Ramsey, J. M. Anal. Chem. 1999, 71, 4173–4177. (19) Krishnadasan, S.; Tovilla, J.; Vilar, R.; deMello, A. J.; deMello, J. C. J. Mater. Chem. 2004, 14, 2655–2660. (20) Guenard, R. D.; King, L. A.; Smith, B. W.; Winefordner, J. D. Anal. Chem. 1997, 69, 2426–2433. (21) Haab, B. B.; Mathies, R. A. Anal. Chem. 1995, 67, 3253–3260. (22) Lyon, W. A.; Nie, S. M. Anal. Chem. 1997, 69, 3400–3405. (23) Ma, Y.; Shortreed, M. R.; Yeung, E. S. Anal. Chem. 2000, 72, 4640–4645. (24) Turner, E. H.; Lauterbach, K.; Pugsley, H. R.; Palmer, V. R.; Dovichi, N. J. Anal. Chem. 2007, 79, 778–781. (25) Ambrose, W. P.; Goodwin, P. M.; Jett, J. H.; Orden, A. V.; Werner, J. H.; Keller, R. A. Chem. Rev. 1999, 99, 2929–2956. (26) Dovichi, N. J.; Martin, J. C.; Jett, J. H.; Keller, R. A. Science 1983, 219, 845–847. (27) Habbersett, R. C.; Jett, J. H. Cytometry, Part A 2004, 60, 125–134. (28) Keller, R.; Ambrose, W.; Goodwin, P.; Jett, J.; Martin, J.; Wu, M. Appl. Spectrosc. 1996, 50, A12–A32. (29) Orden, A. V.; Cai, H.; Goodwin, P. M.; Keller, R. A. Anal. Chem. 1999, 71, 2108–2116. (30) Shera, E. B.; Seitzinger, N. K.; Davis, L. M.; Keller, R. A.; Soper, S. A. Chem. Phys. Lett. 1990, 174, 553–557. (31) Yan, X.; Grace, W. K.; Yoshida, T. M.; Habbersett, R. C.; Velappan, N.; Jett, J. H.; Keller, R. A.; Marrone, B. L. Anal. Chem. 1999, 71, 5470–5480. (32) Werner, J. H.; Cai, H.; Jett, J. H.; Reha-Krantz, L.; Keller, R. A.; Goodwin, P. M. J. Biotechnol. 2003, 102, 1–14. (33) Castro, A.; Williams, J. G. Anal. Chem. 1997, 69, 3915–3920. (34) Ferris, M. M.; McCabe, M. O.; Doan, L. G.; Rowlen, K. L. Anal. Chem. 2002, 74, 1849–1856. (35) Zhao, X. J.; Hilliard, L. R.; Mechery, S. J.; Wang, Y. P.; Bagwe, R. P.; Jin, S. G.; Tan, W. H. Proc. Natl. Acad. Sci. U.S.A. 2004, 101, 15027–15032.

2556

Analytical Chemistry, Vol. 81, No. 7, April 1, 2009

Figure 1. Schematic diagram of the high-sensitivity dual-channel flow cytometer (HSDCFCM): P, half-wave plate; S, polarizing beam splitter; M, mirror; L, achromatic-doublet lens; C, flow cell; A1 and A2, aspheric lenses; Dic-F, dichroic filter; EF, edge filter; BPF, bandpass filter; NDF, neutral density filter; APD, single-photon counting avalanche photodiode; PMT, photomultiplier tube.

