SEARCH

SEARCH BY CITATION

Keywords:

  • ascorbic acid;
  • BrdU;
  • click chemistry;
  • Fenton chemistry;
  • high content analysis

Abstract

  1. Top of page
  2. Abstract
  3. Materials
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgment
  8. Literature Cited
  9. Supporting Information

Ascorbic acid (AA) and copper have been increasingly employed in flow cytometry (FCM) and high content analysis (HCA) since the introduction of “click chemistry” as a non-destructive alternative to classical 5-bromo-2′-deoxyuridine (BrdU) immunodetection for DNA synthesis and proliferation assays. Mixtures of ascorbate and catalytic copper, under certain experimental conditions, act as oxidizing agent, catalyzing the formation of reactive hydroxyl radicals through hydrogen peroxides decomposition via Fenton reaction. We developed a procedure for BrdU incorporation detection based on the use of AA and cupric ions as DNA damaging agents. Optimal DNA damaging conditions were identified and found to provide results comparable with “click” 5-ethynyl-deoxyuridine (EdU) cycloaddition approach and classical BrdU immunodetection. Scavenger agents were found to prevent hydroxyl-induced DNA damages, providing the proof-of-concept for the use of this procedure for DNA denaturation prior to BrdU detection. We demonstrated hydroxyl radicals' reaction to be readily applicable to HCA and FCM assays, for both classical BrdU immunostaining and EdU cycloaddition procedure. This technique was successfully employed for BrdU pulse-chase experiments and in multiparametric immunofluorescence assays for the simultaneous detection of labile phosphoproteins in intact cells. The use of AA/Cu prior to immunodetection for BrdU incorporation assays is a viable alternative to chemical/physical DNA denaturing agents (acids or heat), since it allows preservation of labile epitopes such as phosphoproteins, and over enzymatic agents (digestion with DNases) for its lower cost. © 2013 International Society for Advancement of Cytometry

l-Ascorbic acid (Vitamin C, AA) is a water-soluble antioxidant, which efficiently scavenges free radicals and other radical oxygen species (ROS) produced by cell metabolism. ROS generation is associated with several forms of tissue damage and plays a role in the development of a multitude of chronic diseases [1-3]. AA acts as donor/acceptor in redox reactions, providing a protective activity against free radicals. A mechanism able to control ROS generated during aerobic metabolism is essential for cell viability and AA and glutathione work together to prevent accumulation of oxidative damages that can lead to cell death [4].

Despite its role as radical scavenger, in the presence of free metal ions AA acts as a pro-oxidant and can damage biological molecules, such as nucleic acids. AA reduces cupric (Cu++) to cuprous (Cu+) ion, which is capable of catalyzing the formation of reactive hydroxyl radicals through hydrogen peroxides decomposition via the so-called Fenton's reaction [5-7]. The hydroxyl radicals efficiently react with deoxyribose backbone of DNA [8]: hydroxyl radicals abstract an hydrogen atom from deoxyribose, leading to the formation of a free radical intermediate, and eventually triggering extensive damages such as backbone cleavages, abasic sites, and nucleotide aldehydes [9, 10].

In the last years, AA and cupric ions were employed together in cell-based assays as catalytic agents for “click chemistry” reactions to detect protein, RNA, and DNA synthesis. A common application of click chemistry in cell biology is the quantification of DNA synthesis in cells, which incorporate 5-ethynyl-deoxyuridine (EdU) during S phase: a copper-catalyzed azide-alkyne cycloaddition (CuAAC) has been exploited to covalently bind azide-conjugated fluorescent probes to EdU, without the need of prior DNA denaturation [11]. In previous works, we used 5-bromo-2′-deoxyuridine (BrdU)-azide derivatives followed by anti-BrdU immunodetection and fluorescence amplification by secondary antibody to analyze cells in S phase [12, 13].

Here we show that, at concentrations similar to those employed to catalyze click chemistry reactions in intact cells, AA/copper mixtures also act as DNA damaging agents and introduce double-strand breaks, which render incorporated halogen deoxyuridine derivatives accessible for immunodetection. We exploited this unexpected reactivity to develop a novel procedure for BrdU/EdU detection in multiplexed flow cytometry (FCM) and high content assays.

