Hydroporphyrin-Doped Near Infrared Emitting Polymer Dots for Cellular Fluorescence Imaging

Abstract

Near-infrared (NIR) fluorescent semiconductor polymer dots (Pdots) have shown great potential for fluorescence imaging due to their exceptional chemical and photophysical properties. This paper describes the synthesis of NIR emitting Pdots with great control and tunability of emission peak wavelength. The Pdots were prepared by doping poly [(9,9-dioctylfluorenyl-2,7-diyl)-alt-co-(1,4-benzo-(2,1’,3)-thiadiazole)] (PFBT), a semiconducting polymer commonly used as a host polymer in luminescent Pdots, with a series of chlorins and bacteriochlorins with varying functional groups. Chlorins and bacteriochlorins are ideal dopants due to their high hydrophobicity, which precludes their use as molecular probes in aqueous biological media but on the other hand prevents their leakage when doped into Pdots. Additionally, chlorins and bacteriochlorins have narrow deep-red to NIR emission bands and the wide array of synthetic modifications available for modifying their molecular structure enables tuning their emission predictably and systematically. Transmission electron microscopy (TEM) and dynamic light scattering (DLS) measurements show the chlorin and bacteriochlorin-doped Pdots to be nearly spherical with an average diameter of 46±12 nm. Efficient energy transfer between PFBT and the doped chlorins or bacteriochlorins decreases the PFBT donor emission to near baseline level and increases the emission of the doped dyes that serve as acceptors. The chlorin and bacteriochlorin-doped Pdots show narrow emission bands ranging from 640 nm to 820 nm depending on the doped dye. The paper demonstrates the utility of the systematic chlorin and bacteriochlorin synthesis approach by preparing Pdots of varying emission peak wavelength, utilizing them to visualize multiple targets using wide field fluorescence microscopy, binding them to secondary antibodies, and determining the binding of secondary antibody-conjugated Pdots to primary antibody-labeled receptors in plant cells. Additionally, the chlorin and bacteriochlorin doped Pdots show a blinking behavior that could enable their use in super resolution imaging methods like STORM.

Graphical Abstract

INTRODUCTION

Near Infrared Emitting Chlorins and Bacteriochlorins -

Fluorescence bioimaging methods, which often rely on the availability of fluorescent probes, are noninvasive, do not require harmful radiation, and offer high spatial resolution and temporal resolution when used to visualize cells, tissues and even whole animals.¹ Near-infrared (NIR) fluorescent probes are particularly desirable due to lower levels of light scattering with increasing excitation wavelength, reduced tissue auto fluorescence in the NIR region of the electromagnetic spectrum and increased penetration depth and signal-to-noise ratio, which enables quantitative fluorescence bioimaging of cells and tissues. Several classes of NIR dyes have been examined for in vivo imaging^2,3. Recent examples of small organic molecules used as NIR fluorophores include cyanines⁴ and related dyes⁵, BODIPY⁶ and aza-BODIPY⁷, squaraines⁸, porphyrin derivatives⁹, and benzobisb ([1,2,5] thiadiazole).¹⁰ Organic dyes have certain advantages, but they suffer from significant drawbacks, which include poor water solubility, low emission quantum yield when organic dyes are modified to render them water soluble, low photostability, and limited tunability of optical properties. In addition, optical properties of currently available fluorophores limit their applications for in vivo multicolor imaging², since many near-IR agents feature broad emission bands and relatively narrow absorption bands. There is a growing need for new bright, photostable, tunable NIR fluorophores, with optical properties suitable for bioimaging applications in biological aqueous media.

Among different classes of organic fluorophores, hydroporphyrins (chlorins and bacteriochlorins, Figure 1) appear to have particularly attractive optical properties for multiplexed cellular imaging applications. Chlorins^11,12 and bacteriochlorins¹³ are synthetic analogs of the photosynthetic pigments chlorophylls and bacteriochlorophylls. These tetrapyrrolic macrocycles differ from well-known porphyrins by having one (chlorins) or two (bacteriochlorins) partially saturated pyrroline rings. They are characterized by relatively strong absorption in the deep red (630-690 nm, ε~ 40,000 – 80,000 M⁻¹·cm⁻¹ for chlorins) or NIR (700-800 nm, ε ~ 100,000 – 120,000 M⁻¹·cm⁻¹ for bacteriochlorins). The emission spectra of both are relatively narrow (FWHM ~ 12-25 nm) with quantum yield F_f ~ 0.20 – 0.40 (chlorins) and 0.10 – 0.25 (bacteriochlorins). The absorption and emission wavelengths can be tuned across a broad spectral range by either (a) substitution at 3,13-pyrrolic positions of the macrocyclic ring with conjugated, electron withdrawing, or electron donating substituents,^13,14,15 (b) installation of an additional ring on the periphery of the macromolecule (exocyclic ring),¹⁶ or (c) assembling of the chlorins and bacteriochlorins into strongly conjugated arrays, i.e. arrays where two macrocycles are connected by a linker which provides strong π-conjugation between subunits.¹⁷

Figure 1 -.