biochemical and size information of individual nanoparticles can be characterized simultaneously. Excellent correlation (R2 > 0.998) between the manufacturer-reported concentrations and those measured by HSDCFCM enumeration was obtained for the 210 nm fluorescent latex beads. For particle concentrations ranging from 1.62 × 105 to 3.93 × 107 particles/mL, the measured sample detection efficiencies were ∼91% and ∼85% by fluorescence and dual-channel coincidence measurements, respectively. Sample mixtures with varying proportions of fluorescently labeled and unlabeled nanoparticles were also analyzed with good ratio correspondence. It is believed that the HSDCFCM will find increasing use in a wide range of applications in biomedicine and environmental monitoring by providing rapid, quantitative, sensitive, and multiparameter characterization of individual nanoparticles. EXPERIMENTAL SECTION Instrument Design. The HSDCFCM was designed to collect light emission on two separate channels: one channel for fluorescence and one channel for light scattering. A schematic drawing of the optical layout of the laboratory-built HSDCFCM is shown in Figure 1; an orthogonal configuration was employed. A 532 nm continuous-wave solid-state Nd: YAG laser (CryLas GmbH, Berlin, Germany, maximum output 15 mW) was used as the excitation source. A polarizing beam splitter and a half-wave plate (CVI Laser Corp., Albuquerque, NM) provided polarization control and continuously variable attenuation of the laser light. Laser excitation power (measured after mirror reflection) of 0.75 mW was used in the present study. The 0.9 mm diameter of laser output beam was focused to an ∼10 µm diameter spot (1/e2) by an achromatic-doublet lens of 12.7 mm focal length (AC064013-A1, Thorlabs, Newton, NJ) onto the hydrodynamically focused sample stream inside a 250 µm square quartz flow channel (NSG Precision Cells, Farmingdale, NY). The emitted light from the sample stream was collected by a molded-glass Geltech aspheric lens of 3.1 mm focal length (C330TM-A, 40×, 0.68 N.A., Thorlabs) and then directed by a dichroic beam

splitter (FF555-Di02, Semrock Inc., Rochester, NY) into two distinct light paths for fluorescence and side scatter (SS) detection, respectively. Approximately 98% of a wavelength value from 493 to 548 nm was reflected, and approximately 90% of a wavelength value from 562 to 745 nm was transmitted. To eliminate stray light from the excitation source, the transmitted fluorescence light was passed through a Raman edge filter (LP03-532RS, Semrock). The fluorescence emission was then spectrally filtered by a band-pass filter (FF01-579/34, Semrock) and focused by another molded-glass Geltech aspheric lens (C240TM-A, Thorlabs) onto the active area of a single-photon sensing avalanche photodiode detector (APD, SPCM-AQR-12, EG&G Canada, Vaudreuil, Canada). In the reflected light path for side scatter detection, an OD ) 0.3 neutral density (ND) filter (NE03B, Thorlabs) was placed before the photomultiplier tube (PMT, R3788, Hamamatsu, Japan). The PMT was operated at 400 V with a CC171 Hamamatsu power supply to avoid its saturation and allow operation within its linear gain voltage region. Sample fluid was delivered pneumatically via a precise pressure regulator (8286AMBF2.5, Porter Instrument Inc., Hatfield, PA) to a fused-silica capillary (40 µm i.d., 240 µm o.d., ∼45 cm long, Polymicro Technology Inc., Phoenix, AZ) inserted into the 250 µm square bore flow channel of the flow cuvette. The inserted tip of the capillary, positioned ∼500 µm below the laser beam, was ground to an ∼12° taper (New Objective Inc., New Jersey) to facilitate smooth laminar flow of the sheath fluid around the capillary tip. Ultrapure water (Millipore, Bedford, MA) served as a sheath fluid via gravity feeding, and the flow rate can be regulated by adjusting the relative height between the sheath supply bottle and the waste container. The sheath flow rate can be easily measured by increment method. The sample volumetric flow rate needs to be measured with decrement method. Briefly, a pure water sample with a known volume was put into a 0.6 mL Eppendorf tube and weighed. Then the water sample was run at a consistent pressure (e.g., 0.66 psi for the enumeration experiment) and stopped after a predetermined time period (e.g., 30 min at 0.66 psi). The water sample was weighed again and subtracted from the initial weight to calculate the sample volumetric flow rate. The sheath and sample volumetric flow rate were controlled at ∼15 µL/min and ∼5 nL/min (0.30 psi), respectively, in the present study unless otherwise stated. The corresponding confined core diameter of the hydrodynamically focused sample stream was less than 10 µm. The probe volume was calculated to be less than 0.8 pL by the overlap of the sample stream and the laser beam. Data Acquisition and Processing. The fluorescence and light scattering signals produced from each nanoparticle passing through the illuminated observation volume were detected by the APD and PMT detectors, respectively. The APD signal was counted by a National Instruments DAQ card (PCI-6713, Austin, TX) in which the detected photon pulses in short time intervals (the preset bin width, 10-200 µs) were continuously counted by a high-speed circuit. The signal from the PMT was routed to another National Instruments DAQ card (PCI-6024E, Austin, TX) and recorded at a sampling interval corresponding to the bin width of the APD channel (100 µs in the present work, which is equivalent to 10 kHz sampling rate for the PMT signal). The two

