Ultra-fast cycling for multiplexed cellular fluorescence imaging

Abstract

Rapid analysis of single and scant cell populations is essential in modern diagnostics, yet existing methods are often limited and slow. Here we describe an ultra-fast, highly efficient cycling method for the analysis of single cells based on unique linkers for tetrazine (Tz) / trans-cyclooctene (TCO) mediated quenching. Surprisingly, the quenching reaction rates were more than 3 orders of magnitude faster (t_1/2 < 1 sec) than predicted. This allowed multi-cycle staining and immune cell profiling within an hour, leveraging the accelerated kinetics to open new diagnostic possibilities for rapid cellular analyses.

Graphical Abstract

Gravitationally enhanced?: Black hole quenchers (BHQ) interact with fluorophores to accelerate tetrazine/trans-cyclooctene (Tz/TCO) click reactions by more than three orders of magnitude. Cells stained with fluorophore-TCO labeled antibodies can be efficiently quenched in seconds with a BHQ3-Tz, eliminating time-consuming bleaching/cleavage steps and enabling rapid cellular analyses for cancer diagnostics.

Introduction

Molecular analyses of cancer cells are essential in establishing diagnosis and guiding available treatments^[1] In an ideal world, one would like to harvest cancers frequently and in the least invasive manner so that molecular information can be obtained periodically through treatment and cancer evolution.^[2] “Liquid biopsies”, i.e. the interrogation of circulating tumor cells^[3], extracellular vesicles^[4], or cell-free DNA in the peripheral blood, provide one such option, but detection of actionable events is rare and overall sensitivities can be low.^[5] More importantly, circulating tumor diagnostics cannot currently be traced back to their anatomical origin, whether primary tumor or metastatic site. This limits the ability to correlate molecular events with radiographic/imaging measures of cancer behavior, invasiveness, and progression.

An alternative method is fine needle aspiration (FNA) that yields cells rather than tissue from a tumor, are inherently of known localization, and which can be processed expeditiously, i.e. do not require embedding or sectioning. FNA are obtained with small gauge needles (20–25 G) and are generally well tolerated.^[6] As such, image guided FNA are ideally suited for repeat sampling and have a very low risk of procedural complications. The challenge in processing these cellular samples is that they can be scant (often < 1,000 cells per pass), limiting the number of special stains that can be done, and also delicate, lacking the structural scaffold of intact tissue architecture. Even when processed with fluorescent antibodies, the number of different stains is practically limited to 4–6 and often not sufficient for in depth cancer cell profiling for diagnosis or treatment assessment. This limitation also extends to immune profiling, where significantly more than 4–6 markers need to be interrogated so that analysis reflects the representative immunocyte populations in the tumor microenvironment.^[7] For these reasons, single cell cycling technologies need to be developed that allow repeat staining, destaining, and re-staining of harvested cellular samples for better therapy assessment in both cancer cells and host immune cells.

Most fluorescent cycling methods^[8] were originally developed for paraffin embedded tissue sections that can withstand harsh destaining/quenching conditions. Unfortunately, these harsh conditions typically require oxidants for bleaching at strongly alkaline pH (e.g. 4.5% H₂O₂, 24mM NaOH, pH >12) and are not well suited for cellular FNA samples. Furthermore, it is not uncommon for other antibody-DNA cycling technologies to require a significant investment in nucleic acid tags/technologies and take hours-days of sample processing, including ABCD^[9] and SCANT^[10] Similar technical hurdles accompany other recent methods for antibody-DNA based imaging, including intricate chemical steps for DNA barcode activation and antibody-DNA bioconjugation, and/or complex fluidics required for cycling multiple sequential staining solutions.^[11] We thus set out to explore and develop potentially faster and gentle methods of single-cell cycling. We reasoned that recent advances in bioorthogonal chemistry could be harnessed to develop a new generation of ultra-fast clickable fluorophores and quenchers (FAST probes), superseding the requirement for bleaching or cleavage events. We implemented tetrazine (Tz) / trans-cyclooctene (TCO) chemistry for site-specific delivery of fluorescence quenchers and observed both highly efficient quenching across the color spectrum and a remarkable acceleration in the chemical reaction kinetics. The versatile reagents are capable of ultra-fast (< 1 sec) quenching of fluorescence and allow multichannel imaging of 20–30 markers within an hour.

