Evaluation of PicoGreen variants for use in microscopy and flow cytometry

PicoGreen is a fluorescent probe that binds dsDNA and forms a highly luminescent complex when compared to the free dye in solution. This unique probe is widely used in DNA quantitation assays but has limited application in flow cytometry and microscopy. Here we have investigated various PicoGreen variants for the ability to stain low amounts of cytosolic DNA present in many tumor cells. Analysis of stained cells by flow cytometry and fluorescent microscopy showed that certain variants improved the ability to stain low levels of cytosolic DNA when compared to the commercially available PicoGreen molecule.


Introduction
The integrity of DNA is constantly challenged by endogenous and environmental genotoxic factors [1][2][3]. Cells have evolved elaborate mechanisms to defend the integrity of DNA. Detection of damaged DNA activates complex DNA repair systems and activates innate immune pathways that alert the immune system to the presence of potentially damaged or infected cells [4,5]. In addition, recent studies found that genomic or mitochondrial DNA can accumulate in the cytosol of cells [5]. Similar to damaged DNA, mislocalized DNA triggers innate immune responses including the production of type-I interferons and other inflammatory cytokines [6]. The detection of DNA in the cytosol requires sensitive tools that can be used in conjunction with mitochondrial, nuclear and membrane markers in microscopy and flow cytometry.
Fluorescent probes that specifically bind nucleic acids are useful tools to study DNA metabolism [7]. PicoGreen (PG) is a highly sensitive fluorescent dye developed in the 1990s that detects double-stranded (ds) DNA [8,9]. PG shows significant selectivity for dsDNA and DNA:RNA hybrids over single-stranded (ss) DNA and RNA. Binding to linear dsDNAs results in slightly higher signals when compared to supercoiled plasmids. PG is able to strongly bind to highly polymeric DNA and short DNA duplexes <20 bp. Binding of dsDNA by PG is not affected by the presence of salts, proteins, poly(ethylene glycol), urea, chloroform, ethanol, and agarose, while some ionic detergents and heparin interfere with PG binding.
PG has an excitation peak at 480 nm and an emission peak at 520 nm. Upon binding DNA, PG fluorescence increases >1000-fold while unbound PG has virtually no fluorescence. PG is very stable to photo-bleaching, which allows longer exposure times.

Cell lines
Cervical adenocarcinoma HeLa cells were purchased from ATCC. The TRAMP-

Enhanced stainings of DNA by PicoGreen variants in confocal microscopy analysis
We have recently shown the presence of cytosolic DNA in many human cancers and cancer cell lines including TRAMP-C2 and HeLa cells [11][12][13]. To evaluate the ability of asymmetrical cyanine dyes to label cytosolic DNA, we stained the murine prostate cell line TRAMP-C2 with 65 different PicoGreen variants, all of which share the same core structure (Fig 1). Analysis of live cells staining of TRAMP-C2 cells identified a number of PicoGreen variants that stained cytosolic DNA with higher intensity than the commercially available PicoGreen. The brighter cytosolic DNA signals by PicoGreen variants #2, #34, #41 and #48 correlated with enhanced nuclear DNA staining suggesting that these variants result in stronger fluorescent signals upon binding of dsDNA (Fig 1).

Enhanced stainings of DNA by PicoGreen variants in flow cytometric analysis
Next, we tested the ability of PicoGreen variants to stain dsDNA present in TRAMP-C2 cells by flow cytometry. Similar to our findings using microscopy to analyze the cytoplasmic PicoGreen signals, staining of dsDNA with the PicoGreen variants #34, #41 and #48 resulted in the highest mean fluorescence (MFI) when compared to the commercially available PicoGreen (Fig 2). Interestingly, the MFI of PicoGreen variant #2, which was among the best variants for staining of dsDNA in microscopy, was lower than the commercially available PicoGreen dye suggesting that PicoGreen variant #2 is more sensitive to photobleaching by the more powerful 488 nm laser used in flow cytometry.

Fixation condition impact staining of DNA by PicoGreen variants
In microscopy, fixation is often a prerequisite to further processing of clinical samples. For that reason, we tested the ability of PicoGreen variants to stain dsDNA in paraffin-and paraformaldehyde (PFA)-fixed TRAMP-C2 cells. The PicoGreen variants with the highest normalized MFI (Fig 1) for staining of dsDNA in live cells were chosen. Paraffin-fixed TRAMP-C2 cells were smaller in size and no clear cytosolic dsDNA signal could be detected for any of the PicoGreen variants (Fig 3). In contrast, fixation of cells with 4% PFA slightly enhanced the staining intensity of PicoGreen and also increased the resolution of the cytosolic DNA staining. However, PFA fixation abrogated the ability of the PicoGreen variant #55 to stain cytosolic DNA (Figs 1 and 3). In summary, our data suggest that PFA fixation and the use of certain PicoGreen variants can drastically increase dsDNA staining signals and resolution.