data acquisition cards were controlled with program written in LabVIEW 8.0 (National Instruments Inc.). The data were processed by a program also written in LabVIEW 8.0, which is similar to that performed by Habbersett and Jett27 and Stoffel and co-workers.36,37 Briefly, events were counted using the peak count feature for each channel in the LabVIEW software. The criteria used for peak identification were the threshold levels in both the peak height (a digital discriminator level set to 3 times the standard deviation of the background) and the peak width. Once peaks above the threshold were identified, the appearance times for the APD channel (tFL) and for the PMT channel (tSS) were recorded. The lag times between a fluorescence burst on the APD channel and all peaks within ±100 ms from the side scatter on the PMT channel were calculated (tFL - tSS) and compiled into a lag time distribution histogram.33 Meanwhile, the burst peak height and/or area distribution histograms for both the fluorescence and light scattering were generated dynamically. Reagents and Chemicals. R-phycoerythrin (R-PE) was purchased from Prozyme Inc. (San Leandro, CA). Nonfluorescent carboxyl-polystyrene nanoparticles were purchased from Spherotech (Libertyville, IL). These particles had reported diameters of 180 ± 10, 350 ± 40, and 530 ± 70 nm and concentrations of 7.8 × 1013, 2.1 × 1012, and 6.1 × 1011 particles/mL, respectively. Fluorescent polystyrene nanospheres of 100 ± 6 and 210 ± 21 nm size dimension with excitation/emission maxima of 540/ 560 nm were obtained from Molecular Probes (Eugene, OR). The manufacturer-reported concentrations were 3.6 × 1013 and 3.9 × 1012 particles/mL for 100 and 210 nm fluorescent nanospheres, respectively. The 370 nm doxorubicin-loaded ZrO2 nanoparticles were kindly provided at a stock concentration of 3.0 × 1010 particles/mL by Dr. Zheng (College of Chemistry and Chemical Engineering, Xiamen University, China). SYTOX orange nucleic acid stain was purchased from Molecular Probes Inc. All other chemicals for buffer preparation were obtained from Sigma (St. Louis, MO). Nanoparticle Sample Preparation. The nanospheres were diluted to concentrations around 108 particles/mL with ultrapure water containing 0.05% of Tween 20 to prevent aggregation. The water had been filtered through 20 nm Anodisc filters (Whatman, Clifton, NJ). Before dilution all sphere stock solutions were sonicated for approximately 2 min. All diluted sphere solutions were stored at 4 °C, and sonication was performed immediately before analysis. Escherichia coli Staining Using SYTOX Orange Nucleic Acid Stain. E. coli ER2738 cells were grown overnight in Luria-Bertani (LB) broth (10 g of tryptone, 5 g of yeast extract, and 10 g of NaCl per liter) at 37 °C in baffled flasks with rotary aeration for about 16 h until stationary phase was reached. The concentration of the stock solution was measured to be 2 × 109 cfu/mL using plate counting. The harvested E. coli ER2738 cells were then heat-killed at 75° for 3 min. Approximately 1 × 108 cfu/mL E. coli ER2738 cells were washed three times in 20 nm filtered normal saline (0.9% sodium chloride) by centrifugation and then stained with 50 nM of SYTOX orange nucleic acid stain. SYTOX orange can stain nucleic acids within dead cells or those having compromised membranes. (36) Stoffel, C. L.; Kathy, R. F.; Rowlen, K. L. Cytometry, Part A 2005, 65A, 140–147. (37) Stoffel, C. L.; Rowlen, K. L. Anal. Chem. 2005, 77, 2243–2246.

