Concentric Förster Resonance Energy Transfer Imaging - Analytical

Technical Note pubs.acs.org/ac

Concentric Förster Resonance Energy Transfer Imaging Miao Wu and W. Russ Algar* Department of Chemistry, University of British Columbia, 2036 Main Mall, Vancouver, British Columbia V6T 1Z1, Canada S Supporting Information *

ABSTRACT: Concentric Förster resonance energy transfer (cFRET) configurations based on semiconductor quantum dots (QDs) are promising probes for biological sensing because they offer multiplexing capability in a single vector with robust ratiometric detection by exploiting a network of FRET pathways. To expand the scope and utility of cFRET probes, it is necessary to develop and validate cFRET imaging methodology. In this technical note, we present such a methodology using a protease-sensitive cFRET configuration that comprises a green-emitting QD, Alexa Fluor 555 (A555), and Alexa Fluor 647 (A647). Photoluminescence (PL) images were acquired with three filter-based emission channels to permit measurement of A555/QD and A647/QD PL ratios. With reference to calibration samples, these PL ratios were used to calculate quantitative progress curves for proteolytic activity in regions of interest in the acquired images. Importantly, the imaging methodology reproduces quantitative results obtained with a monochromator-based fluorescence plate reader. Spatiotemporal resolution is demonstrated by tracking the activity of two prototypical proteases, trypsin and chymotrypsin, as they diffuse down the length of a capillary. This methodology is expected to enable the future use of cFRET probes for cellular sensing and other imaging assays.

T

probes are advantageous in that they harness both the optical properties of the QD (e.g., brightness, broad absorption, narrow emission) and its surface area to provide multifunctionality in a single vector. In contrast to the use of two conventional QD-based FRET probes (each comprising a single donor−acceptor pair) for multiplexed sensing, cFRET probes have the capacity to support multiplexing at the ensemble level down to the single-particle level, avoid potential challenges associated with different levels of brightness between probes, and have less spectral crosstalk between emission signals. These features make cFRET probes very promising for cellular sensing applications, where additional advantages are anticipated to include the introduction of less foreign material to the biological system, and negation of differences in cellular uptake, distribution, stability, or other properties that may occur between two separate probes. However, before these prospective advantages can be realized, an important limitation must be addressed: cFRET probes have, to date, only been demonstrated in homogeneous formats with monochromatorbased fluorescence measurements. These formats are not amenable to imaging, which is a requirement for future cellular sensing. Here, we take a critical first step toward realizing image-based sensing with cFRET probes by demonstrating and validating cFRET imaging via fluorescence microscopy. The model

here is widespread interest in the use of photoluminescent nanoparticles such as metal nanoclusters,1 semiconductor quantum dots (QDs),2 semiconducting polymer nanoparticles (Pdots),3 carbon nanomaterials,4,5 and upconverting lanthanide nanoparticles,6 among others, as probes for bioanalysis. Several innovative bioanalysis strategies have been developed with both new and enhanced capabilities compared to strategies with only molecular fluorophores. In this context, we have been developing the concept of concentric Förster resonance energy transfer (cFRET).7−11 In cFRET, a central QD is conjugated with multiple copies of two different biomolecular probes. Each of these probes is directly or indirectly labeled with one of two different fluorescent dyes or other optically active materials that engage in energy transfer with the QD and, in many cases, with one another. These configurations are referred to as “concentric” because the dyes (etc.) are, to a first approximation, located on the surface of a sphere that is slightly larger and concentric with the QD. This nomenclature distinguishes these configurations from other nanostructures with linear arrays or other geometries of fluorophores. One example of a cFRET system has paired a green-emitting QD with a yellow-emitting dye (e.g., Cy3, Alexa Fluor 555) and a deep red-emitting dye (e.g., Alexa Fluor 647).7,11 Another system has paired a red-emitting QD with a luminescent terbium(III) complex and a deep red-emitting dye.9,10 These systems have been used for the multiplexed detection of target nucleic acids,9 the multiplexed detection of protease activity and pro-protease activation,10,11 and, most recently, the parallel detection of protease activity and concentration.7 These cFRET © XXXX American Chemical Society

