Rhodamine-Derived Fluorescent Dye with Inherent Blinking Behavior

Subscriber access provided by UNIV OF DURHAM

Article

Rhodamine-derived fluorescent dye with inherent blinking behavior for super-resolution imaging Patrick J. Macdonald, Susan Gayda, Richard A. Haack, Qiaoqiao Ruan, Richard J. Himmelsbach, and Sergey Y. Tetin Anal. Chem., Just Accepted Manuscript • DOI: 10.1021/acs.analchem.8b01645 • Publication Date (Web): 25 Jun 2018 Downloaded from http://pubs.acs.org on June 25, 2018

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 9 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Analytical Chemistry

Rhodamine-derived fluorescent dye with inherent blinking behavior for super-resolution imaging Patrick J. Macdonald§, Susan Gayda§,†, Richard A. Haack, Qiaoqiao Ruan, Richard Himmelsbach, and Sergey Y. Tetin* Applied Research and Technology, Abbott Diagnostics Division, Abbott Laboratories, Abbott Park, IL 60064 Tel: 224-668-4661 Email: [email protected] ABSTRACT: Super-resolution microscopy enables imaging of structures smaller than the diffraction limit. Single-molecule localization microscopy methods, such as PALM and STORM, reconstruct images by plotting the centroids of fluorescent point sources from a series of frames in which only a few molecules are fluorescing at a time. These approaches require simpler instrumentation than methods that depend on structured illumination, and thus are becoming widespread. The functionalized rhodamine derivative reported in this paper spontaneously converts between a bright and dark state due to pH-dependent cyclization. At pH 7, less than 0.5% of the dye molecules are fluorescent at any given time. Blinking occurs on timescales of seconds to minutes and can therefore be used for single-molecule localization microscopy without sample treatment or activation. The dye is bright and straightforward to use; it is easy to synthesize and functionalize, thus it has potential to become a new and powerful addition to the toolset for super-resolution imaging.

keywords: super-resolution imaging, dSTORM, fluorescence blinking, fluorescent dyes, fluorescence microscopy

Visualization of submicron structures offers invaluable information about live cell machineries. It may also provide critical support for the engineering of various micro- and nanodevices. The power of imaging has been greatly expanded by the advent of super-resolution techniques. For centuries, the optical resolution of microscopy was constrained by the diffraction limit which is proportional to the wavelength of light () and the numerical aperture of the objective (), = . When using visible light microscopes, a typical diffraction limit of ~ 200 nm provides good images of cells. However, many structures and cellular machines should be studied at the level of macromolecular complexes, significantly smaller than the diffraction limit, but not beyond the reach of super-resolution. One group of methods, including stimulated emission depletion (STED) 1 and reversible saturable optical fluorescence transitions (RESOLFT),2 got around this limit by manipulating excitation light in specific patterns to suppress fluorophores around the target region, thus allowing fluorescence only from a small, central section of the point spread function. Another group of super-resolution methods, including photo-activation localization microscopy (PALM),3 stochastic optical reconstruction microscopy (STORM),4 and direct stochastic optical reconstruction microscopy (dSTORM),5 spreads out the signals of otherwise unresolvable fluorophores in time, repeatedly activating only a few molecules at a given moment, so the centers of their point spread functions can be used for recording and reconstructing

a high resolution image. Such techniques have achieved resolutions as low as 10-20 nm6, 7 and granted visual access to some of the engines and scaffolds of the biological world. Indeed, many super-resolution instruments and setups have been commercially available for several years, and superresolution approaches, despite their complexity, are beginning to edge towards the mainstream.6, 8, 9 A comprehensive review of the current methods of single-molecule localization microscopy and their applications for super-resolution imaging in eukaryotic cells was published recently.10 The introduction of STORM and dSTORM, which utilize conventional fluorophores such as cyanines, rhodamines or oxazines, was a big step toward making super-resolution microscopy simpler and expanded its applications in biology and materials research. STORM requires a switching pair of dyes, one of which acts as an activator upon illumination with a second laser beam, 11, 12 while dSTORM is based on spontaneous or light-induced stochastic recovery of individual fluorophores from a dark state.13 Such a process can occur when reducing thiol reagents quench the fluorophore’s triplet state, followed by the recovery of fluorescence after reoxidation with dissolved oxygen. Similar switching between the on- and off- states can also result from intramolecular cyclization. For instance, switchable fluorophores developed in Hell’s group14 and a spirolactam introduced by Kim et al.15 undergo cyclizations triggered by laser activation or metallic ion binding, respectively.