Analytical Chemistry, Vol. 81, No. 7, April 1, 2009

2557

Figure 2. Single-molecule fluorescence burst data from (a) a 20 nm filtered ultrapure water blank and (b) R-PE molecules in an 83 fM solution. The data were binned into 100 µs intervals, and the excitation laser power was 0.75 mW.

RESULTS AND DISCUSSION Before attempting the simultaneous measurements of fluorescence and light scattering from nanoparticles, the detection capabilities on both detection channels of the laboratory-built HSDCFCM system were assessed by analyzing R-PE solutions and mixture of polystyrene nanoparticles, respectively. Performance Evaluation of the HSDCFCM for Fluorescence Detection. R-PE is a highly soluble fluorescent protein. Due to its large extinction coefficients at multiple absorption bands (e.g., 1.96 × 106 M-1 cm-1 at 565 nm, 1.65 × 106 M-1 cm-1 at 546 nm, 1.35 × 106 M-1 cm-1 at 498 nm) and remarkable quantum yield of 0.82, R-PE is 10-20 times brighter than

conventional organic fluorophores. R-PE has been widely used in flow cytometry, live cell staining, and multicolor immunofluorescent staining.38 The reported limit of detection is typically ∼150 PE equiv for the commercial flow cytometers. Figure 2 displays 600 ms of raw fluorescence burst data from ultrapure water and single R-PE molecules obtained on the HSDCFCM. Clearly, the fluorescence bursts of individual R-PE molecules were well above the background. For an R-PE solution at concentration of 83 fM, 13 fluorescence bursts were observed within a 600 ms time interval. According to the Poisson distribution, the probability of two R-PE molecules simultaneously residing in the detection volume was calculated to be less than 0.12%. On the basis of the mean of the R-PE fluorescence burst height and the standard deviation of the background signal, the signal-tonoise ratio (S/N) for this data set was calculated to be 17, which is about 3 orders of magnitude improvement as compared to the conventional flow cytometer. However, the fluorescence of individual rhodamine-6G molecules could not be resolved from the background (data not shown) due to the relatively large probe volume as compared to that of a confocal setup. Performance Evaluation of the HSDCFCM for Light Scattering Detection. Polymer nanoparticles have long been used as instrument calibration standards. Figure 3 displays the side scatter signals measured with a mixture of nonfluorescent polystyrene nanoparticles of three different sizes: 180, 350, and 530 nm. Blank solution of 20 nm filtered ultrapure water was analyzed to evaluate the false-positive signal caused by particulate impurities residing in the solution. The data in Figure 3a indicates that the scattering bursts contributed from the impurities were trivial and thus can be ignored. On the other hand, when the

Figure 3. Side scatter detection of a mixture of three sizes of polystyrene nanoparticles; the concentrations for 180, 350, and 530 nm nanoparticles were 2.6 × 108, 0.7 × 108, and 2.0 × 108 particles/mL, respectively: (a) trace of side scatter signal from a blank of 20 nm filtered water; (b) trace of side scatter signal from the nanoparticle mixture; (c) side scatter burst area distribution histogram derived from data set collected in a 60 s interval; (d) side scatter burst height distribution histogram derived from the same data set as in panel c. 2558

Analytical Chemistry, Vol. 81, No. 7, April 1, 2009

Figure 4. Detection of 100 nm fluorescent polystyrene nanospheres at a concentration of 3.6 × 108 particles/mL: (a) fluorescence burst trace from the APD channel; (b) fluorescence burst size distribution histogram derived from 60 s of data from the APD; (c) trace of the side scatter signal from the PMT channel; (d) side scatter burst area distribution histogram derived from the corresponding 60 s of data from the PMT; (e) histogram of lag times between peak bursts from the APD and PMT channels; (f) scanning electron microscope image of 100 nm fluorescence nanospheres. In panels a and c of the 100 ms time interval, fluorescence burst and side scatter peaks corresponding to the same particle are labeled by the same number. The fluorescence signal was attenuated by a factor of 10 using an OD ) 1.0 ND filter.