Chemical structures of porphyrins (chlorins and bacteriochlorins) used to prepare the dye doped Pdots. The notation PXXX stands for Porphyrin and its peak emission wavelength

NIR Fluorescent Nanoparticles as an Alternative to NIR Molecular Fluorophores –

NIR fluorescent dyes generally suffer from poor water solubility¹¹ and structural modifications to improve their water solubility significantly decrease their emission quantum yield¹⁸. NIR –emitting nanoparticles, for example luminescent semiconductor quantum dots (QDs) were developed as an alternative to NIR-emitting molecular probes to overcome their limitations. Luminescent QDs offer several advantages over organic dyes including higher molar absorptivity, broad absorption, large Stokes shifts, and greater photostability¹⁹. Additionally, luminescent QDs are capable of fluorescing deeper in the IR, with reported emission up to 2000 nm²⁰. A major limitation of QDs, however, is their inherent toxicity, primarily due to (1) leaching of toxic metal ions from the QD core and (2) adverse interactions with the ligands used to stabilize and impart aqueous solubility to the QDs.^21–24 This is especially true for NIR emitting QDs, as most are synthesized using Hg and Pb based semiconductor nanomaterials.^{25, 26} While there has been advances in the development of Hg and Pb-free NIR emitting QDs^{27, 28}, these newer inorganic nanomaterials still exhibit toxicity due to leaching of heavy metal ions that continue to limit their in vivo applications²⁹. As such, attention within the field has recently shifted to the development of new nontoxic fluorescent nanoparticles.

Semiconducting polymer nanoparticles, named polymer dots (Pdots) have generated increased interest as a new type of fluorophore which could provide a viable alternative and overcome some of the major drawbacks of molecular fluorophores and luminescent QDs in the visible and NIR regions³⁰. Pdots are composed of semiconducting polymers that offer several advantages over inorganic QDs and organic dyes, including facile synthesis, high photostability, tunable absorption and emission, and exceptionally high absorptivity and brightness.³¹ On a per particle basis, fluorescent Pdots are several times brighter than similar emitting QDs and organic dyes.³⁰ This brightness is the result of fast emissive rates, large extinction coefficients, and high emission quantum yield.³² Additionally, Pdots are resistant to photobleaching³³, making them highly applicable to fluorescence imaging. Pdots can be synthesized by adding highly emitting hydrophobic polymer molecules dissolved in an organic solvent like tetrahydrofuran (THF) to aqueous solution under vigorous sonication.³⁴ Initial studies using this approach resulted in Pdots characterized with broad emission and low emission quantum yield.³⁵ Modifications of the polymer molecular structure such as the inclusion of π bridges³⁶ and donor-donor-acceptor systems³⁷ have recently improved the emission quantum yield of Pdots made using this approach. Another approach, which is also used in our study involves either doping or covalently attaching a NIR emitting dye to the Pdots.^38–42 Doping alters the emission properties of the Pdots due to energy transfer between the polymer and the doped dye molecules. The polymer molecules absorb the excitation light and transfer the energy to the dye molecules. This approach takes full advantage of the properties of both materials: the strong, broad absorption of the polymer, and the narrow NIR emission of the dye. Pdots synthesized with this method are significantly brighter on a per particle basis than the dye alone or similar emitting QDs. A key limitation of NIR-emitting Pdots is the limited availability of NIR dyes and, consequently, the truncated range of possible emission peak wavelengths of NIR-emitting Pdots. Our study successfully increases the selection of NIR emitting dyes that could be used as efficient acceptors in Pdots, which is necessary to realize the multiplexing potential of NIR emitting Pdots. This paper describes the synthesis of a series of chlorin and bacteriochlorin-doped Pdots which can all be excited by a single excitation source. Systematic modifications of the porphyrin ring result in emission wavelength tenability between 640 and 820 nm. We demonstrated the utility of these red and NIR emitting Pdots as bioimaging probes in cellular imaging applications and their potential as super resolution imaging NIR probes.

EXPERIMENTAL

Chemical Reagents –

Poly[(9,9-dioctylfluorenyl-2,7-diyl)-alt-co-(1,4-benzo-(2,1’,3)-thiadiazole)] (PFBT) (37,000 MW, 3.3 polydispersity, Poly[2-methoxy-5-(3,7-dimethyl-octyloxy)-1,4-phenylenevinylene] (ADS104) (100,000 MW, 3.3 polydispersity), Poly[{9,9-dihexyl-2,7-bis(1-cyanovinylene)fluorenylene}-alt-co- {2,5-bis(N,N’-diphenylamino)-1,4-phenylene}] (ADS11) (48,000 MW, 2.9 polydispersity), and Poly[{2-methoxy-5-(2-ethylhexyloxy)-1,4-(1-cyanovinylenephenylene)}-alt-co-{2,5-bis(N.N’-diphenylamino)-1,4-phenylene}] (40,000 MW, 3.5 polydispersity) were purchased from American Dye Source. Poly (styrene/maleic anhydride) (PSMA) [67:33] (MW 7,500) was purchased from Polysciences. Tetrahydrofuran (anhydrous, >99.8%), Phosphate Buffered Saline (PBS), and Bovine Serum Albumin (BSA) was purchased from Fisher Scientific. Dulbecco’s Modified Eagle Medium, Fetal Bovine Serum, HEPES buffer and Sodium Pyruvate were purchased from ThermoFisher Scientific. Rat endothelial cells (REC) were previously isolated in lab⁴³. Anti-Rabbit IgG (H+L), F(ab) fragment antibody produced in goat, N-(3-Dimethylaminopropyl)-N′-ethylcarbodiimide, Tris-HCl, NaCl, Driselase, MES, Sodium Azide, Tween 20 and Triton x-100 were purchased from Millipore Sigma. GST-tag antibody [HRP] produced in rabbit was purchased from GenScript. Nitrocellulose membrane and western blot chemiluminescent substrate kit were purchased from Fisher Scientific. Paraformaldehyde (PFA) was purchased from Electron Microscope Sciences. Murisage and Skoog mixture was purchased from Caisson. DYKDDDDK Tag Recombinant Rabbit monoclonal antibody was purchased from Invitrogen. All commercial materials were used as received.