Received: May 25, 2015 Accepted: July 14, 2015

A

DOI: 10.1021/acs.analchem.5b01946 Anal. Chem. XXXX, XXX, XXX−XXX

Technical Note

Analytical Chemistry

published protocol. 17 These peptides are denoted as SubChT(A555) and SubTRP(A647), with the subscript identifying the target enzyme (ChT, chymotrypsin; TRP, trypsin). Fluorescence Measurements. Fluorescence images were acquired with a 4.0× magnification (NA = 0.16) objective lens using an IX83 inverted epifluorescence microscope (Olympus, Richmond Hill, ON, Canada) equipped with MetaMorph/ MetaFluor software (Molecular Devices, Sunnyvale, CA), an XCite 120XL metal-halide light source (Excelitas Technologies, Mississauga, ON, Canada), an Orca-Flash 4.0 V2 sCMOS camera (C11440; Hamamatsu Photonics, Hamamatsu, SZK, Japan), and a motorized emission filter wheel (Olympus). The filter wheel housed 520/40 (center line/bandwidth in nm), 565/30, and 665 long-pass filters (Chroma Technology Corp, Bellow Falls, VT) for sequential measurement of QD, A555, and A647 emission, respectively. The excitation filter was always 405/20 and paired with a 425 nm cutoff dichroic mirror (Chroma). Additional fluorescence measurements were performed with an Infinite M1000 Pro multifunction plate reader (Tecan Ltd., Morrisville, NC). This instrument is equipped with a xenon flash lamp, excitation and emission monochromators, and a photomultiplier tube detector. Emission was measured at 520, 570, and 670 nm with excitation at 400 nm (5 nm monochromator bandwidth in all cases). Data and Image Analysis. Epifluoresence images were acquired in the QD, A555, and A647 filter channels, and A555/ QD and A647/QD PL ratios were calculated as described in detail in the Supporting Information. For plate reader measurements, the PL emission intensity was measured at 520, 570, and 670 nm and used to calculate analogous PL ratios. Both of these PL ratios were empirical in that the values depended on the instrument and its acquisition settings. When necessary, correction factors can be employed to calculate true physical PL ratios;11 however, this is not necessary for most applications. Here, an experimental calibration (vide inf ra) was used to correlate A555/QD and A647/QD PL ratios against the average number of A555 and A647 per QD, M and N, and thus the number of each peptide sequence per QD. Progress curves were plotted as the change in the number of SubChT(A555) and SubTRP(A647) per QD, ΔM and ΔN, as a function of time.

cFRET system (Figure 1A) comprises green-emitting CdSe/ CdS/ZnS QDs and peptides labeled with Alexa Fluor 555

Figure 1. (A) Schematic of the QD-based cFRET configuration: A555 and A647 dyes are assembled to the QD through peptide substrates for chymotrypsin, 1 = SubChT(A555), and trypsin, 2 = SubTRP(A647). The dashed lines indicate recognition sites where the proteases hydrolyze the peptides. The weights of the arrows qualitatively indicate the relative efficiency of each FRET pathway. (B) Emission spectra for the QD, A555, A647, and a fully assembled cFRET configuration, [SubChT(A555)]7-QD-[SubTRP(A647)]7. Shaded areas represent the transmission region of the optical filters used for imaging. (C) Schematic of the microscope configuration used for cFRET imaging (EX, excitation filter; DM, dichroic mirror; OBJ, objective lens; EM, emission filter; FW, filter wheel).

(A555) and Alexa Fluor 647 (A647). The two peptide sequences are selected as substrates for two prototypical proteases, trypsin and chymotrypsin, which are a reliable test system.10,11 As described previously, hydrolysis of the peptides by the proteases modulates cFRET by decreasing the number of A555 and A647 per QD.11 We now show that three-color fluorescence microscopy (Figure 1B,C) can be used for measurement of cFRET. This imaging methodology can reproduce results obtained with a monochromator-based fluorescence plate reader and permit quantitative analysis of spatially and temporally heterogeneous proteolytic activity. As such, it is an essential foundation upon which future imagebased sensing applications of cFRET will build, with prospective application in cellular analysis.