1 ACS Paragon Plus Environment

Analytical Chemistry 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

In this paper, we report on a spontaneously switching dye from the rhodamine family which, at physiological pH, cycles between a closed, non-fluorescent form and an open, fluorescent form. We discovered spontaneous blinking of this newly-synthesized, rhodamine secondary amide while working on developing single-isomer rhodamines for protein labeling. The benefit of such dyes is that they can be employed for dSTORM without any dependence on activation/suppression laser excitation or reducing/oxidizing agents in solution. This, in turn, may lead to reduced photobleaching and more flexible applications. The singleisomer rhodamine is simple to synthesize making it a practical choice for super-resolution imaging. Moreover, such simplicity opens the possibility of further structural modifications for tuning photophysical properties of the dye. In early 2014 we presented a poster reporting on spontaneous blinking of this new rhodamine dye and its application for producing super-resolution images.16 In the same year, Uno et al.17 published a comprehensive paper describing the preparation and evaluation of silicon containing rhodamine (SiR) derivatives that spontaneously blink upon intramolecular spirocyclization. One of them, a hydroxymethyl substituted analog (HMSiR), was optimized and conjugated with self-labeling protein tags for superresolution imaging of microtubules in live cells. In a recent paper, Takakura et al. transformed HMSiR into an environment-sensitive membrane probe. 18 We will compare HMSiR and our rhodamine dye later in this paper. Overall, our single-isomer, functionalized rhodamine derivative (FRD) is prepared using off-the-shelf components and simple chemical reactions. Both, the blinking (FRD-B) and non-blinking (FRD-ON) versions can be easily produced in large quantities. We present the properties of FRD and describe its pH-dependent transition from the non-fluorescent to fluorescent form. We made use of the blinking dye behavior to generate super-resolution images of the antibody-coated microparticles that are commonly used in diagnostic immunoassays. We also show images of the tubular structures of HIV-1 p24 capsid and of actin in stained HeLa cells.

dye powder. Fluorescence lifetime measurements were performed on a ChronosBH fluorometer (ISS, Champaign, IL) using fluorescein solution (pH 8) as a reference. pH was measured with an Orion 370 (Thermo Fisher Scientific) pH meter. pH Buffered Solution Set: 50 mM citric acid was titrated with 50 mM Na2HPO4 to prepare solutions with pH 3-6. 50 mM NaH2PO4 was titrated with Na2HPO4 to prepare solutions with pH 6.5-8. We alkalized 50 mM boric acid with 6N NaOH, and titrated back down with boric acid to make solutions with pH 11-8.5. We thus generated a series of buffers with pH increasing by 0.5 from 3 to 10. NaCl was added to all solutions to a final concentration of 150 mM. Single-molecule analysis: FRD-B was conjugated to an antibiotin antibody. The antibody conjugate was diluted to 300 pM in HBS-EP (GE Healthcare Life Sciences, Marlborough, MA) and incubated overnight at room temperature in separate wells of a PEG-biotin-coated glass coverslip. After washing, samples were incubated for 10 minutes in solutions with different pH, then imaged using total internal reflection fluorescence (TIRF) microscopy. An intensity threshold was selected, and each frame of the single-molecule intensity trace was identified as being in the on-state (above threshold) or offstate (below threshold). Off-transitions were logged across all single-molecule traces for above-threshold frames, n, in which the next frame, n+1, fell below the threshold, and ontransitions logged for below-threshold frames with the next frame above the threshold. On-times and off-times were calculated by subtracting the subsequent off-transitions from on-transitions, and vice versa, then using the 50-ms exposure time plus 29.45-ms readout time to convert frame numbers to times. Single-molecule TIRF images were taken on an Olympus IX81 microscope (Center Valley, PA) using objective-based TIRF. A LightHub laser combiner (Omicron, Rodgau, Germany), connected to the microscope via optical fiber, provided four laser wavelengths: 405, 488, 561, and 638 nm. Excitation and emission light passed through a quad filter cube (U-N84000v2; Chroma, Bellows Falls, VT), and was focused into the sample with 100x/1.49 oil immersion TIRF objective. Samples were typically excited with laser powers of approximately 0.5 kW/cm2, and 50-ms or 100-ms acquisitions were recorded on an iXon Ultra EMCCD camera (Andor, Belfast, UK). Super-resolution imaging: The following model systems were imaged using dSTORM super-resolution. 5-µm polystyrene magnetic microparticles were coated with goat anti-mouse antibody (Abbott Laboratories, 200 ug/mL antibody for 1% w/w microparticles). Subsequently, the pentafluorophenyl active ester, FRD-B2 (60 µM) was incubated with the microparticles (0.1% solids) for 3 hours at room temperature and then washed, thus directly labeling the antibodies on the microparticles. HIV-1 core antigen tubular assemblies were generously provided by B. Pornillos and coworkers (University of Virginia), and were similarly directly labeled with the active ester FRD-B2, washed, sonicated to resuspend and assemble, and then transferred to the imaging slide. HeLa cells were obtained in-house (Abbott Laboratories), fixed, and then stained. The fixed cells were incubated with 0.5 µM phalloidin-biotin for 1 hr; washed; incubated for another hour with 5 µg/mL of FRD-B3-labeled