Synthesis and Characterization of NIR Porphyrin Dyes –

The synthesis and characterization of red emitting chlorins P660⁴⁴, P665⁴⁵, P640,⁴⁶ and P690⁹ as well as near-IR emitting bacteriochlorins P710,^47,13 and P820¹⁷ were reported previously. The notation PXXX stands for porphyrin (P) (chlorin and bacteriochlorins are porphyrins) and the dye’s emission peak wavelength. Microwave reactions were performed in CEM Discover CEM, Mathew, NC) microwave instrument. All reactions were performed in 10 mL CEM pressurized microwave vessel, with continuous monitoring of pressure and temperature. Temperature was monitored using built-in IR sensor. All NMR spectra were acquired on 400 MHz JOEL ECX-400 NMR.

Polymer Dots (Pdots) Synthesis and Characterization –

Porphyrin-doped Pdots were synthesized via a nanoprecipitation method.³⁸ PFBT, PSMA, and the porphyrin dyes were dissolved in anhydrous THF overnight, each in separate 25 ml round bottom flasks. Solutions were stored under nitrogen at room temperature, with the dyes stored away from light. The three precursors were filtered through a 0.22 μm PTFE syringe filter, then mixed together in 1 ml THF. The concentrations of each of the precursors in the mixture were 100 ppm (PFBT), 100 ppm (PSMA), 2-20 ppm (dye) depending on the dye. The solution was then injected into 10 mL DI water under vigorous sonication at room temperature for 1 minute. THF was removed through vacuum evaporation on a rotary evaporator at 60℃. All samples were filtered through a 0.22 μm cellulose acetate syringe filter prior to further studies. The size and morphology of the synthesized Pdots were measured using a FEI Morgagni 268 100kV TEM. Samples were prepared by placing a drop of Pdot solution on a Ted Pella copper supported grid and drying at room temperature. Hydrodynamic size and surface zeta potential were measured with a DTS1070 folded capillary cell in a Malvern Zeitasizer Nano (Model No. ZEN3600) instrument. UV-Vis absorption spectra of the Pdots in aqueous solution were measured with an Aligent Cary UV-Vis multicellular Peltier spectrophotometer (Model No. G9864A). Fluorescence spectra of the dye doped Pdots were measured using a Photon Technology International fluorimeter. All samples were measured in DI water and excited at 450 nm. Corrected spectra were collected from 475 to 850 nm. Fluorescence quantum yields were measured in air-equilibrated solvents using tetraphenylporphyrin (TPP) in air-equilibrated toluene (Φ_f = 0.070)^DH as a standard. Fluorescence microscopy experiments were performed using an Olympus ix73 fluorescence microscope equipped with multiple ports. Fluorescence and brightfield images were captured with a Hamamatsu ORCA-Flash4.0 digital CMOS camera. Samples were illuminated by a CoolLED pE-300^ultra microscope illuminator. Typical exposure time was 300 ms. For fluorescence spectroscopy images, a Samrock QDLP-B-000 filter cube containing a FF01-435/40-25 excitation filter, a FF01-500/LP-25 emission filter, and a FF510-Di02-25x36 dichroic filter was utilized. Sample spectra of selected areas were captured using an Andor Shamrock SR303i spectrograph equipped with a 150L/MM-500NM grating. The spectra were acquired using a Zyla 4.2 sCMOS camera.

Conjugation of secondary antibodies to Pdots –

IgG secondary antibody conjugation was achieved through a previously reported method⁴⁸. A 1 ml aliquot of 100 μg/ml Pdot solution was mixed with 20 μl of 5% wt PEG solution (3350 MW) and 20 μl of 1 M HEPES buffer. To this mixture we added 10μl of 2 mg/ml IgG secondary antibody solution and stirred the mixture on a vortex to ensure complete mixing. Finally, 20 μl of 5 mg/ml N-(3-Dimethylaminopropyl)-N′-ethylcarbodiimide were added and the mixture was placed on a rotary shaker for 2 hours. Afterward, the mixture was transferred to a 100 kDa centrifugal filter unit alongside 10 μl of 10% wt Triton x-100 and centrifuged at 5000 RCF for 10 min to remove unbound antibody. Conjugation was confirmed by using dynamic light scattering and zeta potential measurements and Dot Blot. In Dot Blot experiments, Goat anti-Rabbit IgG secondary antibody (original 2 mg/ml, used as positive control), pure Pdots (used as negative control), and antibody conjugated Pdots were diluted indicated times and 2 μl of each sample was spotted onto the nitrocellulose membrane. After blocking non-specific sites by soaking in 5% skim milk powder dissolved in TBST buffer (20 mM Tris-HCl pH 7.5, 150 mM NaCl, and 0.1% Tween 20) for 0.5-1 hr, the nitrocellulose membrane was incubated with primary antibody conjugated with HRP (GST-tag antibody [HRP] produced in rabbit) for 1 hr at RT. Afterward, the nitrocellulose membrane was washed three times with TBST buffer (3 x 5 min) and incubated with western blot chemiluminescent substrate for 1 min, and then was used to expose X-ray films in the dark room.