■

RESULTS AND DISCUSSION Calibration. The cFRET probes depicted in Figure 1A provide quantitative data through measurement of the A555/ QD and A647/QD PL ratios. In an imaging format, these ratios are calculated from three PL images acquired with three filters that select for QD, A555, and A647 PL, as shown in Figure 1B. Calibration of these PL ratios permits determination of the average numbers of A555 and A647 dyes per QD, abbreviated by M and N. These dyes were linked to peptide substrates for proteases, thus permitting quantitative assays of proteolytic activity. Figure 2A shows images of 25 calibration samples acquired using the QD, A555, and A647 filter channels. The image-derived PL ratios are plotted in Figure 2B. As expected, QD PL decreased with increasing amounts of either A555 or A647 per QD, A555 PL increased with a greater number of that dye per QD and decreased with more A647 per QD, and A647 PL increased with increasing numbers of both A555 and A647 per QD. These trends reflected the QD-to-A555, QD-to-A647, and A555-to-A647 FRET pathways illustrated in Figure 1A. The values of the PL ratios depended on filter selection and varied predictably with changes in image acquisition parameters

■

EXPERIMENTAL SECTION Detailed experimental methods are available in the Supporting Information. Reagents and QD−Peptide Conjugates. Bovine trypsin and chymotrypsin were from Sigma-Aldrich (Oakville, ON, Canada). Borate buffered saline (BBS; pH 8.5, 50 mM, 50 mM NaCl) was prepared in-house and sterilized by autoclaving. CdSe/CdS/ZnS QDs were synthesized using established methods12,13 and coated with hydrophilic glutathione (GSH) ligands and self-assembled with dye-labeled, polyhistidineterminated peptides (Bio-Synthesis, Inc., Lewisville, TX; see Table S1 in the Supporting Information) as described previously.14−16 The peptides were labeled at their distal terminus with A555 or A647 (Invitrogen, Carlsbad, CA) using a B

DOI: 10.1021/acs.analchem.5b01946 Anal. Chem. XXXX, XXX, XXX−XXX

Technical Note

Analytical Chemistry

Figure 2. (A) PL images acquired in the QD, A555, and A647 filter channels for calibration samples in microtiter plate wells. Each calibration sample had a different (M, N) combination of (A555, A647) per QD as [SubChT(A555)]M-QD-[SubTRP(A647)]N cFRET conjugates (including 7 − M and 7 − N equivalents of predigested peptide substrates). (B) Corresponding calibration plots of A555/QD and A647/QD PL ratios from imaging.

such as excitation intensity (i.e., lamp intensity) and camera integration time (see the Supporting Information for details). Although we recommend that these parameters be kept constant between calibration experiments and assays, correction factors can potentially be utilized to accommodate changes in settings. Homogeneous Assays. The first tests of the microscopebased cFRET imaging methodology were homogeneous assays in 96-well microtiter plates, which were compared to replicate assays measured with established fluorescence plate reader methodology.7,11 The calibration data in Figure 2 was used to convert PL ratios into progress curves for proteolysis of the cFRET probes (i.e., the change in the average numbers of each dye-labeled peptide substrate per QD, ΔM and ΔN at each measurement time point). Examples of assays with only trypsin and only chymotrypsin are shown in the Supporting Information (Figure S5). Figure 3A shows (i) PL ratio data and (ii) progress curves for imaging assays of proteolytic activity for different mixtures of both trypsin and chymotrypsin. For each assay, there was good agreement between progress curves derived from plate reader and microscope imaging measurements. Importantly, there was a near-unity correlation between the initial proteolytic rates in Figure 3B, thus validating the imaging methodology. Figure 3B includes data from Figure 3A and Figures S5 and S6 in the Supporting Information. Imaging Assays. The value of imaging is not in assaying homogeneous samples but rather in resolving spatially heterogeneous distributions of biological activity. Here, we tested this capability by filling a glass capillary tube with cFRET probes and introducing protease at one or both ends of the tube. Figure 4 shows representative data from introducing trypsin at one end of the capillary. As the trypsin diffused through the capillary, the A555 PL intensity increased while that of A647 decreased, resulting in increasing A555/QD and