Anti-fade reagents fail to enhance the staining of PicoGreen variants
Anti-fade reagents were shown to suppress photobleaching and cause little or no quenching of the initial fluorescent signal [14,15]. Slowfade and Slowfade Gold are glycerol-based mountants of several commonly used fluorophores, but both reagents failed to enhance PicoGreen staining of cytosolic DNA (Fig 4). Slowfade

PicoGreen variants specifically recognize dsDNA
To analyze the specificity of the different PicoGreen variants towards different nucleic acid types, we measured the fluorescence enhancement in vitro by spectroscopy when the different dyes bound to dsDNA, ssDNA, ssRNA or dsRNA ( Fig 5). All variants showed specificity towards dsDNA when compared to ssDNA, ssRNA and dsRNA. Some dyes such as PicoGreen variant #34, #41, #48 and #55 also weakly bound to dsRNA and to a lesser degree ssDNA. In summary, the PicoGreen variants #34, #41 and #48 showed the greatest fluorescence enhancement when bound to 500 ng/ml dsDNA, which was more than 2-fold higher than the commercially available PicoGreen (Fig 5A). These results are in agreement with our earlier observation that the PicoGreen variants #34, #41 and #48 resulted in the highest fluorescence when analyzed by microscopy (Fig 1) or flow cytometry (Fig 2). The differences in fluorescence enhancement were less pronounced when smaller amounts of dsDNA were measured (Figs 5B and 5C), but most PicoGreen variants still outperformed the commercially available PicoGreen.

In vivo specificity of PicoGreen variants
To examine the specificity to the PicoGreen variants towards dsDNA in cells, we

Discussion
To visualize DNA present in the cytosol of many tumor cells, we have previously used commercially available PicoGreen [11][12][13]. The fluorescence enhancement of PicoGreen when bound to dsDNA is very high with negligible background signal from unbound PicoGreen molecules [16]. PicoGreen fluorescence signals remain stable during the image acquisition as it is relatively insensitive to photobleaching.
We have successfully used PicoGreen for up to 90 minutes in live-cell imaging [11]. PicoGreen stainings of tumor cells suggested a granular distribution of cytosolic DNA in tumor cells, while stainings using dsDNA-specific antibodies indicated a more homogenous distribution of DNA in the cytosol possibly in part due to greater sensitivity of the antibody staining [11]. PicoGreen was initially developed for quantification of DNA by spectrofluorometer [16]. We therefore Furthermore, it is possible that other nucleotide species such as RNA:DNA hybrids [11][12][13] and long dsRNAs [17] are stained by the PicoGreen variants.
The increased fluorescence emission of dsDNA-bound PicoGreen depends on the structure and binding mode of PicoGreen [9,18]. PicoGreen has a 4-[[2,3dihydro-3-methyl-(benzo-1,3-thiazol-2-yl)-methylidene]-quinolini-um]+ core structure. In addition, it contains a phenyl group at position one of the quinolinium ring and a N-bis-(3-dimethylaminopropyl)-amino residue at position two of the quinolinium ring. PicoGreen carries three positive charges, which are likely to contribute to its high binding affinity for dsDNA. We observed that electron donating substituents on the delocalized benzothiazolyl-quinolinium system common to PicoGreen variants correlate with brighter signals from cellular dsDNA.
Electron-donating groups were previously found to form more stable ligand-DNA complexes [19]. In addition, the electron donating substituents can also induce an increase in the molar absorption coefficient and a shift in the fluorescence spectra, which was not observed for the different PicoGreen variants analyzed in this study ( Fig 5) [20].
Photobleaching is an irreversible photochemical process that stops the emission of photons by fluorophores [21]. The bleaching process is likely to involve reaction of excited-dye molecules with reactive oxygen species (ROS) that are produced when molecular oxygen interacts with different electronic states of the dye [22]. Most antifade reagents are reactive oxygen species scavengers.
Glycerol-based Prolong was reported to have among the highest anti-fading properties for many applications while good antifading properties have also been reported for the other reagents used in this study [23,24]. Surprisingly, none of the anti-fading reagents including Prolong improved the staining of cytosolic DNA in tumor cells. In fact, Prolong and Slowfade Gold decreased the staining intensity of some PicoGreen variants. Consistent with this finding, it was previously found that antifade reagents can sometimes negatively affect fluorophores by quenching the fluorescence of the dye [14].
In summary, we show in this study that the addition of electron donating

Conflict of interest
The authors declare no conflict of financial interests.       Alexa647 + cells including nuclear PicoGreen signals. Scale bar denotes 10 μm.