Materials

  1. Top of page
  2. Abstract
  3. Materials
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgment
  8. Literature Cited
  9. Supporting Information

Reagents and Antibodies

Boric acid, Click-IT® AlexaFluor® 488 kit, DAPI, dichlorodihydrofluorescein diacetate (DCFH), FBS, McCoy's 5A medium, PBS, ΦX174 (5.4 Kb) supercoiled DNA plasmid, PenStrepto solution, L-glutamine, AlexaFluor 488 or 647 goat anti mouse and rabbit IgG antibodies, agarose, TBS, TO-PRO®-3, Vybrant® Dye Cycle™ Green were purchased from Invitrogen. DRAQ5™ was purchased from Biostatus. BrdU, bathocuproinedisulfonic acid (BCSA), camptothecin, l-cysteine, copper (II) sulfate, DMSO, ethidium bromide (EB), hydrogen peroxide, luminol (5-amino-2,3-dihydro-1,4 phthalazinedione), methanol-free-Formaldehyde 37%, methanol, paclitaxel (PTX), propidium iodide (PI), RNAse A from bovine pancreas, sodium ascorbate (AA), terephthalic acid (TPA), trypsin/EDTA solution, Triton® X-100, and Tween® 20, were purchased from Sigma-Aldrich. DNA ladder molecular weight marker (N3232) was purchased by Biolab. Aldehyde site (ARP, DNA and protein) detection kit (catalog # 600170) was purchased from Cayman chemicals. Anti-BrdU mAb (clone B44) was purchased from BD Biosciences. BrdU Labeling and Detection Kit (cat.11296736001) nuclease-based was purchased from Roche. Anti-8-hydroxyguanosine mAb (clone 15A3) was purchased from Abcam. Phospho-histone H3 (Ser10) rabbit mAb (06–570) was purchased from Millipore. EdU (5-Ethynyl-2′-deoxyuridine, PY7562) was purchased from Berry & Associates, MI. 5′-BMA used for CuAAC experiments was synthesized as previously described [12].

Cell Cultures

Cell lines were obtained from ATCC (American Type Culture Collection), European Collection of Cell Cultures (ECACC), DSMZ (Leibniz Institute DSMZ—German Collection of Microorganisms and Cell Cultures GmbH) and authenticated by comparison of STR fingerprint with reference profiles as shown in Supporting Information [14]. U-2 OS (ATCC) cells were grown in McCoy's 5A medium supplemented with 10% FBS, L-glutamine 2 mM, 1% Pen/Strep solution and were maintained in 5% CO2 at 37°C. A2780 (ECACC), HCT-116 (ECACC), and MOLM-13 (DSMZ) cells were grown in complete RPMI1640 medium. Cells were grown as monolayers in CellStar™ T75 flasks (Greiner Bio-One, Germany) and routinely subcultured twice a week.

Methods

  1. Top of page
  2. Abstract
  3. Materials
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgment
  8. Literature Cited
  9. Supporting Information

DNA Gel Electrophoresis

To analyze ascorbate/copper (AA/Cu) effects on DNA, 1 μg of plasmid DNA (ΦX174, 5.4 Kb) in 50 μl Tris HCl (pH 7.4) was treated with different concentrations of AA (from 10 μM to 50 mM) in the presence of 20 μM copper (II) sulfate for 20 min at room temperature. The reaction was stopped by placing the reaction mixture in ice. Loading buffer (5 μl; 0.25% bromophenol blue, 40% sucrose) was added and samples (20 μl) were analyzed by electrophoresis in 1% agarose gel in TAE buffer. For low molecular weight fragments detection followed by DNA degradation in cells, a modified ladder procedure was applied [15]. Methanol fixed U-2 OS cells (1.5 × 106) were treated with AA/Cu mixtures in the same conditions already described, then collected and washed twice in cold PBS containing 5 mM EDTA (pH 8.0) and embedded in agarose plugs prepared with 1% agarose in TAE buffer (40 mM Tris/Acetate, 1 mM EDTA, pH 8.0). Plugs were kept on ice for 2 h, incubated with 0.5 M EDTA (pH 8.0), 1% Sarkosyl, 1 mg/ml proteinase K solution at 0°C (in ice) for 1 h and then at 48°C overnight, and finally rinsed four times in 1× TAE. Each plug was loaded onto a 2% agarose gel. The run was performed in 1× TAE buffer with a constant voltage of 100 V until samples resolution. These conditions were helpful to separate low molecular DNA fragments (on the bottom of gel) from high molecular fragments or undamaged DNA. At the end of run, both gels were stained in TAE containing 0.5μg/ml EB and visualized by UV transillumination.