Figure 3. (A) PL ratio data (i) and progress curves (ii) showing the hydrolysis of cFRET conjugates by different mixtures of trypsin and chymotrypsin. (B) Correlation between the rates derived from microscope imaging measurements and fluorescence plate reader measurements. This plot includes data from panel A and Figures S5 and S6 in the Supporting Information.

decreasing A647/QD PL ratios, consistent with loss of the QDto-A647 and A555-to-A647 FRET pathways. The per-pixel PL intensity values in regions of interest (ROIs) were averaged and used to calculate local PL ratios and progress curves from a time-series of PL images. Progress curves calculated for different ROIs along the length of the capillary showed that the onset of trypsin proteolysis correlated with the distance of the ROI from the end of the capillary where trypsin was introduced. As expected, only SubTRP(A647) was hydrolyzed by trypsin, yielding ΔN < 0 and ΔM ≈ 0. The small decrease in ΔN for ROI4 (see Figure 4C) can be attributed to photobleaching of the A647. The very small decrease in ΔM may also have a contribution from slight photobleaching of the A555 or possibly from very low levels of a nonspecific effect associated with trypsin. Both of these trends were easily distinguished from the target proteolytic activity. An analogous imaging assay was done with chymotrypsin (see Figure S7 in the Supporting Information), where diffusion of chymotrypsin along the length of capillary was evident between ROIs. In this case, the QD PL intensity increased while both the A555/QD and A647/QD PL ratios decreased. Only SubChT(A555) was hydrolyzed, with ΔN ≈ 0 and ΔM < 0. Limited photobleaching of the A647 was again observed but easily distinguished from proteolytic activity. Figure 5 shows representative data from adding both trypsin and chymotrypsin at one end of the capillary. The QD PL intensity increased dramatically as these enzymes diffused from the end of the capillary, and both the A555/QD and A647/QD C

DOI: 10.1021/acs.analchem.5b01946 Anal. Chem. XXXX, XXX, XXX−XXX

Technical Note

Analytical Chemistry

Figure 4. Spatiotemporally resolved trypsin activity. (A) Experimental design: one end of a glass capillary filled with a solution of cFRET probes is exposed to a trypsin solution, which diffuses along the length of the capillary. (B) False-color images of QD, A555, and A647 PL intensity within the capillary, as well as raw A555/QD and A647/QD PL ratio images (not corrected for crosstalk and background), at various time points after adding trypsin. The scale bar is 400 μm. (C) Progress curves for hydrolysis of SubChT(A555), ΔM, and SubTRP(A647), ΔN, for different regions of interest (ROIs).

Figure 5. Spatiotemporally resolved trypsin and chymotrypsin activity. (A) Experimental design: one end of a glass capillary filled with a solution of cFRET probes is exposed to a mixed trypsin and chymotrypsin solution, which diffuses along the length of the capillary. (B) False-color images of QD, A555, and A647 PL intensity within the capillary, as well as raw A555/QD and A647/QD PL ratio images (not corrected for crosstalk and background), at various time points after adding protease. The scale bar is 400 μm. (C) Progress curves for hydrolysis of SubChT(A555), ΔM, and SubTRP(A647), ΔN, for different regions of interest (ROIs).