nanoparticle mixture was analyzed, intense side scatter bursts showed up with peak heights mainly residing around three different levels, namely, ∼1200, ∼3400, and ∼8000 mV for 180, 350, and 530 nm nanoparticles, respectively (Figure 3b). In Figure 3, parts c and d, three distinct peaks with good resolution can be observed in both the side scatter burst area and burst height distribution histograms. The side scatter signals generated from 180 nm nanoparticles can be easily distinguished from the background. Slightly better resolution was achieved in the burst height distribution histogram (Figure 3d) as compared to the burst area distribution histogram (Figure 3c). Performance Assessment for Two-Channel Detection. Commercially available fluorescent polystyrene nanospheres were utilized to assess the performance of the two-channel system. The fluorescent nanospheres, 100 and 210 nm in diameter, were diluted to the concentrations of 3.6 × 108 and 3.9 × 108 particles/mL, respectively, and analyzed on the HSDCFCM. Because the (38) http://probes.invitrogen.com/media/pis/mp00800.pdf.

fluorescence emitted from both sizes of fluorescent nanospheres was too intense, an OD ) 1.0 ND filter was inserted in the optical path of fluorescence detection channel to avoid APD saturation. A representative data set for a sample of 100 nm fluorescent nanospheres is given in Figure 4. Panels a and c of Figure 4 show fluorescence burst data from the APD channel and corresponding side scatter signal from the PMT channel, respectively. It is apparent that the fluorescence burst and the side scatter signal generated from the same particle correlated quite well in the time frame. The peak marked by X in Figure 4c could be attributed to the impurity particle in water. Panels b and d of Figure 4 are the peak area distribution histograms for fluorescence burst and side scatter burst, respectively. The measured coefficient of variation (CV) for a defined population was 8.5% and 15% for the fluorescence and side scatter, respectively. Cross-correlation analysis yields a large peak at zero lag time (Figure 4e). To evaluate the instrument reproducibility of the HSDCFCM, the 210 nm fluorescent nanospheres were analyzed Analytical Chemistry, Vol. 81, No. 7, April 1, 2009

2559

Figure 5. Detection of the 370 nm doxorubicin-loaded ZrO2 nanoparticles at a concentration of 3.0 × 108 particles/mL: (a) fluorescence burst data from the APD; (b) side scatter signal from the PMT; (c) scanning electron microscope image of the ZrO2 nanoparticles; (d) histogram of lag times between peaks from the APD and PMT channels; (e) dot-plot of fluorescence burst area vs side scatter burst area.

repeatedly for 15 times at the same experimental conditions. The relative standard deviations (RSDs) were 0.80% and 0.63% for the means of fluorescence burst area and side scatter burst area, respectively. The average of the fluorescence histogram CV was 4.90% with an RSD of 1.60%, and the average of side scatter histogram CV was 8.49% with an RSD of 1.20%. These results demonstrated that HSDCFCM instrument can provide stable analysis in both fluorescence and side scatter detection channels. Application in Drug-Loaded ZrO2 Nanoparticle Analysis. Doxorubicin is a widely used and effective chemotherapy agent for cancer treatment. However, it exhibits severe dose-limiting toxicity to heart muscle, and therefore technology development for improved drug delivery and release is urgently needed. Nanoparticles, nanocapsules, and micellar systems have been applied to enhance drug efficiency through target-directed drug delivery.39,40 The fluorescence characteristics of doxorubicin have been often used to monitor localization of the drug within lipid bilayers and liposomal delivery systems41 or to observe the efficiency of drug release from polymeric nanoparticles.42,43 Doxorubicin is fluorescent, and its excitation and emission spectra (39) Packhaeuser, C. B.; Schnieders, J.; Oster, C. G.; Kissel, T. Eur. J. Pharm. Biopharm. 2004, 58, 445–455. (40) Jayanth, P.; Vinod, L. Adv. Drug Delivery Rev. 2003, 55, 329–347. (41) Karukstis, K. K.; Thompson, E. H. Z.; Whiles, J. A.; Rosenfeld, R. J. Biophys. Chem. 1998, 73, 249–263.