Results and Discussion

Tz-TCO based antibody labeling and quenching

The overall principle of the technology involves efficient, fast and maximum quenching of antibody-associated fluorochromes of different wavelengths (Fig. 1). This is achieved via a modular linker between fluorochromes and antibodies with an embedded TCO for clicking with a tetrazine-quencher. In prior work^[12], we developed a BODIPY-TCO and Tz-black hole quencher (BHQ) pair as a tool to study mechanistic aspects of the rapid kinetics of click-to-release reactions. Those probes exhibited substantial (>90%) quenching of the dye fluorescence in a 1:1 complex, leading us to speculate that cooperative (multivalent) dye-quencher interactions in the context of a labeled antibody could produce useful degrees of quenching. In a proof of principle study, we labeled a commercially available Alexa-Fluor 488 labeled secondary antibody with TCO (4–5 TCO/antibody) and observed a marked reduction in antibody fluorescence after click-reaction with the BHQ10-tetrazine (Supporting Information, Fig S1).

Fig. 1: Chemical scheme of synthesis and strategy.

A. Synthesis route for preparation of TCO-linked fluorophores (FAST probes) built on a lysine scaffold with a PEG₄ linker for efficient antibody conjugation: i. TSTU, DIPEA; ii. H₂N-PEG₄-CO₂H; iii. DCM/TFA (20%); iv. rTCO(axial)-PNP, DIPEA; v. piperidine (7.5%). The core linker can be functionalized with any amine-reactive fluorophore of choice. Inset: Excess TSTU can be used for rapid activation of the PEG4-CO₂H and neutralized immediately thereafter with ENBA, circumventing a range of purification and antibody-conjugation obstacles. B. BHQ3-amine is coupled with HTz-PEG₅-NHS to yield BHQ3-Tz in one step. C. Structural schematic of the BHQ3-fluorophore quenching interaction after TCO-Tz click; AF647 is depicted here.

We proceeded to synthesize a series of TCO-fluorophore (TCO-Fl) reagents built around lysine as a ternary scaffold, equipped with a PEG₄ linker for antibody conjugation (Fig 1A). While the synthesis was straightforward, the final TCO-fluorophore conjugates exhibited atypically poor stability relative to our extensive past experience.^[13] Both commercially available TCO and the hydrophilic alternative dTCO^[14] proved prone to degradation to the unreactive cis-cyclooctene during storage and subsequent handling of the NHS esters, particularly when exposed to any acidified reverse phase conditions during purification. To mitigate this issue we adopted the axial 3-OH-functionalized TCO (release-TCO, rTCO), which has enhanced stability under a range of biochemical exposures, for further testing.^[15] In prior work, we and others^[16] had observed that H-tetrazines induce minimal release in aqueous solution and that the net release for aryl-H tetrazines is negligible, particularly at physiologic pH and above (vide infra). Given the goal of rapid image cycling, the fast click kinetics of aryl-H tetrazines also harmonize well with our overall objectives.

While isolation of TCO-fluorophore-activated esters under gently buffered reverse phase conditions (pH 4.5–5) was feasible, we elected to develop a new method for rapid microscale activation immediately prior to antibody conjugation (Fig 1A, inset, and Supporting Information, Fig S2). We recognized that the negligible reactivity of secondary amines with NHS esters could be exploited to neutralize excess activating reagents, such as TSTU (N,N,N′,N′-Tetramethyl-O-(N-succinimidyl)uronium tetrafluoroborate). Rapid intramolecular ring closure converts 4-(ethylamino)butanoic acid (ENBA) succinimidyl ester to inert N-ethyl-2-pyrrolidone to quench the reaction with no impact on the TCO-fluorophore-NHS. This allows use of TSTU in a significant molar excess to i) achieve rapid conversion to the NHS ester (seconds); ii) reduce the impact of trace amounts of water on reliable dye activation; iii) simplify microscale reagent titration for applications in a biological context. Importantly, this approach also enables kinetic discrimination between the PEG₄ carboxylic acid and the comparatively hindered 2’-carboxylic acid on many xanthene dyes (e.g. AF488, AF594).^[17] The resulting dye solution contains no competing reactive species (i.e. no residual activator or contaminating active esters that could react with the antibody) and can be aliquoted for immediate use or stored (≤ −20°C) for subsequent labeling reactions.