Cellular Imaging and Toxicity Studies –

Rat endothelial cells cultured in Dulbecco’s Modified Eagle Medium (DMEM) High Glucose supplemented with 10% FBS, 1% HEPES and 1% Sodium pyruvate were used in our cellular imaging studies. Cell viability measurements were carried out by seeding ~50,000 cells in each well of a 24-well plate and allowing the cells to grow for 24 hours. The cells were then washed twice with phenol red and serum free DMEM. Pdots (~ 10 μg/ml and 50 μg/ml) were then added to the cells after being filtered and diluted in phenol red free and serum free DMEM and incubated for 1 hour. Calcein AM and ethidium homodimer were then added to the cells at concentrations of 4 μM and 2 μM respectively. For the negative control, cells were incubated with 70% ethanol to ensure complete cell death. The medium was removed from the well before imaging so the excess Pdots in solution do not interfere with the fluorescing cells, but the wells were not washed so as not to remove any dead cells. The cells were then imaged and processed using the Cytation 5 fluorescence microscope with Gen5 imaging software (Biotek, USA). Cell quantification was performed using imageJ, counting with the multipoint tool. Cellular imaging and spectroscopy measurements with Pdots were carried out by seeding about 100,000 cells in each well and allowing the cells to grow for 24 hours. The cells were then washed twice with phenol red and serum free DMEM. Single type or mixed pdots were added to the wells and incubated with the cells for 30 min at 37ºC. The cells were then washed twice with phenol red and serum free DMEM to remove excess pdots not taken up by the cells. Cells were then imaged using the Olympus ix73 fluorescence microscopy system described above.

Construction of FLAG-FLS2 transgenic plants –

The FLAG sequence-fused FLS2 genomic fragment was generated by a two-fragment polymerase chain reaction (PCR) approach, in which primers including the FLAG sequence were used in two separate PCR reactions. Each fragment of FLS2 was amplified by PCR from genomic DNA of wild-type plants using gene specific primers containing the FLAG sequence. An overlapping PCR reaction was then performed to generate the whole FLS2 genomic fragment with the FLAG sequence inserted behind the signal peptide sequence of FLS2. The PCR products were cloned into pDONR-Zeo (Invitrogen) by BP reaction and subsequently cloned into the binary vector pGWB1 (Invitrogen) by LR reaction. The construct was then transformed into A. thaliana fls2 knockout mutant plants via the Agrobacterium strain GV3101 using the ‘floral-dip’ method⁴⁹ The fls2 knockout mutant was obtained from the Arabidopsis Biological Resource Center (ABRC, Ohio State University, Columbus, OH). The transgenic plants were selected by germination on 35 μg/mL hygromycin-containing 1/2 MS medium. Homozygous T3 generation plants were used in this study. The functionality of the FLAG tagged-FLS2 receptor was confirmed by its ability to complement the phenotype of the fls2 knockout mutant.

Fluorescence imaging of antibody conjugated Pdots when targeting the FLS2 receptor in plant cells –

Arabidopsis thaliana seeds, expressing FLAG-tagged FLS2 (FLAG-FLS2) at the N terminus, were grown on 0.5% phytagel with Murisage and Skoog media for 10-14 days with photocycles of 16 hours of light followed by 8 hours of darkness. Plants were removed from the medium and fixed in 4% Paraformaldehyde for 1 h under vacuum before being stored in PBS. Prior to immunostaining, a scalpel was used to slice the leaves to allow some chlorophyll release for clearer imaging. The plant cell wall was partially dissolved by 0.2% driselase in 2 mM MES at pH=5.7 at 37^oC for 15 minutes. The membrane was then permeabilized with 1% Triton X-100 in PBS for 20 minutes at room temperature. Plants were then soaked on blocking buffer (1% BSA, 0.1% tween, 2 mM sodium azide in PBS) at room temperature for 20 minutes. Primary Rabbit anti DYKDDDDK-tag (anti-FLAG) antibodies were diluted to 1:250 in blocking buffer, and plants were incubated with primary antibodies overnight at 4 ^oC. Plants were rinsed in blocking buffer three times for 20 minutes at room temperatures. Pdots conjugated to the secondary antibody (goat anti rabbit IgG) were then diluted to 30 μg/ml in blocking buffer and incubated for 3 hours at room temperature. Plants were rinsed in PBS three times for 10 minutes at room temperature.

