Sungmoon
Choi
a,
Junhua
Yu†
a,
Sandeep A.
Patel
a,
Yih-Ling
Tzeng
*b and
Robert M.
Dickson
*a
aSchool of Chemistry and Biochemistry and Parker H. Petit Institute for Bioengineering and Bioscience, Georgia Institute of Technology, 901 Atlantic Drive, Atlanta, GA 30332-0400, USA. E-mail: dickson@chemistry.gatech.edu
bDepartment of Medicine, Emory University School of Medicine, Atlanta, GA 30322, USA. E-mail: ytzeng@emory.edu
First published on 9th November 2010
Through tailored oligonucleotide scaffolds, Ag nanocluster syntheses have yielded thermally and cell-culture medium stable silver cluster-based emitters. Optimizing ssDNA stability has enabled creation of highly concentrated and spectrally pure nanocluster emitters with strong intracellular emission. Both fixed and live-cell staining become possible, and intracellular delivery is demonstrated both through conjugation to cell-penetrating peptides and via microinjection.
Silver nanodots are generally synthesized by reduction of a mixture of ligand and silver nitrate by sodium borohydride.1 The optimal nucleobase : silver ratio to yield maximized emission is usually found to be approximately two bases per silver.1,2,14,16,20–22 For a given sequence, however, variations in this ratio can bias fluorophore creation toward one of multiple emitters formed in a given sequence. Gwinn et al., for example, nicely illustrated this concept using DNA hairpins of varying length, but at constant Ag concentration, thereby yielding mixtures of multiple Ag nanodot emitters.4,16 Although recent successes at creating spectrally pure emitters have been achieved,2,23 most studies continue to result in low concentration mixtures of green and red emitters. Further, the extreme insolubility of Ag salts in biological media, coupled with poor spectral purity and low overall obtainable nanodot concentrations (i.e. sub-micromolar)4,16,21,22,24,25 has precluded more widespread use of these otherwise promising fluorophores. Herein, through facile optimization of creation conditions, we create and utilize a wide range of multicolor fluorophores with excellent chemical stability and spectral purity, at sufficiently high concentrations for use in complex biological environments. These emitters can be delivered into living cells either by microinjection or through conjugation to a cell penetrating peptide (CPP),26,27 thereby opening paths toward intracellular labeling.
DNA sequence | T m/°C | λExc/λEm/nm | Lifetime/ns | Φ (%) | Ε/105 M−1 cm−1 | Half life time in water at 20 °C (hours) | Half life time in PBS at 37 °C (hours) |
---|---|---|---|---|---|---|---|
C20 | n/a | 467/523 | 3.6 ± 0.1 | 27 ± 2 | 1.2 ± 0.3 | >80 | 20 |
AATTC12 | 9 | 480/562 | 4.3 ± 0.1 | 38 ± 3 | 2.0 ± 0.4 | 6 | <1 |
CGAAC12 | 32 | 543/590 | 2.0 ± 0.1 | 36 ± 3 | 3.5 ± 0.7 | 7 | <1 |
CGCGC12 | 57 | 560/615 | 2.5 ± 0.1 | 42 ± 3 | 9.5 ± 1.9 | 19 | <1 |
CGAACGCGC12 | 72 | 560/618 | 2.3 ± 0.1 | 16 ± 2 | 2.8 ± 0.6 | 40 | 6 |
ATATC8 | 3 | 570/635 | 2.7 ± 0.1 | 26 ± 3 | 3.2 ± 0.7 | 5 | <1 |
GGGGC8 | 64 | 600/670 | 3.5 ± 0.1 | 34 ± 3 | 2.5 ± 0.5 | >40 | <1 |
Fig. 1 Emission and multicolor staining of various Ag nanodots. A, Normalized emission spectra of various emitters, from left to right: 523 nm (green C20), 562 nm (AATTC12), 590 nm (CGAAC12), 615 nm (CGCGC12), 635 nm (ATATC8), 670 nm (GGGGC8). (Inset) Different color Ag emitters in cuvettes, each excited near their excitation maximum. B, Emission intensity comparison of Ag nanodots between sequence with a hairpin structure (CGCGC12: 5′-CGCGCCCCCCCCCCCCCGCG-3′) and that with a simple polycytosine structure (C12: 5′-CCCCCCCCCCCC-3′), each at their highest intensities. C, Typical absorption spectrum of concentrated 615 nm emitter (1 cm light path). D, Thermal stability of CGAACGCGC12CGAACGCG-protected Ag nanodots compared to ATATC12-protected Ag nanodots. E, Excitation and emission spectra of aqueous penetratin-C12 silver clusters solution. Emission (solid) was excited at 575 nm, and excitation (dotted) was detected at 640 nm. F, MALDI mass spectrum of penetratin-C12 conjugate. |