EXPERIMENTAL SECTION Fluorescent Dyes: The development and synthesis of the core dye is described in a previously published work.19 A derivative of the core dye was functionalized incorporating either a polyethylene glycol linker, FRD-B, or an aliphatic hydrocarbon linker, FRD-B2 (see Synthesis section). Both exhibit blinking behavior and were used in imaging experiments. We also prepared FRD-ON, a dye containing a hydrocarbon linker in which the nitrogen of the amide bond was capped with a methyl group to prevent spirolactam formation and the associated blinking, thereby existing always in the “on” fluorescent state. For comparison experiments, we employed commercially available dyes Alexa Fluor 546 and Alexa Fluor 568 (Thermo Fisher Scientific, Waltham, MA). Spectroscopy: Absorbance, excitation and emission spectra were acquired with aCary 4000 UV-Vis spectrophotometer (Agilent, Santa Clara, CA) and an SLM 8100 steady-state spectrofluorimeter (SLM Instruments, discontinued). Extinction coefficients were determined by measuring absorbance of a series of analytically prepared solutions using


Page 2 of 9

Page 3 of 9 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Analytical Chemistry anti-biotin antibody conjugate, in which the FRD-B3 had been conjugated to DBCO-labeled antibody via “click” chemistry. Through light images taken every 1000 frames (~100s) were used to correct image drift. Super-resolution and TIRF imaging was automated using Metamorph Advanced software (Molecular Devices, Sunnyvale, CA). The results were analyzed using ImageJ (NIH, Bethesda, MD) employing the GDSC-SMLM (University of Sussex, UK) and ThunderSTORM20 localization plugins, with additional data processing in Mathematica (Wolfram Research, Champaign, IL).

FRD-B contains a PEG-linker between the core and the terminal pentafluorophenyl active ester. In the FRD-B2 variant, the PEG-linker was replaced by an aliphatic hydrocarbon chain.

RESULTS Fluorophore Design. Xanthene dyes (fluoresceins, rhodamines and rhodols) are extensively used to study biological systems.21 Classic rhodamine synthesis produces symmetrical dyes as a mixture of two regioisomers, which leads to an undesirable heterogeneity of fluorescent conjugates with different three-dimensional configurations and quenching properties. To alleviate the need for tedious isomer separation, we utilized a different approach to synthesize a single rhodamine isomer and showed that the carboxyl in the 2’-position on the phenyl ring is much more reactive than expected, making this group suitable for conjugation.19 The synthesis of this unsymmetrical, single-isomer rhodamine core dye (FRD-core) is outlined in Figure 1A. The synthesis of the blinking secondary amide active esters of FRD-core is shown in Figure 1C and is described as follows: FRD-core was converted to the NHS ester 1 by treatment with N,N,N’,N’-tetramethyl-O-(N-succinimidyl)uronium tetrafluoroborate (TSTU). 1 was independently reacted with amino-PEG-t-butyl ester 2, aminocaproic acid 3 and N-methyl aminocaproic acid 422 providing amides 5,6 and 7, respectively. The resulting amido-t-butyl ester 5 was hydrolyzed with TFA/methylene chloride providing acid 8. Using pentaflurophenyl trifluroacetate, compounds 6, 7 and 8 were independently converted to the pentafluorphenyl active esters FRD-B, FRD-B2, and FRD-ON for subsequent conjugations. 1 was also reacted with aminoazide 9 to form FRD-B3 for “click” chemistry labeling methods. As shown in Figure 1B (top), the parent rhodamines exist in the highly fluorescent open form over a broad pH range (pH 314) in polar solvents.23 However, in nonpolar solvents such as benzene it exists as the non-fluorescent closed spirolactone form.23 The spirolactone formation disrupts the electron extended conjugated system of the xanthene chromophore and eliminates fluorescence. Unlike the parent rhodamines (Figure 1B, top), secondary amides (Figure 1B, bottom) produce non-fluorescent spirolactams under neutral and basic conditions, rendering them undesirable for conventional fluorescent studies.24-26 However, such secondary amides spontaneously and reversibly convert between the open fluorescent form and the closed non-fluorescent form, which makes them valuable for dSTORM. The spontaneous nature of this transition eliminates the need for an additional light-induced ring opening step to produce a fluorescent species.14 Our paper describes the use of a functionalized secondary amide of the FRD-core that displays spontaneous blinking behavior.