Leaves and root hairs were removed from the plant and placed in a glass-bottom dish with 10 μL of PBS under a 10 cm glass coverslip. These were imaged using an inverted fluorescence microscope (Olympus IX 73). These were illuminated using a 440 nm solid state laser (Crystalaser) at 0.1 mW. Samples were observed using a 60× magnification, water objective lens (NA 1.2, Olympus UPlanSApo). Images were collected through a 647 nm band pass filter with 57 nm width and 710 nm band pass filter with 40 nm width. Images were captured by an EMCCD camera (Princeton Instrument Photonmax) at 0.3 seconds exposure time.

Fluorescence imaging of single Pdots –

Pdots were diluted in water, placed on a glass coverslip and imaged using the inverted fluorescence microscope and laser described in the previous section. Pdot samples were observed using a 100x magnification, oil immersion objective lens (NA 1.4, Olympus UPlanSApo). The laser illumination covered an area with a diameter of ~80 μm. Emission was acquired using a band pass filter centered at 647 nm with 57 nm width.

RESULTS AND DISCUSSION

Chlorin and Bacteriochlorin, Molecular Design, Synthesis and Characterization –

Chlorins and bacteriochlorins are uniquely characterized with narrow emission bands compared to other NIR emitting dyes with the full-width-at-half-maximum (FWHM) ~ 15 nm for chlorins and ~ 20 nm for bacteriochlorins.^{11–13, 50}In addition, unlike with other NIR emitting dyes, it is possible to position and tune the emission band of chlorins and bacteriochlorins with nearly nanometer precision.^{11–13, 50} Tuning the absorption band of chlorins and bacteriochlorins is realized by changing their molecular structure, specifically by substituting functional groups at the periphery of their macrocycles. Conjugated or electron withdrawing/electron donating substituents affect the HOMO-LUMO energies and consequently S₀ → S₁ transitions energies.^{11–13, 50}A second way to tune the emission wavelength of chlorins and bacteriochlorins involves arranging their macrocycles into conjugated arrays. Conjugation of the macrocycles delocalizes molecular orbitals over adjacent macrocycles, thus increases the size of the π-system. This ultimately leads to a reduction of the S₀ → S₁ energy bandgap.¹⁵ In the current study, we prepared NIR emitting Pdots using a series of hydrophobic chlorins and bacteriochlorins we previously synthesized (Figure 1), all share a common hydroporphyrin macrocyclic structure. The chlorin and bacteriochlorin dyes differ in their peripheral substituents, or they are arranged into strongly conjugated arrays. The set includes a simple chlorin P640, which possesses the shortest emission wavelength, styrene-substituted P660, and a conjugated dyad 690, for which the emission wavelength is progressively bathochromically shifted. Similarly, we also used a bacteriochlorin series which contains a simple bacteriochlorin P720, chalcone-substituted P790, and a conjugated dyad P820. Chalcone substituent,⁴⁴ which is both conjugated and electron-withdrawing significantly shifts the absorption of P790 compared to P720. The strongly-conjugated dyads P690 and P820 are the most-bathochromically-shifted derivatives in both the chlorin and the bacteriochlorin series. As described in detail in the Experimental section, substituted chlorins and bacteriochlorins were prepared using palladium-catalyzed cross coupling reactions, starting from the corresponding monobromo-chlorins and dibromo-bacteriochlorins respectively.^9,34–36 For example, Scheme S1 shows the synthesis of P790 using a microwave-assisted Heck reaction^44,51 of a known bacteriochlorin-Br₂⁴⁷ and methyl vinyl ketone with a 46% yield. Structures of the six porphyrin dyes used in this study are shown in Figure 1. The full synthetic procedure to prepare P790, its NMR characterization, as well as UV-Vis, and fluorescence spectra for all dyes are shown in the supporting information document (Figures S1–S6).

Polymer Selection for NIR Emitting Pdots –

An appropriate selection of a semiconducting polymer for fluorescent Pdots is imperative for their optical properties. In our NIR-emitting Pdots, the Pdots are excited at the peak absorption maximum of the polymer molecules. The emission maximum is obtained at the emission peak wavelength of the doped NIR dye. High absorption cross and at the excitation wavelength and efficient energy transfer between the polymer and dye molecules are imperative for high NIR emission of the Pdots. Additionally, the emission band of the polymer should minimally bleed into the emission peak of the doped dye to minimize spectral interference between the polymer and dye molecules in imaging experiments. Figure 2 shows the normalized UV-Vis and fluorescence spectra of four semiconductor fluorescent polymers PFBT (Figure 2a), ADS104 (Figure 2b), ADS 111 (Figure 2c), and ADS114 (Figure 2d). The three polymers from the ADS series have been used previously in Pdots⁴⁷ but they show lower emission than PFBT and more importantly, their emission peak tail bleed into the NIR region. PFBT was chosen as the polymer for our NIR-emitting Pdots because of its high absorption cross section and sufficient overlap of its fluorescence peak with the absorption peaks of our porphyrin dyes, conveniently excitable absorbance peak, and most importantly, its minimal overlap of the emission peak red tail with the dyes’ emission. Using PFBT as the polymer component of the Pdots has enabled the formation of Pdots with emission peak wavelength ranging from 640 nm to 820 nm. It is possible to further tune Pdots’ emission by using blue shifted emitting polymers for visible emitting Pdots or by using red shifted emitting polymers like ADS104, ADS 111, and ADS 113 for Pdots that emit even deeper into the NIR region.

Figure 2 -.