CuAAC Cycloaddition (Click Chemistry) and “Fenton” Chemistry for FCM

U-2 OS cells were pulsed for 30 min with 10 μM EdU (for CuAAC) or 30 μM BrdU (for Fenton reaction or standard acidic conditions), washed with PBS, detached with trypsin, fixed in methanol 70%, and stored at −20°C. Since in previous work [13, 16], we observed differential anti-proliferative activity of EdU and BrdU in asynchronously growing cells, pulse-and-chase experiments were performed using nontoxic EdU and BrdU concentrations, but still effective for tracking cells during DNA synthesis.

For CuAAC staining, PBS buffer containing cells was mixed for 15 min with 0.1 mM BMA, 10 mM AA and 2 mM cupric sulfate. AA stock solution prepared fresh (1 mg/ml) was discarded after use. For Fenton reaction, we proceeded with a similar protocol, without the use of BMA and using 10–50 mM AA in presence of 0.2–2 mM cupric sulfate.

DNA cleavage with AA/Cu mixtures worked worse when AA solutions were left to air since ready-to-use solutions should be discarded and fresh solutions should be prepared from AA powder stored in modified atmosphere; in fact AA is readily oxidated in browning products due to its peculiar chemical behavior leads to furfural and pyrone species [17]. Cells were washed with PBS containing Tween 20 0.5% (PBST) for 15 min and incubated for 30 min at RT with 150 μl anti-BrdU mAb diluted 1:50 in PBST. After washing with PBST, cells were incubated for 30 min with AlexaFluor 488 conjugated F(ab)2 fragment goat anti-mouse IgG diluted 1:300 in PBST at RT in the dark. Finally, cells were stained in PI 5 μg/ml in PBS plus RNase A and incubated for 1 h at RT before DNA content analysis. For double staining with AlexaFluor 488 azide using Click-IT, azide concentration was employed as previously reported [13].

Image Acquisition and High Content Analysis

For high content analysis (HCA), U-2 OS cells (6,000/well) were seeded in ViewPlate™ 96F (PerkinElmer Life Sciences, Boston, MA) and cultured for 24 h. Cells were pulsed with 10 μM EdU or 30 μM BrdU for 30 min, then fixed for 15 min by 3.7% formaldehyde and stored in PBS. Before nucleotide revealing, cells were permeabilized with 0.3% Triton® X-100 in PBS for 15 min at 4°C and washed twice with PBS. For CuAAC treatments, cells were exposed for 20 min to 0.1 mM BMA, 10 mM AA, and 2 mM cupric sulfate in 100 μl/well. For Fenton reaction, cells were exposed for 30 min to 10 mM AA and 2 mM cupric sulfate in 100 μl/well. For assay setup, different combinations of AA/Cu were employed according to the experimental design. Finally, cells were incubated for 30 min with anti-BrdU mAb diluted 1:100 in PBST, washed with PBS, and incubated with 100 μl anti-mouse AlexaFluor® 488 IgG mAb diluted 1:1000 and 1 μg/ml DAPI in PBST at room temperature for 30 min in the dark.