2560

Analytical Chemistry, Vol. 81, No. 7, April 1, 2009

match with our instrument setup. The application of the HSDCFCM in the analysis of doxorubicin-loaded nanoparticles is given in Figure 5. As displayed in Figure 5, relatively large variation in signal intensity can be observed on both the fluorescence channel (Figure 5a) and side scatter channel (Figure 5b) for the 370 nm doxorubicin-loaded ZrO2 nanoparticles. Supported by the SEM image (Figure 5c), the signal variation could be due to the intrinsic size variation, fragmentation, and the aggregation of ZrO2 nanoparticles. However, a good cross-correlation between the two signals was identified in the lag time distribution histogram (Figure 5d). And the dot-plot of fluorescence burst area versus side scatter burst area (Figure 5e) indicates a relatively linear correlation between the intensities of side scatter and fluorescence. Present work demonstrated the potential application of the HSDCFCM in the characterization of drug-loaded nanoparticles, such as to evaluate the drug’s content distribution loaded into each nanoparticle, to assess the size distribution of nanoparticles, and to examine the correlation between the two properties. Besides, the HSDCFCM could be especially useful in monitoring the dynamic drug release process during which the fluorescence (42) Missirlisa, D.; Kawamurab, R.; Tirellic, N.; Hubbell, J. A. Eur. J. Pharm. Sci. 2006, 29, 120–129. (43) Husseini, G. A.; Rosa, M. A. D.; Gabuji, T.; Zeng, Y.; Christensen, D. A.; Pitt, W. G. J. Nanosci. Nanotechnol. 2007, 7, 1028–1033.

Figure 6. Detection of the SYTOX orange dye labeled E. coli ER2738 cells at a concentration of 1.0 × 108 cfu/mL. (a1, b1, and c1) Fluorescence burst data from APD for a 50 nM SYTOX orange diluted in saline solution, E. coli ER2738 cells without SYTOX orange staining, and stained E. coli ER2738 cells, respectively. (a2, b2, and c2) Side scatter signal from the PMT for the above three samples, respectively. (a3, b3, and c3) Histograms of lag time between peaks from the APD and PMT channels for the above three samples, respectively. The fluorescence signal was attenuated by a factor of 10 using an OD ) 1.0 ND filter.

associated with each individual nanoparticle shall decrease while the background fluorescence from the solution is expected to increase. Application in Bacteria Detection. The detection and enumeration of nanometric biological particles such as bacteria and virus is vital for food safety, clinical diagnosis and therapies, and environmental monitoring. Currently, enumeration of bacteria or virus is most commonly accomplished by a plaque assay technique that was tedious and time-consuming to conduct.44 In this context, a technique enabling rapid bacteria or virus enumeration will have a vast number of potential medical applications. Here we used bacteria E. coli ER2738 as an example to demonstrate the ability of the HSDCFCM in single nanometric biological particle detection. Because the fluorescence signal from each stained bacterium was too intense, an OD ) 1.0 ND filter was placed in the fluorescence detection optical path to protect the detector and avoid APD saturation. Figure 6 shows the results of the nucleic acid dye SYTOX orange stained E. coli ER2738 cells at a concentration of 1.0 × 108 cfu/mL and the results of two control samples. SYTOX orange is an intercalating dye with excitation/emission maxima of 547/570 nm and 500-fold fluorescence enhancement upon nucleic acid binding, and there is no background fluorescence in the unbound state.45,46 As we can see from Figure 6, the background signal from SYTOX orange stain diluted in saline solution was very low in both the (44) Samoylova, T. I.; Smith, B. F. BioTechniques 1999, 27, 356–361.