In parallel, we screened three different types of quencher (BHQ10, BHQ3, and IRDye QC-1^[18]) with relevant fluorophores and selected BHQ3 (Fig 1B), which showed the best quenching efficiency with all the dyes that we tested in this study (Supporting Information, Fig S3). BHQ2 probes were synthesized as well, but proved insufficiently soluble for use in this format. From the broad wavelength compatibility, we infer a significant contribution from static/contact quenching, which does not require spectral overlap between the fluorophore and quencher.^[19] Our design objectives included adequate aqueous solubility (≥ 25μM in PBS) to achieve rapid quenching given the expected click reaction kinetics, and a linker architecture compatible with cooperative quenching, such that a BHQ3-Tz tethered to one TCO-Fl site may be within quenching range of adjacent dye molecules. We thus selected an HTz-PEG₅-NHS linker, which conferred sufficient solubility and linker length (rendered to scale in an extended conformation, Fig 1C). With a series of reagents synthesized, we set out to systematically explore the labeling of model antibodies and their quenchability.

Ultra-fast and highly efficient quenching

We performed a series of experiments to quantitate: i) the antibody labeling efficiency of these TCO-Fl conjugates; ii) antibody brightness as a function of degree of labeling (DOL); iii) effect of the TCO-Tz quenching method on different fluorochrome-modified antibodies across the visible wavelength spectrum (Fig 2). The FAST linkers displayed excellent labeling characteristics with all the fluorophores and antibodies tested, with a predictable and efficient DOL as a function of dye concentration. Mouse splenocytes stained with a FAST647 anti-CD4 antibody matched or exceeded the staining performance and brightness of a conventional commercial standard (Fig 2A). The brightness of FAST labeled antibodies was substantially quenched upon treatment with BHQ3-Tz: pre/post fluorescence emission spectra of a FAST647-labeled anti-EGFR antibody (cetuximab) illustrate a >99% decrease of the dye emission (Fig 2B). We went on to assess the quenching dynamics in the cellular/imaging context, staining A431 cells with the respective cetuximab conjugates and collecting images before and after a three minute incubation with 20μM BHQ3-Tz at pH 9 (Supporting Information, Fig S4). Quenching was best for the far red dye AF647, which has the greatest degree of spectral overlap with BHQ3, but the quantitative reduction in brightness was similar for different fluorochromes spanning the visible range, with a general trend toward superior quenching at higher DOL. To further delineate the dynamics of the rTCO/aryl-HTz ligation and assess any impact from release (which would liberate the quencher from the antibody), we measured the brightness of FAST647-Cetuximab/BHQ3-Tz quenching over a four hour time course as a function of pH (Fig 2C). Addition of BHQ3-Tz produced a uniform ~99% reduction in the baseline fluorescence intensity, irrespective of pH (see below for detailed studies of quenching kinetics). At pH 5–6, the fluorescence rebounded to approximately 15% of the initial intensity, with a faster rate at the lower pH, indicative of partial release of the quencher. In contrast, the net rebound in brightness was minimal at physiologic pH (≤ 5% in PBS, pH 7.4), and negligible at pH 9, with no significant change in the observed quenching at t = 4 hrs.

Fig. 2: Quantitative data: labeling and quenching efficiency.

A. (i) AF647-rTCO (FAST-AF647) labeling of an anti-CD4 antibody produced bright conjugates with excellent efficiency and staining brightness proportional to DOL, matching or exceeding the brightness of the commercial antibody (Fl-MAb). (ii) Representative images of mouse splenocytes stained with FAST-AF647 anti-CD4 and the Fl-MAb reference. B. Quenching efficiency assessed by fluorimeter and by microscopy. (i) The fluorescence emission spectrum of FAST-AF647 anti-CD4 demonstrates >99% reduction in signal after treatment with BHQ3-Tz in PBS. (ii) Residual MFI in A431 cells after quenching of FAST-labeled anti-EGFR antibody (cetuximab) staining for four different dyes (AF488, AF555, AF594, and AF647). (iii) Representative images of EGFR staining and quenching in A431 cells. Cells were stained with AF488-rTCO or AF594-rTCO cetuximab (5μg/mL) and imaged before and after (at right) quenching with 20μM BHQ3-Tz for 5 minutes. C. To quantify any impact of rTCO release, BHQ3-Tz quenching of FAST-AF647 labeled cetuximab was monitored longitudinally as a function of pH. While a slow rebound in brightness is evident at mildly acidic pH, consistent with partial release of BHQ3-Tz/rTCO from the antibody, quenching is exceptionally stable at pH 9.