Phospho-histone-H3 (p-HH3) staining was performed as previously reported [18]. After incubation, cells were washed twice and resuspended in 200 μl PBS. Image acquisition and analysis was performed by means of a ThermoFisher ArrayScan™ VTI HCS reader equipped with a 200 W mercury-xenon lamp. A quadruple band filter XF93 was used to acquire fluorescence in the DAPI channel and green channel with a 10× objective; at least 1,000 cells were analyzed for each sample. The nuclear fluorescence intensity was measured with the Cyto-Nuc Translocation Bioapplication (Thermo Fisher Scientific) [12].

Cuprous Ion Determination by Colorimetric Assay

Cuprous ion quantification was performed by BCSA colorimetric assay which forms stable red Cu+-bathocuproine complex, mixing 160 μl of PBS solution containing 0.36 mM BCSA with 20 μl of AA and 20 μl copper sulfate solutions at reported final concentration in 96 well plates. After 5 min, reaction was stopped by the addiction of 50 μl of 0.01 M EDTA and samples were analyzed using Fusion plate reader (Perkin Elmer) at 490 nm [19].

Cell-Free Assay for Hydroxyl Radical and ROS Quantification

Hydroxyl radicals react with luminol solution generating chemiluminescence [20]: this assay was employed to monitor ROS production during Fenton reaction for BrdU detection. Briefly, in 140 μl of luminol 30μM dissolved in borate buffer (0.05 M, pH 10) were added 20 μl AA (14×) solution, 20 μl cupper sulfate (14×) solution and 100 μl hydrogen peroxide 0.004% solution in 96 well black plates. Chemiluminescence determination was performed within 5 min using a Fusion luminometer (Perkin Elmer) at 430 nm.

Measure of Oxidative Stress by DCFH, ARP, and Anti-8-hydroxy-deoxyguanosine Detection Assays

U-2 OS cells were treated with AA and copper solutions in 1 ml PBS buffer at the reported concentrations for 20 min. For DCFH assay, we followed a protocol previously reported [21]; hydrogen peroxide 0.03% was used as positive control. Aldehyde sites originating from protein and DNA modification after oxidative damage were detected using ARP probe, o-(biotinylcarbazoylmethyl) hydroxylamine, and FCM using fluorescein avidin. Epigallocatechin gallate (EGCG), used as positive control, was used to cell cultures at final dilution of 1:100 for 24 h. After treatments, cells were fixed for 10 min with the solution provided in the commercial kit. Cells were then incubated in 20 μl ARP binding solution and 1 μl ARP probe for 1 h at 37°C. After three washes, cells were denatured for 30 min at 37°C, and reaction was blocked by incubation with 200 μl of blocking solution at room temperature for 45 min. Finally, labeled cells were stained with 10 μl of avidin-FITC staining solution for 1 h at 37°C. Cells were washed in washing buffer and analyzed by FCM [22, 23]. For anti-8-hydroxy-dG assay, cells were washed with PBS and prefixed by adding 1% formaldehyde solution in PBS for 15 min [24]. Cells were washed with PBS, permeabilized with ice-cold methanol for 15 min, and rehydrated in PBS before blocking with PBS containing 10% normal goat serum. The blocking solution was replaced with PBS containing 0.2% goat serum. 8-hydroxy-dG DNA modifications were revealed with anti-8 hydroxyguanosine mAb followed by AlexaFluor 488-conjugated secondary mAb (1:400 in PBS for 1 h) and green fluorescence was analyzed with a BD FACSCalibur [25, 26].

DNA Damage Assay

A modified PI binding assay was used for DNA damage analysis in U-2 OS cells based on the formation of a fluorescent complex between double-strand DNA and PI [27]. One million fixed cells were stained with 1 ml of PI at saturating concentration (50 μg/ml PI, plus RNAse and NP-40 at 37°C for 1 h) [28]. Cells were washed twice with PBS and incubated in solutions containing different concentrations of AA plus copper for 1 h. Cells were washed again with PBS and then analyzed for relative DNA content in linear/logarithmic scale. PBMCs (peripheral blood mononuclear cells) were processed with the same protocol and added to each experiment as calibrator of DNA content and analyzed separately during FCM acquisition.