Absorption (black) and fluorescence (red) spectra of (a) PFBT, (b) ADS104, (c) ADS111, and (d) ADS113.

Synthesis and Characterization of NIR Emitting Pdots –

Chlorin and bacteriochlorin dyes used in our study are hydrophobic and cannot be used as molecular probes in aqueous media without incorporating them into Pdots. The synthesis of chlorins and bacteriochlorins-doped Pdots is shown in scheme 1. As described in the experimental section, the NIR emitting Pdots are prepared by co-injecting PFBT, PSMA and fluorescent dye solutions all dissolved in THF into aqueous solution under vigorous sonication following a previously described Pdots’ synthesis protocol.³⁸ Large aggregates which form during the short 1-minute-long reaction are removed from the Pdots’ solution by filtration.

Scheme 1 -.

Preparation of dye doped NIR emitting polymer dots (Pdots) (shown here with P720).

The structural properties of the formed Pdots are shown in Figure 3. A representative TEM image (Figure 3a) and size distribution (Figure 3b) of the formed Pdots show near spherical morphology with some disorder in structure and average diameter of 46±12 nm. While the obtained Pdots are fairly polydisperse, their emission peak wavelength is independent of their size. As expected, DLS measurements show a slightly larger average hydrodynamic diameter of 52 nm (Figure 3c). The surface zeta potential, shown in Figure S7 of the synthesized Pdots was −37 mV; indicating the surface is functionalized with negatively charged carboxyl groups.

Figure 3 -.

(a) Size and morphological characterization of PFBT polymer dots (loaded with P640) by TEM. The representative TEM image shows the near spherical Pdots to average 46±12 nm in diameter with some disorder. (b) Histogram of Pdots size distribution from TEM images (n=69) shows that the Pdots to be somewhat poly dispersed and the formation of larger aggregates. (c) Dynamic light scattering (DLS) measurements show the Pdots hydrodynamic diameter to average 52 nm which is in close agreement with the TEM results.

Pdots NIR Emission properties –

Due to the spectral overlap between the PFBT emission and porphyrin absorption, and the close proximity of the dyes to the polymer when doped into a Pdot, we predicted a highly efficient energy transfer from the polymer to the dye upon excitation of the porphyrin–doped PFBT Pdots at 450 nm. As expected, doping PFBT Pdots with optimal dye concentration resulted in complete quenching of PFBT fluorescence and in an increase in NIR emission from the dopant dyes. Figure 4 shows normalized fluorescence spectra of all six NIR-emitting Pdots. The formed porphyrin doped Pdots feature narrow emission peaks in the NIR range with an average full width at half maximum (FWHM) of 25 nm. Table 1 summarizes the emission peak wavelength and emission quantum yield of the porphyrin dyes (termed PXXX) and porphyrin-doped Pdots (termed PPDXXX). An emission red shift is observed when the dyes are doped into the Pdots most likely due to changes in their chemical environment. The magnitude of the red shift is larger at higher emission peak wavelengths. Consistent with previous studies³⁸, the emission quantum yields of the dye-doped Pdots are lower than the emission quantum yield of dye-free PFBT Pdots but the emission peaks are significantly narrower and in the NIR region. The emission quantum yields of the formed NIR-emitting Pdots range from 0.1 to 0.5 with no apparent trend. The amount of dye dopant in the Pdots was optimized by varying the dopant concentration in the Pdots preparation solution to maximize their fluorescence intensity. The NIR emission intensity of the Pdots increases with increased dye concentration up to a certain optimum level. Within this concentration range the polymer emission is drastically reduced, and the dye emission increases. Above a certain doping percentage, the dye emission begins to decrease due to self-quenching. For example, for PPD640 a P640 concentration in the reaction mixture of 6.9% resulted in the highest emission, and the NIR emission decreased above this level (data not shown). The PFBT native emission also increases when the excitation intensity is above the saturation level of the dye-doped Pdots. An example for the saturation effect is shown in Figure 3 for PPD790 (purple curve). While energy transfer between PFBT and P790 molecules results in NIR emission at 790 nm, the emission of PFBT is only partially quenched even at a P790 level of 9.1%, the level required to maximize the PPD790 NIR emission. This may be attributed to molecular interactions that inhibit energy transfer from the polymer to the dye molecules that need to be further studied. To prevent dye fluorescence quenching and saturation we set the dye loading percentage to an optimal level of 5% in all the Pdots used in this study.

Figure 4 -.

Normalized emission spectra of porphyrin doped polymer dots (λ_ex = 450 nm).

Table 1 -.

Peak emission wavelengths and fluorescence quantum yields of porphyrin dyes in THF and PPDs in water.