Having quantified the desired magnitude and stability of quenching, we sought to determine the minimum incubation time required to achieve high (>98%) quenching efficiency after addition of BHQ3-Tz (20μM) to stained cells. The fast click kinetics of the benzylamino-H-Tz with rTCO (k = 7200 M⁻¹s⁻¹ in PBS at 25°C, unpublished data) enabled rapid quenching within a minute, indistinguishable from our initial five minute benchmark (Fig 3A). To our surprise, however, we observed equally efficient quenching as we progressively decreased the BHQ3-Tz incubation time to as little as ten seconds, as short as experimentally feasible. Conventionally stained cells exhibited no discernible change in brightness when exposed to BHQ3-Tz under these conditions (Supporting information, Fig S5A).

Fig. 3: Quantitative data: quenching kinetics for FAST probes.

A. A431 cells stained with FAST-AF647 cetuximab and DAPI for nuclear reference were treated with 20μM BHQ3-Tz in PBS-bicarb (pH 9) for progressively shorter amounts of time. Quenching remained near-quantitative after incubation for as little as 10 seconds. B. Experimental design for fluorescence kinetics: FAST-AF488 was selected for initial studies to minimize spectral interference between ex/em wavelengths and the BHQ3 absorbance. At just 2 μM BHQ3-Tz in PBS, quenching of labeled Ab fluorescence was exceptionally rapid (shaded interval, 2 seconds). C. Systematic kinetics for AF647, AF594 and AF488 probes in PBS-bicarb (pH 9) as a function of BHQ3-Tz concentration revealed remarkable dye-specific accelerations in click rates relative to the expected rate for the Tz-TCO pair. D. Further kinetic studies of FAST-labeled antibodies in PBS-bicarb (pH 9) demonstrate that the cumulative acceleration is even greater for the multivalent antibodies and dependent on dye, DOL, and BHQ3-Tz concentration.

To investigate further, we implemented a fluorimeter-based assay to directly observe the click kinetics for BHQ3-Tz and FAST-labeled antibodies/probes. At nanomolar fluorophore concentrations this allows real-time observation of pseudo first-order click rates with negligible optical interference, provided the BHQ3-Tz concentration in the cuvette is kept appropriately low. Addition of BHQ3-Tz produced no effect on the brightness of conventionally-labeled secondary antibodies (Supporting Information, Fig S5B); in contrast, at a matched BHQ3-Tz concentration of 1–2 μM (≤ 1/10th the amount used in our initial imaging experiments) we found that FAST-AF488 antibody fluorescence was quenched remarkably rapidly, with a half life of less than half a second (Fig 3B). The presence of a 100-fold excess of unlabeled antibody (relative to FAST-Ab, used to block nonspecific adsorption to the cuvette, see methods section 6A in the Supporting Information), produced no effect on the rate, suggesting that the kinetics are not being driven by BHQ3-protein interactions. Given the multivalency of antibody labeling, these rates reflect a cumulative multi-click quenching process, averaged across multiple FAST label (dye-TCO) pairs, not a single chemical event. To better assess the intrinsic rate constants, we assessed the quenching rate for the free FAST probes in solution as a function of BHQ3-Tz concentration. Click rates were dye-dependent, exhibited classical concentration-dependent (pseudo first order) kinetics, and were again dramatically accelerated relative to the parent Tz-TCO (Fig 3C). Measured second-order rate constants ranged from ~30-fold faster than expected for FAST-AF488 to nearly 440-fold faster than expected for FAST-AF647.