PL ratios decreased. The apparent A555 intensity did not appear to decrease because of crosstalk with the QD PL. Progress curves calculated from different ROIs again revealed the diffusion of these enzymes along the length of the capillary, with hydrolysis of both SubTRP(A647) and SubChT(A555), and ΔN < 0 and ΔM < 0. Limited photobleaching of A647 was again observed in the absence of proteolytic activity (see ROI5). Results from another experiment, with trypsin and chymotrypsin introduced sequentially from opposite ends of

the capillary, are shown in the Supporting Information (Figure S8). As trypsin diffused from one end, the A555/QD PL ratio increased while the A647/QD PL ratio decreased. At the opposite end, both the A555/QD and A647/QD PL ratios decreased. These results, as well as progress curves for ROIs, corresponded to the expected diffusion and activity of trypsin and chymotrypsin from opposite ends of the capillary. Cumulatively, the results in Figures 4 and 5 and Figures S7 and S8 in the Supporting Information demonstrate quantitative D

DOI: 10.1021/acs.analchem.5b01946 Anal. Chem. XXXX, XXX, XXX−XXX

Technical Note

Analytical Chemistry

only possible with a monochromator-based fluorescence plate reader. The imaging methodology makes quantitative analysis of multiplexed, heterogeneous biological activity possible with cFRET. This capability was demonstrated by imaging the proteolytic activity of trypsin and chymotrypsin as these enzymes diffused through a capillary, including extraction of quantitative progress curves from regions of interest. These capabilities are the principal requirements for cellular, microfluidic, and other imaging assays and will permit the future use of cFRET probes for studying biochemical processes with spatiotemporal resolution and quantitative measurement of activity.

spatial resolution of proteolytic activity with good selectivity through QD-based cFRET imaging. Discussion. The main technical requirement of this method is an automated emission filter wheel or other device for rapid switching between three emission filter channels. Alternatively, beam splitter-based quad-view technology18 could be used, although the older and more widespread dual-view technology has too few channels. The quad-view technology would acquire all three PL images simultaneously, potentiating better time resolution than a filter wheel (which acquires images sequentially), but at the expense of a smaller field of view. Many filter wheels have switching times on the order of 100 ms or less, such that time resolution on the order of seconds remains possible. Many emission filter wheel systems are commercially available for upgrading microscope systems that do not have this capability. Care should be taken to select emission filters that minimize crosstalk between QD, A555, and A647 (or other dye) PL and efficiently block excitation light. In our experience, many red emission filters transmit violet light because they have been designed with the assumption of greenorange excitation due to the small Stokes shifts of typical red fluorescent dyes. Notably, with the use of a QD as an initial and central donor, only a single excitation filter and dichroic mirror are required. As noted in the introductory information and highlighted in a recent review,19 an important advantage of cFRET probes is that they offer multiplexed detection in a single vector, where that detection is ratiometric and thus much less sensitive to dilution and instrumental variations and fluctuations than measurements based on absolute PL intensities. In the context of prospective use as a cellular imaging probe, these features are very attractive but required the development of an imaging methodology to obtain spatial resolution, as done here in this technical note. Importantly, we have shown that dynamic and quantitative cFRET imaging is possible. While we derived progress curves for particular regions of interest in PL images, the same analysis could be done on a pixel-by-pixel basis for high-resolution mapping of protease activity. Adaption of cFRET measurements from monochromator-based instruments to a filter-based microscope also offered a technical advantage: the camera sensitivity (e.g., gain, binning, integration time) was readily adjustable between detection channels to optimize signal magnitudes. In our experience, such adjustments have not been practical with monochromator-based spectrofluorimeters and fluorescence plate readers. Porting the cFRET methodology to a filter-based imaging format also confirms that cFRET is compatible with filter-based fluorescence plate readers, which are equally if not more common than their monochromator-based counterparts. Overall, this cFRET imaging methodology will enable a new generation of assay and sensing applications that benefit from spatial and temporal resolution, including microfluidic systems, arrays or other patterned interfaces, and cellular analysis. With the cFRET imaging methodology now developed and validated, such applications can now be explored.

■

ASSOCIATED CONTENT

S Supporting Information *

Additional experimental methods, details of data analysis, additional results and discussion on the effect of excitation intensity and integration time, plate reader calibration data, comparison of homogeneous plate reader and microscope assays, and additional capillary imaging assays. The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acs.analchem.5b01946.