Figure 1. Dye structure and synthesis. (A) Synthesis of the single-isomer core dye (FRD-core). Overall yield (11%).19 (B) Rhodamines (top) can exist in both the closed spirolactone, nonfluorescent state or the open fluorescent state, though most of the rhodamine molecules will be found in the open form in aqueous buffers.23 When the 2’ carboxyl group is used as a conjugation point (bottom), secondary amides derived from primary amines can exist in both the closed spirolactam, non-fluorescent state or the open fluorescent state. Under neutral to basic conditions, the non-fluorescent spirolactam predominates.24-26 (C) Synthesis of the blinking dye FRD-B, its linker variant, FRD-B2, an azide version FRD-B3, and non-blinking FRD-ON. Overall yields from



FRD- core: FRD-B (38.5%), FRD-B2 (25%), FRD-B3 (77%), FRD-ON (51%). All new compounds exhibited analytical data (NMR and HRMS) consistent with their assigned structures. The yields reported are for chromatographically pure samples.

Figure 2. FRD pH sensitivity. (A) Normalized absorbance, excitation, and emissions curves are shown for FRD-core. (B) Absorbance and fluorescence spectra were recorded for FRD-B equilibrated in a series of buffers with pH values ranging from 3 to 10. (C) The change of absorbance and fluorescence amplitudes from panel B are plotted as a function of pH, showing both measurements reflect the chemical transition of a pKa of ~ 4.3. (D) Fluorescence lifetime curves measured in a series of buffers with pH ranging from 3 to 8 show a single lifetime.

This demonstrates flexibility in the blinking FRD dye family for accommodating various conjugation needs. Both dyes exhibit blinking behavior, and the following results focus on the characterization of FRD-B, as no obvious functional differences between the two variants were observed. Spectral Properties. The FRD core dye is a bright fluorophore. Its quantum yield was determined in a dilution experiment by plotting the integrated fluorescence versus absorbance measured for each known and unknown sample in the series. Using the known quantum yield of the Alexa Fluor568 as a standard ( ), the quantum yield of the FRD core ( ) was calculated from the ratio of the slopes under the assumption that both dyes are absorbing the same number of photons for equivalent ODs at a given wavelength: =

easily synthesized, well-behaved dye can be readily converted to its blinking derivative. During the initial testing, we found that the FRD core dye is pH-insensitive, so it was interesting to discover that functionalizing the carboxyphenyl group with a primary amine led to a closed-ring, non-fluorescent spirolactam exhibiting a pH-driven equilibrium between the closed and open forms. At acidic pH, the FRD-B dye exists predominantly in the open, fluorescent state. As the solution becomes neutral, the equilibrium shifts to the closed, non-fluorescent state. The most thrilling feature of FRD-B is that at pH 7.4, less than 0.5% of the molecules fluoresce at any given time as the fluorescent and non-fluorescent populations remain in dynamic equilibrium. Apparently, all dye molecules have a small probability of transitioning to the fluorescent “on” state for a short time before returning to the longer-lived dark state. In effect, the FRD dyes of this type exhibit a “blinking” behavior, which makes them useful for dSTORM. These dyes exhibit low density, inherent, stochastic fluorescent activation without adding any oxidizers/reducers, or external activators other than a fixed pH. Moreover, it is possible to functionalize the carboxyl group with a secondary—rather than a primary— amine, which prevents transition to the dark, closed state. This steadily fluorescent version (FRD-ON) of essentially the same dye can be useful for control experiments.

where is the slope and is the index of refraction of the solvent for the unknown sample and the standard. The results of this analysis are shown in Table 1 along with literature values for Alexa Fluor 546 and 568. FRD-core has a better quantum yield and thus a slightly higher brightness. This


Page 4 of 9

Page 5 of 9 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Analytical Chemistry The characteristics of the blinking dye FRD-B are shown in Figure 2, including normalized absorbance, excitation, emission (Figure 2A) and fluorescence lifetime data (Figure 2D). Dye stock, dissolved in DMSO, was diluted into a series

of buffers (see Experimental section) ranging from pH 3 to pH 10. After equilibration, spectra of each sample were taken on a spectrophotometer and fluori meter (Figure 2B). For Figure 2C,