When we extended these methods to reassess kinetics for the BHQ3-Tz reaction with FAST-labeled antibodies, we unexpectedly observed even more extreme acceleration of the quenching rates, rising in tandem with increasing DOL (Fig 3D). The fluorescence of the AF488-labeled Ab was quenched with an apparent second order rate constant as high as 7.4 × 10⁶ M⁻¹s⁻¹ (t_1/2 = 3.7s at 25nM BHQ3-Tz, ~1000-fold accelerated). Mirroring the trend for the free dyes, the apparent quenching rate for the AF647-labeled Ab was even faster at 2.4 × 10⁷ M⁻¹s⁻¹ (t_1/2 = 0.6 seconds at 50nM BHQ3-Tz), accelerated by >3300-fold compared the expected bimolecular click rate. Concordant rate accelerations and DOL/concentration dependence were observed for additional control antibodies, confirming the generality of this effect (Supporting information, Fig S6). While the free FAST probes exhibited excellent single exponential fits under pseudo first order conditions (Fig 3C), the FAST-antibody quenching was better fit to a double exponential. Neither did the observed rate vary linearly with quencher concentration, with the apparent second order rate rising as the concentration of BHQ3-Tz decreased, suggestive of a cooperative quenching interaction between dyes and in keeping with the observed accelerating effects of DOL (Supporting information, Fig S6). While the efficiency of quenching in the biomolecular context is evident from the imaging experiments, to further explore the general biocompatibility of the FAST-acceleration we measured the reaction kinetics between FAST-labeled antibodies and BHQ3-Tz in PBS with added BSA (to further assess the impact of high protein concentration) and in cell culture media. Click/quenching rates remained profoundly accelerated, with only subtle changes in the temporal profile (Supporting Information, Fig S7).

Intramolecular contact quenching for fluorophore-BHQ pairs is well known for dual-labeled oligonucleotide^[20] and peptide^[21] cleavage probes. In the oligonucleotide context where this has been studied more quantitatively, fluorophore-quencher interactions subtly perturb melting temperatures^[19] and can promote intramolecular dimer formation with an affinity sufficient to form a stemless hairpin.^[22] We therefore hypothesize that dye-dependent fluorophore-BHQ3 interactions drive a reversible molecular association of sufficient affinity/duration to promote Tz-TCO ligation, even at nanomolar concentrations. The temporal dynamics of this process are intriguing, given that no fluorescence quenching (a process occurring on the nanosecond timescale) is observed in the absence of the click reaction. Future experiments that vary the linker geometry, reaction temperature, and buffer/solvent will be required to further delineate mechanistic details of this effect. The scope of the kinetic acceleration and its impact on other bioorthogonal reaction pairs will also be of interest.

Imaging of multiple targets in single cancer cells

To move from single-channel benchmark studies to the clinically relevant objective of imaging/ quenching multiple protein markers within individual cancer cells, we next sought to validate multichannel FAST staining. We selected targets of known importance in cancer diagnosis and complementary spatial distributions, including cell surface, cytoplasmic, and nuclear markers. In one illustrative example, we stained A431 epidermoid cancer cells for EGFR, S6 and phosphor-S6 (pS6) with three antibodies simultaneously and imaged the same field of view before and after quenching with BHQ3-Tz (Fig 5). Quantitative intensity profiles for each channel span the stained and quenched images to allow visualization of the quenching efficiency, which proved to be excellent across all cellular compartments. The residual background signal in the quenched images is <5% of the peak staining intensity, inclusive of channel-specific autofluorescence (higher at green wavelengths than red). Given the complexity of normalizing absolute fluorescence intensity between individual clinical samples, quantitative co-expression and the ratio of paired markers (e.g. pS6/S6 as demonstrated here) are particularly important for both cellular classification and as a tool to measure therapeutic protein inhibition.^[10] With a working method for multi-color cycling, we set out to profile co-expression of multiple protein markers in a murine model of clinical FNA samples.

Fig. 5: Immune cell profiling of tumor FNA.

Representative images obtained by cyclic imaging of immune markers in a mouse tumor FNA sample are shown. A. Twelve markers were imaged using three FAST-probe fluorophores (AF647: red, AF594: magenta, AF488: green in the images) in four imaging cycles. All images show the same set of cells within a zoomed-in area of a single field of view in order to appreciate the patterns of fluorescence signal of immune markers expressed in individual cells. DAPI was used to stain the nuclei of all cells imaged in each cycle for cycle-to-cycle alignment (scale bar: 20 μm). B. From a total of 1846 cells analyzed, the frequencies of different immune cell types (CD8+ T cells, CD4+ T cells, macrophage, dendritic cells, neutrophils) and key subsets (PD-1, CD163/CD206) in CD45⁺ immune cells were quantified. Each immune cell type was identified using the selected combinations of markers as indicated.