APD and PMT channels (Figure 6, panels a1 and a2). For the unstained E. coli ER2738 cells, only scattering signals generated from individual bacteria can be observed in the PMT channel as expected (Figure 6b2), whereas no fluorescence burst was observed in the APD channel (Figure 6b1). For the SYTOX orange dye stained E. coli ER2738 cells, fluorescence and side scatter burst signals can be observed on the APD and PMT channels (Figure 6, panels c1 and c2), respectively, and crosscorrelation analysis yields a large peak at zero lag time (Figure 6c3). The height of the spikes was not uniform on both the fluorescence (APD) and the light scattering (PMT) channels, which might be caused partly by the disparity in nucleic acid content and in cell size (0.7-1.5 µm in length under the microscope), as each individual bacteria may reside in a different phase point of the cell cycle. On the other hand, the varied paths and the rotation of the rod-shaped bacterial cells as they transit the probe region could be another source of the signal variation on both the detection channels. Nanoparticle Enumeration. For the enumeration of nanoparticles, the sample introduction pressure was maintained at 0.66 psi and the measured sample volumetric flow rate was 47 nL/ min (n ) 3). Fluorescent nanoparticles of 210 nm diameter were analyzed with sample concentrations ranging from 1.62 × 105 to (45) Yan, X.; Habbersett, R. C.; Cordek, J. M.; Nolan, J. P.; Yoshida, T. M.; Jett, J. H.; Marrone, B. L. Anal. Biochem. 2000, 286, 138–148. (46) http://probes.invitrogen.com/media/pis/mp11368.pdf.

Analytical Chemistry, Vol. 81, No. 7, April 1, 2009

2561

Table 1. Enumeration of 210 nm Fluorescent Nanoparticles on the HSDCFCM MRC (particles/mL)a

event rate (particles/min)b

FLC (particles/mL)b

FLC/MRC (%)

event rate (particles/min)c

DCC (particles/mL)c

DCC/MRC (%)

1.60 × 105 4.86 × 105 1.46 × 106 4.37 × 106 1.31 × 107 3.93 × 107

6 ± 0.9 21 ± 2.5 57 ± 8.6 167 ± 9.7 590 ± 55 1839 ± 79

(1.39 ± 0.4) × 105 (4.59 ± 0.5) × 105 (1.24 ± 0.2) × 106 (3.63 ± 0.2) × 106 (1.28 ± 0.1) × 107 (4.00 ± 0.2) × 107 av ± SD

86 94 85 83 98 102 91 ± 8

5 ± 1.3 19 ± 2.6 54 ± 6.8 161 ± 18 573 ± 43 1811 ± 95

(1.14 ± 0.3) × 105 (4.13 ± 0.6) × 105 (1.18 ± 0.1) × 106 (3.49 ± 0.4) × 106 (1.25 ± 0.1) × 107 (3.94 ± 0.2) × 107 av ± SD

70 85 81 80 95 100 85 ± 11

a Manufacturer’s reported concentration (MRC) based on dry weight analysis. b Enumeration by fluorescence channel (FLC). c Enumeration by dual-channel coincidence (DCC).

Table 2. Particles Mixture Data Obtained with 530 nm (Nonfluorescent) and 210 nm (Fluorescent) Particles on the HSDCFCM concentration mixture ratio (530 nm/210 nm)

530 nm/210 nm (experiments)a

RSD (n ) 8, %)

1/1 9/1 20/1

0.98/1 9.01/1 20.34/1

4.8 2.2 4.1

a

Figure 7. Correlation between the manufacturer’s reported concentrations and those measured by HSDCFCM enumeration of 210 nm fluorescent latex beads.