Figure 3. Bulk transition kinetics. A pH jump study of FRD-B was conducted. The peak absorbance (570 nm) is plotted as function of time showing (A) the fast transition to the closed-state (acidic-to-basic) and (B) the slow transition to the open-state (basic-to-acidic). (C) Absorbance scans of the “opening” transition show an increase in amplitude without any shift in shape. (D) Steady-state, full spectrum absorbance scans in acidic and basic conditions reveal fairly stable UV absorbance and emphasizes the drop in fluorescence absorbance range as >99.5% convert to a true dark state.

the normalized, integrated absorbance (540-580 nm) and emission (560-700 nm) are plotted as a function of the buffer pH, showing the conversion, with pKa ~ 4.3, from fluorescent at acidic conditions to non-fluorescent state at higher pH values. At pH 7.4, less than 0.5% of the dye population is fluorescent, which complies with the low-density fluorophore conditions required for super-resolution imaging. Table 1 – Comparison of spectral properties. (*literature values)27 Dye

FRDcore

Alexa 546*

Alexa 568*

Abs (max)

560

556

572

Em (max)

583

573

603

Extinction coefficient, ε (M-1cm-1)

102,000

112,000

88,000

Quantum yield, Φ

0.88

0.79

0.69

Molecular brightness, ε∙Φ (x103)

90

88

61

While performing the pH titration of FRD-B, we observed a rapid transition (< 1 second) to the non-fluorescent state, but a slow transition (tens of minutes) back to fluorescence upon the addition of acid. To quantify this reaction, we performed stopped-flow experiments rapidly mixing buffer to change the dye environment from about pH 2.5 to pH 9 and the reverse, while monitoring the absorbance. FRD-B stock was diluted into a low molarity citric acid solution with pH 2. Using stopped flow, this solution was rapidly mixed with an equal volume of 0.1 M dibasic sodium phosphate solution, resulting in a final pH of 8.4. The sample absorbance was monitored by repeated scans. The process was mirrored, preparing the equivalent dye stock dilution into a low molarity Na2HPO4 solution (pH 9.2) which was mixed with 0.1 M citric acid to achieve a final pH of 2.1 (Figure 3C). Figure 3 shows the peak absorbance (570 nm) as a function of time during the transition from basic to acidic conditions (Figure 3A) and acidic to basic (Figure 3B). The data fit well to single-exponential models confirming earlier observations; the conversion to the dark state is quick, = , while the return to the fluorescent state upon acidification is much slower, = . . We suggest that this fluorescence recovery is slow because the lactam ring-opening mechanism requires the coincidence of two events: (i) a shift in the electron distribution, and (ii) the



dissociation of protonated aryl amines (pKa 3-5)28 which is disfavored under acidic conditions (See Figure S1). We also examined the lifetime of FRD-B as a function of pH. The data (Figure 2D) fit well to a single-species lifetime and show essentially no change as a function of pH. This confirms that the dark state at pH 8 is a non-fluorescent state, rather than a dim state with a shorter lifetime. This is further supported by the lack of shift in the absorbance spectra (Figure 3C & D). Figure 3D shows full spectrum scans of FRD-B stock diluted directly into either acidic solution (pH 2) or basic solution (pH 8.4) and allowed to equilibrate. The absorbance of the benzene rings in the far UV region only exhibits slight shifts, while the absorbance in the visible spectral region, caused by conjugated ring structures, disappears for >99% of the population. It is worth noting that for FRD-ON, which is always fluorescent since it cannot form the lactam, the absorbance spectra at pH 2 and pH 12 are identical (Figure S2). As mentioned, the FRD-B dye blinks at neutral and basic pH. However, the term “blinking” lacks specificity. Therefore, in the interests of a more quantitative definition, we performed additional analysis at the single-molecule level, using FRD-B dye conjugated to anti-biotin antibody. The labeled antibody was directly immobilized on a PEG-biotin surface for singlemolecule imaging. The incorporation ratio of FRD-B to the antibody was 0.6, leaving most antibody molecules with either zero or one FRD-B label, which is the desirable scenario for a single-molecule experiment. A 10-minute, 8000-frame time series of 50-ms single-molecule TIRF images was taken of the FRD-B/antibody-coated surface. Antibodies without an FRDB label were not visible. Antibody molecules with 2 or more FRD-B labels (

Rhodamine-Derived Fluorescent Dye with Inherent Blinking Behavior

Recommend Documents