■

AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected]. Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS The authors acknowledge support for this research from the Natural Sciences and Engineering Research Council of Canada (NSERC), the Canada Foundation for Innovation (CFI), and the University of British Columbia. W.R.A. is grateful for a Canada Research Chair (Tier 2) and a Michael Smith Foundation for Health Research Scholar Award.

■

REFERENCES

(1) Zheng, K. Y.; Yuan, X.; Goswami, N.; Zhang, Q. B.; Xie, J. P. RSC Adv. 2014, 4, 60581−60596. (2) Pisanic, T. R.; Zhang, Y.; Wang, T. H. Analyst 2014, 139, 2968− 2981. (3) Wu, C. F.; Chiu, D. T. Angew. Chem., Int. Ed. 2013, 52, 3086− 3109. (4) Lim, S. Y.; Shen, W.; Gao, Z. Q. Chem. Soc. Rev. 2015, 44, 362− 381. (5) Yang, N.; Chen, X.; Ren, T.; Zhang, P.; Yang, D. Sens. Actuators, B 2015, 207, 690−715. (6) DaCosta, M. V.; Doughan, S.; Han, Y.; Krull, U. J. Anal. Chim. Acta 2014, 832, 1−33. (7) Wu, M.; Petryayeva, E.; Algar, W. R. Anal. Chem. 2014, 86, 11181−11188. (8) Kim, H.; Ng, C. Y. W.; Algar, W. R. Langmuir 2014, 30, 5676− 5685. (9) Algar, W. R.; Wegner, D.; Huston, A. L.; Blanco-Canosa, J. B.; Stewart, M. H.; Armstrong, A.; Dawson, P. E.; Hildebrandt, N.; Medintz, I. L. J. Am. Chem. Soc. 2012, 134, 1876−1891. (10) Algar, W. R.; Malanoski, A. P.; Susumu, K.; Stewart, M. H.; Hildebrandt, N.; Medintz, I. L. Anal. Chem. 2012, 84, 10136−10146. (11) Algar, W. R.; Ancona, M. G.; Malanoski, A. P.; Susumu, K.; Medintz, I. L. ACS Nano 2012, 6, 11044−11058. (12) Yu, W. W.; Peng, X. Angew. Chem., Int. Ed. 2002, 41, 2368− 2371.

■

CONCLUSIONS In conclusion, we have demonstrated and validated a new imaging method suitable for use with assays that utilize cFRET probes. Using the proteolytic activity of trypsin and chymotrypsin as a test system, we have shown that threecolor, filter-based imaging with a fluorescence microscope can reproduce results from multiplexed assays that were previously E

DOI: 10.1021/acs.analchem.5b01946 Anal. Chem. XXXX, XXX, XXX−XXX

Technical Note

Analytical Chemistry (13) Li, J. J.; Wang, Y. A.; Guo, W. Z.; Keay, J. C.; Mishima, T. D.; Johnson, M. B.; Peng, X. G. J. Am. Chem. Soc. 2003, 125, 12567− 12575. (14) Sapsford, K. E.; Pons, T.; Medintz, I. L.; Higashiya, S.; Brunel, F. M.; Dawson, P. E.; Mattoussi, H. J. Phys. Chem. C 2007, 111, 11528− 11538. (15) Petryayeva, E.; Algar, W. R. Anal. Chem. 2014, 86, 3195−3202. (16) Aldeek, F.; Safi, M.; Zhan, N. Q.; Palui, G.; Mattoussi, H. ACS Nano 2013, 7, 10197−10210. (17) Algar, W. R.; Blanco-Canosa, J. B.; Manthe, R. L.; Susumu, K.; Stewart, M. H.; Dawson, P. E.; Medintz, I. L. Methods Mol. Biol. 2013, 1025, 47−73. (18) Photometrics. Dual View and Quad View Simultaneous Imaging Systems, http://www.photometrics.com/products/multichannel/ (accessed May 2015). (19) Algar, W. R.; Kim, H.; Medintz, I. L.; Hildebrandt, N. Coord. Chem. Rev. 2014, 263−264, 65−85.

F

DOI: 10.1021/acs.analchem.5b01946 Anal. Chem. XXXX, XXX, XXX−XXX