Although a promising green emitter is obtained using C20, sequence modifications are even more crucial than base : Ag ratio for tuning Ag nanodot emission color and spectral purity. Fine-tuning of creation conditions, coupled with optimizing the DNA hairpin sequence and thermal stability yield significant improvements in nanodot properties that enable intracellular use. Building on others’ successful hairpin studies,4 adding four complementary bases at the 3′ end of oligocytosine to yield CGCGC12 (5′-CGCGCCCCCCCCCCCCCGCG-3′) forms a stem-loop structure in which direct nanodot synthesis produces a spectrally pure 615 nm emitter at a loop base : Ag+ ratio of 1.8, yielding a ΦF nearly twice that of C12 nanodots (42% vs. 23%), and improved overall emission intensity relative to that obtainable in C12 (5′-CCCCCCCCCCCC-3′) scaffolds (Fig. 1B). Further, in contrast to C12-encapsulated nanodots, the improved stability enables these emitters to be directly concentrated through either solvent evaporation or ethanol precipitation without destroying their stability or spectral purity. Such concentration was not possible with previous emitters that report the μM-encapsulating DNA concentration, but have very low nanodot yields.22 These improved methods yield solutions with optical densities well above 3 (Fig. 1C), indicating nanodot concentrations in excess of 10 μM. Analogous sequences with terminal hairpin structures, such as AATTC12, 5′-CGAACCCCCCCCCCCCTTCG-3′ (CGAAC12), ATATC8, and 5′-GGGGCCCCCCCCCCCC-3′ (GGGGC8), also produce various spectrally pure, concentratable Ag nanodot solutions with very bright, spectrally pure emission with a ratio versus other observed emission of more than 100 and much higher concentrations than reported with any other hairpin or non-hairpin equivalent (Table 1). Interestingly, mismatched stem pairs (e.g. 5′-CGGCC12CGCG-3′) fail to produce Ag nanodots at high concentrations (not shown), suggesting that the hairpin structure is beneficial to Ag nanodot formation,4 as also suggested with Ag-induced hairpin formation within polycytosine.29 Tuning the hairpin structure yields multiple spectrally pure emitters with emission wavelengths varying from the yellow to the red (Fig. 1A). Each of these emitters exhibits excellent photophysics (Table 1) with photostabilities and sustained emission rates improving on the best available organic fluorophores.30 As we report here, maintaining the oligocytosine loop, while increasing thermal stability of the stem and optimizing creation conditions has yielded buffer- and culture medium-stable emitters that maintain their strong emission for many hours at 37 °C (Fig. 1D). With sub-3-nm hydrodynamic radii (data not shown) dominated by the encapsulating DNA strands, all emitters are likely composed of fewer than ∼10 Ag atoms,31 although chemical analysis of these hairpins remains elusive. However, no nanoparticle formation is observed in any of these samples either with darkfield imaging or centrifugation at 23000 × g. Although concentration enables highly absorbing, spectrally pure (>100-fold higher than any other emitter) solutions to be obtained, chemical synthetic yields for these emitters only range from 1% to 5%, depending on the sequences. Clearly higher yield synthetic methods will further improve imaging applications.
Even without hairpin-based stabilization, the spectrally pure, green-emitting C20:Ag nanodots selectively stain the nuclei of fixed NIH3T3 cells (Fig. 2) and show excellent photostability under both one- and two-photon excitation. These cells were co-stained with HCS red cell stain (Invitrogen), which stains the cytoplasm, as shown in red. Under confocal scanning, both labels show excellent one-photon-excited stability, with less than 10% fluorescence intensity decay of either dye after 2 h of continuous imaging (Fig. 2A, C). However, under two-photon-excitation (TPE), the fluorescence intensity of HCS red decays to 60% after just 10 min of scanning, while Ag nanodots retain 90% fluorescence intensity after scanning for 1 h (Fig. 2B, D). The in vitro one- and two-photon excited nanodot emission spectra, excited at 458 nm (OPE) or at 720 nm (TPE), respectively, are indistinguishable (not shown).