Sample	λem Dye (nm)	λem PPD (nm)	QY Dye	QY PPD
P640	637	641	0.29±0.01	0.11±0.01
P660	657	662	0.52±0.02	0.49±0.01
P690	686	692	0.35±0.01	0.21±0.01
P720	711	718	0.31±0.01	0.20±0.01
P790	778	794	0.25±0.01	0.20±0.01
P820	800	821	0.32±0.01	0.15±0.01

Cellular Imaging and Toxicity Studies –

A common concern associated with the use of fluorescent nanoparticles, in our study NIR emitting Pdots, is whether they by themselves affect cell viability during incubation. Cell viability stain assays using Calcein AM and ethidium homodimer-1 were carried out to determine whether our NIR-emitting Pdots adversely impact cell viability upon exposure and incubation. Calcein AM, a green fluorescent dye, is readily endocytosed by live cells; while ethidium homodimer, a red fluorescent dye, is only capable of penetrating dead cells. Fluorescence images of cells labeled with our Pdots compared to a positive control are shown in Figure 5. Figure 5 a–c shows that cells incubated for 1 hour with PPD640 concentration 0μg/mL, 10μg/mL, and 50μg/mL respectively had almost no fluorescence from the homodimer; indicating that most of the cells remained viable including at high concentrations of Pdots. Cell viability assays conducted following incubating the cells with NIR-emitting for 24 hours indicate that the Pdots do not adversely impact the cells when used as cellular imaging probes in our imaging studies.

Figure 5 -.

Live/Dead of RECs incubated for 1hr at 37° C in the presence of P640-loaded Pdots at (a) 0 μg/mL, (b) 10 μg/mL, (c) 50 μg/mL (Scale bar 200 μm). Calcein-AM (green) stain is used to stain live cells and ethidium homodimer-1 (red) is used to stain for dead cells. (d) Survival percentage of RECs with Pdots at different concentrations.

To demonstrate the multiplexing capabilities of our dye doped Pdots, we incubated rat endothelial cells with a mixture of Pdots of varying emission wavelengths. Endothelial cells have been shown to readily endocytose nanoparticles of similar dimensions to our Pdots^{52, 53}. Bright field and fluorescence microscopy images of cells incubated with PPD640, PPD660, and PPD690 Pdots are shown in Figure 6. Figure 6a and 6b show bright field and fluorescence images of cells labeled with one type of PPDs, PPD640. The bright field image in Figure 6a shows that the PPD-labeled cells maintain their structural integrity. The fluorescence image in Figure 6b shows some of the PPDs attached to the cells’ membranes while others may have permeated into the cells. The insert in Figure 6B shows the fluorescence spectra of sub-section of the fluorescence image with high Pdot density. As expected, it shows a single fluorescence peak at 640 nm. Figures 6c and 6d show bright field and fluorescence images of cells labeled with two types of NIR-emitting Pdots, PPD640 and PPD660. The insert in Figure 6d shows the presence of the two types of PPDs in the cells. Figures 6d and 6e show bright field and fluorescence images of cells labeled with three types of NIR-emitting Pdots, PPD640, PPD660 and PPD690. The insert of Figure 6f shows the presence of the three types of PPDs in the cells. While a baseline separation of emission peaks is not realized over the narrow emission wavelength range of about 70n nm used in these measurements, our results clearly indicate that due to their narrow emission peaks our NIR-emitting Pdots could be used for multiplex cellular fluorescence imaging of multiple targets either in the cell or on the cell membrane.

Figure 6 -.

Bright field and fluorescence images of rat endothelial cells incubated with (a,b) one type of NIR-emitting Pdots PPD640, (c,d) two types of NIR-emitting Pdots PPD640 and PPD660, and (e,f) three types of NIR-emitting Pdots PPD640, PPD660, and PPD690. Inserts show the fluorescence spectra of sub-sections of the image with high Pdot density.

Fluorescence Imaging of Targeted Receptors in Plant Cells –

NIR spectrum fluorescent dyes are a promising avenue for bioimaging in plants, as the presence of chlorophyll limits the types of fluorescent dyes available within the visible spectrum. The emission spectrum for Chlorophyll-b includes a peak from ~600-670 nm, thus NIR dyes that emit at longer wavelengths may not be obscured by a strong chlorophyll background, making them ideal candidates for fluorescence imaging in leaves. In order to demonstrate the potential for the use of NIR pdots for bioimaging in plants, we first conjugated Anti-Rabbit IgG (H+L), F(ab) fragment antibody molecules to P690 containing Pdots (PPD690). The conjugation of antibody molecules to the Pdots was confirmed by dynamic light scattering (DLS) which showed an increase in hydrodynamic diameter from 46±12 nm to 70±10 nm and zeta potential from −47.8 to −34.5 mV. The activity of the antibody-Pdots conjugates was confirmed using Dot Blots, which showed a clear recognition of the primary antibody produced in rabbit by the Anti-Rabbit IgG (H+L), F(ab) fragment antibody-conjugated Pdots. We then stained the transmembrane receptor kinase FLAGELLIN SENSITIVE2 (FLS2) in both leaf cells and root hair cells of Arabidopsis thaliana. Both cell types express the FLS2 receptors, which plays a key role in recognizing the bacterial flagellin protein and, hence, generating a defense reaction to the invading bacterial pathogen.^54,55 The Arabidopsis thaliana plants were genetically engineered to express a DYKDDDDK peptide, known as FLAG, tag at the FLS2 receptor N terminal. Functionality of this transgenic receptor construct was confirmed by showing that it restored flagellin recognition to A. thaliana fls2 mutant plants (see methods).