Tumor immune cell profiling by cyclic imaging

To show applicability of the method for tumor immune cell profiling, we imaged 12 immune markers (CD45, CD8, CD3, CD4, PD-1, CD11b, F4/80, MHCII, CD163, CD206, Ly6G, CD11c; Supporting information, Table S1) in cells directly harvested from MC38 mouse colon cancer, a highly immunogenic tumor model. The choice of the tumor model and immune cell markers was based on our previous studies on tumor-infiltrating immune cell populations.^[23] FNA samples were obtained from subcutaneously implanted MC38 tumors (Supporting Information, Biological Methods) and imaged in successive cycles of FAST staining. Cell nuclei were stained with DAPI for alignment of images across cycles. CD45 was imaged in cycle 1 as a pan-hematopoietic marker for identification of the immune cells in each field of view and selection of optimal imaging locations across the stained slide. In the FNA sample presented in Fig. 5, 11 different field of views with a sufficient number of cells were selected during cycle 1, and images of subsequent cycles were taken from the same positions to record the same set of cells. For image analysis, CD45⁺ immune cells were first identified among DAPI⁺ cells by cell segmentation of images to create a mask for all immune cells to analyze (Supporting Information, Fig S8). Image sets of each field of view were aligned using normalized cross correlation^[24], and the CD45 mask was applied to measure the fluorescence intensity of immune markers in each CD45⁺ cell. Cell Profiler (https://cellprofiler.org/) was used for the image analysis. A total of 1846 CD45⁺ immune cells were detected from images of 11 fields of view. Distinct immune cell populations within CD45⁺ cells were identified based on their expression level of the immune markers (Fig. 5B). Thresholds for the positivity of a cell for each marker were determined from the fluorescence intensity distribution. The frequency of relevant immune cell populations (CD8⁺ T cells, CD4⁺ T cells, macrophages, dendritic cells, and neutrophils) was quantified as shown in Fig. 5B. We were also able to analyze the frequency of key immune population subsets, such as CD8⁺ T cells expressing immune inhibitory receptor PD-1, an important target of cancer immunotherapy with immune checkpoint inhibitors. In addition, the expression levels of CD163 and CD206 in macrophages were assessed to analyze the immunosuppressive phenotype, which has been reported to be observed in tumor-associated macrophages.^[25]

Conclusions

Here we present an ultra-fast, highly efficient but yet gentle quenching technology for multiplexed protein profiling in single cells. We use TCO-Tz click chemistry for staining, quenching, and cycling of fluorescence imaging and introduce a new method for rapid dye activation immediately prior to antibody labeling. Using this technology, we successfully stained 12 immune markers within one hour from the same cells, which enabled profiling of immunocyte populations in the tumor microenvironment.

The ultra-fast quenching is a key feature of the technology as it allows rapid repeat multi-color staining. An incubation of less than ten seconds at low micromolar concentrations is sufficient to remove > 95% of the fluorescence signal from the previous cycle. One interesting aspect is that the observed quenching speed is unexpectedly much faster than the predicted bimolecular TCO-Tz reaction rate; this was observed for all fluorochromes tested (Fig. 3C). The reasons for this ultrafast quenching remain partially speculative at this point, but the experimental evidence suggests that transient complexation between dye and quencher markedly accelerates the TCO-Tz click through an effect on local concentration. The acceleration of other bioorthogonal reaction rates in this architecture is currently under study. Irrespective of the exact mechanism, the ultrafast quenching is a highly distinct feature compared to alternative methods where at least 30 minute incubation with chemicals or water is necessary. This current method can be practically used to image many different sample types including FNA and tissues as the process does not involve usage of any strong chemicals nor other harsh destaining conditions.

Going forward, we aim to expand a set of markers of interest to profile important proteins that contain molecular information for the development of diagnostics and therapeutics. For example, cancer and host immune cells from specific tumor sites could be profiled at different time points during cancer treatment to monitor and predict response. This will provide a better representation of immunologic dynamics that cannot be easily detected by analyzing peripheral immune cell populations that do not reflect the tumor microenvironment.

Fig. 4: Imaging of multiple targets in single cells.

Three FAST probes (AF488, AF594, AF647) were used to stain multiple targets in single cells. Anti-pS6 was labeled with AF488-rTCO, anti-EGFR with AF594-rTCO, and anti-S6 with AF647-rTCO to validate the multi-target imaging ability of the FAST probes. After antibody staining (1–5μg/mL; see the Supporting Information) and washing, the cells were incubated with BHQ3-Tz (10μM in PBS-Bicarb, pH 9, one minute) for quenching. Intensity profiles before and after BHQ3-Tz quenching demonstrate quantitative reduction of the fluorescent intensity of all three targets to background levels.