Fig. 2 Confocal fluorescence imaging and photo-stability of C20 Ag nanodots (green) in NIH 3T3 cells co-labeled with HCS red cell stain (red) under one photon excitation (OPE, 458 nm, A) or two photon excitation (TPE, 720 nm, B). The photostability can be easily seen by the color decay of images in nuclei (green C20 Ag nanodots) and cytoplasm (HCS red cell stain), and illustrated in C (for OPE) and D (for TPE) by their mean fluorescence intensity of individual emitters in the corresponding scanning images. The green Ag nanodots show similar photostability to HCS red cell stain under OPE, but much greater stability under TPE. E and F, Red and green channels from fixed NIH 3T3 cells co-stained with anti-OxoPhos/ATATC8 Ag nanodots (red) and anti-Heparin Sulfate/AATTC12 Ag nanodots (green). Scale bar, 30 μm. |
Upon conjugation to an antibody, Ag nanodots do not significantly impair specificity. AATTC12 was conjugated to anti-HS to mark the cell membrane, and red emitting ATATC8 was conjugated to anti-OP to mark the mitochondria. Methods analogous to those previously reported were followed. Upon conjugation and nanocluster formation in AATTC12, emission is blue-shifted by 40 nm to 522 nm, but the conjugated ATATC8 Ag nanodot emission is unchanged. As shown in Fig. 2E & F, methanol-fixed NIH3T3 cells treated with anti-OP/ATATC8 Ag nanodots (5 μM, concentration of DNA ligand) and anti-HS/AATTC12 Ag nanodots (5 μM, concentration based on DNA ligand) stained mitochondria (red) and the cell membrane (green), respectively. The strong staining of nuclei observed with unconjugated green nanodots was circumvented with antibody conjugation.
Ag salts are notoriously insoluble, but all of these Ag nanodot emitters are highly buffer stable. Prepared directly in PBS, the green C20 nanodots show excellent chemical stability in both PBS and in DMEM cell growth medium under ambient conditions. This C20-encapsulated green emitter is actually preferentially stabilized in PBS. The combination of thermal and buffer stability suggest that the DNA performs a significant protecting role in this nanodot emitter, as other C20-encapsulated emitters are destabilized under these buffer conditions, leading to spectral shifts and eventual cluster decay.22 The balance between improved DNA stabilization at high ionic strength,32–34 and the insolubility of most Ag salts preferentially forms this green emitter in a variety of buffers composed from sodium bicarbonate, sodium acetate, or sodium perchlorate. The stability decreases somewhat with increased temperature, retaining half of its emission in DMEM after 4 h at 60 °C. We should note that shorter oligocytosine-protected Ag nanodots (e.g.C12) possess similar thermal and chemical stability only in the absence of high concentration of chloride.
Live cell imaging with Ag nanodots also poses significant challenges. The excellent chemical stability offered by hairpin-encapsulated Ag nanodots, however, enables direct preparation in DMEM and significantly improved stability in cell culture without apparent silver salt precipitates during nanodot formation. For these syntheses, AgNO3 is directly added to DMEM solutions containing the hairpin of interest, followed by chemical reduction. Using ATATC12, direct nanocluster preparation in DMEM produces emitters with greatly improved stability compared to those prepared with the same sequence in regular phosphate buffer at 4 °C. It is likely that those with stable DNA-Ag+ complexes are chemically pre-selected during the incubation of DNA-Ag+ prior to chemical reduction, yielding stable emission. However, those unstable DNA-Ag+ complexes could not form silver nanodots. To further improve the nanodot thermal stability in high salt conditions, we increased the stem length in the hairpin structure for nanodot preparation directly in DMEM. Employing 5′-CGAACGCGC12CGCGTTCG-3′ (hairpin melting temperature (Tm) of 72 °C),35 bright 618 nm emission with excellent stability in PBS as well as DMEM at 37 °C is observed with a half life of 5 h. This is more than two-fold improved over the ATATC12 encapsulated silver nanodots (Fig. 1D).