To demonstrate the utility of our Anti-Rabbit IgG (H+L), F(ab) fragment antibody-conjugated Pdots, plants expressing FLAG-tagged FLS2 receptor were stained with a DYKDDDDK Tag Recombinant Rabbit Monoclonal Antibody. Plants were then rinsed and stained with Anti-Rabbit IgG (H+L), F(ab) fragment antibody-conjugated PPD 690 Pdots. This staining was performed alongside a wild-type control plant lacking the FLAG-tagged FLS2 receptor. We imaged these plants on an inverted fluorescence microscope illuminated by a 440 nm laser in both 647 nm and 710 nm channels. To decrease chlorophyll interference in the leaf, small cuts were made in the leaves to allow release of some chlorophyll, which enabled us to resolve fluorescent Pdots along the leaf cell walls in both channels (Figure 7A, left). In contrast, no specific staining was observed in the wild type (Figure 7A, right). The same results occurred when imaging root hair cells, showing specific fluorescence signals in both channels in the FLAG-tagged FLS2 plants but not in the wild-type plants (Figure 7B). These results demonstrate that the antibody-coated PPD690 Pdots were able to specifically bind to FLAG-tagged receptors within fixed plant tissues. While ideally NIR Pdots in longer wavelengths will be used to circumvent the presence of chlorophyll within leaves by emitting at wavelengths beyond the chlorophyll emission spectrum, these results demonstrate as a proof-of-concept that antibody conjugated Pdots within our experimental system are able to specifically recognize their target in plant tissues.

Figure 7 -.

Examples of fluorescence images showing leaf cells (A) and root hair cells (B) taken from Arabidopsis thaliana expressing Flagellin Sensitive2 (FLS2) receptor kinase tagged with FLAG at the N-terminal (left panel) and from a wildtype plant expressing FLS2 with no tag serving as negative control (right panel). The receptor is labeled using a primary antibody against FLAG, followed by PPD690 (Pdots encapsulating P690 dye), decorated with a secondary antibody against the primary antibody. Images were taken in two channels to capture the emissions of the polymer (647±57 nm, upper) and the emission of the dye (710±40 nm, lower). Scale bars represent 20 μm.

Fluorescence Blinking of Pdots –

While NIR fluorescence imaging microscopy is advantageous due to reduced autofluorescence background, its spatial resolution is diffraction limited. Therefore, the increase in emission wavelength decreases the spatial resolution of microscopy experiments compared to fluorescence microscopy measurements in the visible region. To overcome this spatial resolution limitation, we conducted single particle imaging studies to determine whether our 50 nm NIR-emitting Pdots could be used as super resolution imaging probes. When an aqueous solution of PPD690 Pdots were diluted and imaged with high magnification, it was possible to capture emissions of individual Pdots as they went through on-off cycles, or blinking. This blinking behavior is demonstrated in Figure 8, where consecutive images are selected from a series, available under Supplementary Information (Video 1). Such a blinking property makes these Pdots valuable for single molecule-based super resolution fluorescence imaging, such as STORM⁵⁶ or PALM⁵⁷, where photo-switching fluorophores are required. Together with the observation that these Pdots are not toxic to living cells, their ability to blink makes them valuable for biological imaging with nm spatial resolution.

Figure 8 -.

Demonstration of Pdot blinking. Consecutive images of PPD640 Pdots selected from a series shown in the video under Supplemental Information. The Pdots were diluted and imaged on a glass coverslip using 440 nm excitation wavelength and ~600–700 nm emission band. The arrows point to two Pdots as an example for their blinking through the series.

Summary and Conclusions

This paper describes the synthesis of deep red and NIR emitting Pdots that were prepared by doping Pdots with a series of chlorin and bacteriochlorin dyes with systematically tuned molecular structures. The resulting Pdots are excited by a single excitation source, are highly emitting with distinct narrow emission peaks and high emission quantum yield over a broad range of emission wavelengths. Doping the Pdots with chlorins and bacteriochlorins minimally alters the emission properties of the free dyes and enables their use in aqueous biological buffers. It is not possible to use chlorins and bacteriochlorins in a free form in aqueous media because of their low aqueous solubility. The study reveals that the resulting Pdots do not show measurable cytotoxicity under our experimental conditions of a relatively short Pdots-cells exposure of one hour. We demonstrated the utility of the porphyrin-doped Pdots in multiplexed cellular imaging. We also show that when labeled with secondary antibodies, the NIR emitting Pdots can be used effectively for targeted imaging of receptors in plant cells. NIR imaging is especially important in plant cells where autofluorescence of chlorophyll and other cellular constituents in the visible range of the electromagnetic spectrum is often a limiting factor. Finally, the data show that the porphyrin-doped NIR emitting Pdots photoblink and therefore could potentially be used as super resolution cellular imaging probes. However, our study also reveals that while chlorin and bacteriochlorin-doped Pdots are sufficiently photostable and could be used effectively in wide field fluorescence imaging applications, their limited photostability hinders their use in super resolution imaging studies. Currently we are exploring a new series of dyes with greatly improved photostability for preparing Pdots for super resolution imaging applications.

The supporting information document contains detailed information about the synthesis and characterization of bacteriochlorin P790 (scheme S1 and section 1.1), the absorption and emission spectra of the chlorin and bacteriochlorin dyes used in the study (figures S1–S6 in section 1.2), and zeta potential of the Pdots (figure S7 in section 1.3). A video file is attached as a second supporting information file.