3.93 × 107 particles/mL. With data acquisition time of 60 s, multiple data sets (n ) 8) were consecutively collected for each sample. Table 1 summarizes the detection efficiencies measured at each nanosphere concentration (assuming the manufacturer’s reported concentration is accurate). Two approaches, fluorescence and dual-channel coincidence measurements, were employed for the nanoparticle counting. As we can see in Table 1, for the 210 nm fluorescent latex beads, the detection efficiencies varied from 83% to 102% for the fluorescence, and from 70% to 100% for the dual-channel coincidence approaches, respectively. In general, higher particle concentration yields relatively higher detection efficiency. Particle adsorption onto the capillary surface becomes significant at lower particle concentrations which results in relatively low detection efficiency. When the particle concentration is above 3 × 107 particles/mL, nearly 100% detection efficiency can be obtained within both approaches. Figure 7 shows that excellent correlations were obtained between the manufacturer’s reported concentrations and those measured by HSDCFCM enumeration. For both fluorescence and dual-channel coincidence approaches, the R2 was 0.9982 and 0.9992, respectively. If nanoparticles of concentration less than 1 × 105 particles/mL need to be enumerated, increased sample delivery pressure, prolonged acquisition time, or a combination of both approaches needs to be applied. However, it is worth noting that the increase of the sample pressure will yield a corresponding higher interrogating rate of the particulate impurities residing in the sample solution due to the increased sample volumetric flow rate. As the particulate impurities are usually nonfluores2562

Analytical Chemistry, Vol. 81, No. 7, April 1, 2009

Average 530 nm/210 nm ratio collected from eight 60 s runs.

cent, the augmented impurity counting will be reflected primarily in the light scattering channel. Therefore, fluorescence or dual-channel coincidence measurement will be advantageous in resolving the fluorescent nanoparticles for accurate enumeration. For fluorescent nanoparticles, reliable enumeration can be attained for a concentration as low as 103 particles/mL when higher sample pressure and 2 min of acquisition time were applied. For nonfluorescent nanoparticles with a light scattering signal well resolved from the particulate impurities, reliable counting can be achieved at a concentration of ∼104 particles/mL with 10 min or longer acquisition time. In order to evaluate the feasibility of the HSDCFCM in microsphere mixture enumeration, 530 nm nonfluorescent nanoparticles and 210 nm fluorescent nanoparticles were mixed in 1/1, 9/1, and 20/1 ratios and analyzed. The total particle concentration for each ratio was maintained at ∼6.0 × 107 particles/mL. As Table 2 indicates, it is feasible to distinguish between fluorescently labeled and unlabeled nanoparticles of different sizes and varying proportions based upon the measured side scattering and fluorescence properties. In conclusion, simultaneous detection and cross-correlation of the side scattering and fluorescence signals of individual nanoparticles has been achieved on a laboratory-built high-sensitivity dual-channel flow cytometer. Information on size, concentration, and biochemical properties of nanosized particles can be obtained concurrently. Clearly, the development of HSDCFCM extends the application of HSFCM instrumentation from DNA fragment sizing27,31 into nanoparticle analysis. It is envisioned that the HSDCFCM will find many important applications in the new frontiers of biomedicine involving nanoparticles. For example, the HSDCFCM may provide new scientific insights into the submicrometer world of biological entities through the individual analysis of bacteria, viruses, and subcellular organelles. These new discoveries could lead to breakthroughs in medical diagnosis and

treatment. The HSDCFCM can be used in the rapid characterization of newly synthesized nanoparticles and in studying nanoparticle mediated drug delivery. ACKNOWLEDGMENT This work was supported by the National Natural Science Foundation of China (Nos. 20645001 and 20675070), the Department of Science and Technology of Fujian Province (No. 2005NZ1013), the Program for New Century Excellent Talents in University (NCET), the Program for New Century Excellent Talents in Fujian Province University (NCETFJ), and by the

Scientific Research Foundation for the Returned Overseas Chinese Scholars, State Education Ministry (SRF for ROCS, SEM) for which we are most grateful. The authors gratefully acknowledge Dr. Nanfeng Zheng of the Xiamen University for providing doxorubicin-loaded ZrO2 nanoparticles.

Received for review November 20, 2008. Accepted February 13, 2009. AC802464A

Analytical Chemistry, Vol. 81, No. 7, April 1, 2009

2563