Since the high charge density of DNA limits intracellular availability, and endosomal escape when using transfection reagents is often a problem,36,37 both microinjection and cell penetrating peptide (CPP) conjugation were utilized for cytosolic delivery. Although highly cargo dependent in subcellular localization, CPPs have been demonstrated to effectively import small cargoes ranging from fluorophores to oligos and small proteins.38,39 Large cargo such as quantum dots, however, are unable to reach the cytosol upon CPP conjugation.40 As a proof-of-concept, we conjugated a commonly employed CPP, penetratin,41 to oligocytosine ssDNA scaffolds for nanodot delivery. Penetratin is derived from the third helix of the Antennapedia homeodomain and has been widely investigated for cargo delivery due to its low cytotoxicity and high delivery efficiency.38,39,42,43 Three glycines and one cysteine were added to the C-terminus of the penetratin peptide to yield RQIKIWFQNRRMKWKKGGGC, in which the thiol group can be tethered to amino-group-functionalized oligocytosines (Integrated DNA Technologies, Coralville, IA, USA) through the heterobifunctional cross-linker sulfo-SMCC to yield conjugate penetratin-C12. The oppositely-charged penetratin and ssDNA tend to form aggregates due to electrostatic interaction, mandating use of high NaCl concentration (1 M) and high temperature (60 °C) during synthesis. The resulting conjugate and emissive nanodots also require high ionic strength for solubility (1 M sodium acetate). Only short (and therefore less anionic) strands yielded good conjugation efficiencies, dictating conjugation with C12 for nanodot formation. Red-emitting Ag nanodots were formed within this penetratin-C12 conjugate by mixing with silver nitrate in aqueous sodium acetate solution, followed by borohydride reduction at a base : silver ratio of 2:1. The resulting penetratin-C12Ag nanodots possess the essential photophysical properties of ssDNA-encapsulated Ag nanodots, i.e., 640 nm emission at 575 nm excitation (Fig. 1E),21 indicating that the electrostatic interaction between penetratin and ssDNA does not interfere with the formation and photophysics of the Ag nanodots. Cargo delivery was examined using NIH 3T3 cells and bovine pulmonary artery endothelial cells (BPAEC). Penetratin-C12Ag nanodots (10 μM) initially bind with and eventually cross the cell membrane, with silver nanodots being readily detected in nuclei within 10 min either in NIH 3T3 or BPAEC cells (Fig. 3A and D, respectively), with some expected residual membrane staining.44 The DNA-encapsulated Ag nanodots alone showed no staining of cells (data not shown), indicating penetratin-dependent cellular uptake. The internalization of penetratin-C12 Ag nanodots ceases at 4 °C, which is in line with reports that endocytosis may be involved in CPP translocation.39 Further, the penetratin-C12 Ag nanodots entered cells with intact cell membranes, as demonstrated by co-incubation of penetratin-C12 Ag nanodots with the cell membrane-impermeable dye Sytox Green (Invitrogen) which only stains nuclei of cells with compromised membranes (Fig. 3B, C). In all the Ag nanodot-stained cells, less than 15% of cells showed poor cell membrane integrity. Though the challenge of charge compensation limits solubility and size of penetratin-DNA conjugates, this demonstrates that intracellular imaging studies with ssDNA encapsulated Ag nanodot fluorophores indeed hold promise.
Fig. 3 Intracellular delivery of DNA-encapsulated Ag nanodots. A–D, Penetratin-enhanced Ag nanodot internalization. A to C, Colocalization of NIH 3T3 cells co-incubated with penetratin-C12 Ag nanodots and Sytox Green nuclear staining dye which only stains nuclei of cells with compromised membranes. The merge (C) of red emission (A) from penetratin-C12 Ag nanodots and green emission (B) from Sytox Green indicates that only a few cells show poor membrane integrity. Panels A–C are at the same scale. Penetratin-C12Ag nanodots were excited at 543 nm and Sytox Green at 488 nm. D, Merge of a bright field image and a fluorescent image of BPAECs incubated with penetratin-C12 Ag nanodots for 10 min. E and F, 615 nm emitter was micro-injected into NIH 3T3 cells, shown in phase contrast (E) and fluorescence after microinjection (F). All scale bars are 25 μm. |
Conversely, direct microinjection circumvents difficulties in membrane transport and the problem of endosomal escape. Microinjection (Eppendorf FemtoJet) of concentrated 615 nm (CGCGC12) emitting nanodots into live NIH 3T3 cells, for example, demonstrates the ability of these Ag nanodots to show bright red fluorescence in the cytosol, and indicates promise for more generalized cellular staining using these or similar nanodot probes (Fig. 3E and F), when stability is also further improved. Although eventually higher fluorescence intensity is observed in the nucleus even with microinjected Ag nanodots after 10 min, the distribution of fluorescence is initially almost exclusively cytosolic. This distribution appears uninfluenced by the Ag cluster, as fluorescein-tagged ssDNA of the same sequence exhibits a similar distribution inside living cells.
Footnote |
† Current address: Department of Chemistry & Education, Seoul National University, Seoul 151-742, South Korea. |
This journal is © The Royal Society of Chemistry and Owner Societies 2011 |