When AA/Cu mixture is added to stained cells, orange/red fluorescence resulted from interaction with dye (i.e., PI) and nuclear DNA was significative decreased after DNA damage. This effect defined in this article as loss of stainability or fluorescence bleaching (Fig. 3 and Supporting Information Figs. 7 and 8) was measured as Foldstandard/Foldsample ratio, where each term corresponds to the relative MFI of untreated or treated sample versus MFI of PBMCs. In order to evaluate different DNA damages, additional DNA dyes were used: 7-AAD (GC base specific), Hoechst-33342 (AT base specific), TO-PRO-3, DRAQ5, and Vybrant Dye Cycle Green.

For DNA binding assays, different cell lines (A2780, HCT-116, and MOLM-13 cells) were processed as already described and compared to U-2 OS cells. Three compounds were used to test hydroxyl radical scavenging activities: pantopazole 2 mM [29], TPA 10 mM [24], and cysteine 10 mM [30]. Pantoprazole (PP) was extracted and purified from commercial drug preparations [31]. In order to evaluate dye binding capabilities after Fenton reaction, U-2 OS cells were stained after incubation with AA plus copper sulfate solutions as previously reported, and DNA content was analyzed. For Hoechst 33342 fluorescence determinations, a 355 nm UV laser equipped BD FACSAria SORP was employed, otherwise for visible excitable dyes a BD FACSCalibur was used as previously described. Additional experimental details are reported in Supporting Information.

FCM Setup

Cells were analyzed by dual laser 488/633 nm BD FACSCalibur. Fluorescence signals were detected using a band-pass filter 530/30 nm and 620/35 nm in combination with a dichroic mirror at 570 nm for the 488 nm excitation and band-pass filter 661/16 nm for the 635 nm excitation. For DNA damage experiments (Fig. 3), relative MFI were measured based on G1 phase peak, after FSC-H/SSC-H and Area/Width fluorescence gating, acquiring either linear (for cell cycle analysis and CV% evaluation) or logarithmic (for DNA damage assay) scales. For cell cycle studies (Fig. 4), aggregates were excluded from DNA analysis by gating on area/width fluorescence, acquiring at least 10,000 events inside singlets gating. UV fluorescence of cells stained with Hoechst 33342 was detected using a BD Special Order Product FACSAria I equipped with three laser at 355, 488, 633 nm. UV fluorescence were detected at 450/50 and 530/30 nm pass-band trigon.

Data Analysis and Hierarchical Clustering

FCM gating was performed using BD CellQuest™ (BD Biosciences), DNA histograms were fitted using Modfit® 3.0 (Verity Software House), and heat maps were produced using Spotfire™ (TIBCO). Three-dimensional graphs were created using SigmaPlot™ software (Systat Software).

UPGMA (Unweighted Pair Group Method with Arithmetic Mean) clustering and Euclidean distance were used for hierarchical clustering. Cellular assay results were normalized as logarithmic ratio [log(raw data)/average of data set] for direct comparison of non-homogeneous datasets (Fig. 5 and Supporting Information Excel file).

Results

  1. Top of page
  2. Abstract
  3. Materials
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgment
  8. Literature Cited
  9. Supporting Information

AA/Cu Mixtures Cause Oxidative DNA Damage Via Fenton Reaction

Figure 1 shows the proposed model of chemical reactions involving “click” and Fenton chemistry. AA and cupric/cuprous ions (Cu2+/Cu+) act as cleavage agent for BrdU detection reacting with EdU or BrdU azide for copper catalyzed (CuAAC) addition, as well as for Fenton reaction. While AA/cupric ions (AA/Cu couple) are able to generate Cu+ ions to efficiently catalyze alkyne/azide cycloaddition [13], in BrdU but also generate secondary hydroxyl radicals [7, 32]. Triazole-linked bromodeoxyuridine moiety or BrdU detection by hydroxyl radical attack both provide accessible epitopes for the subsequent anti-BrdU mAb staining. The observed DNA cleavages mainly originate from hydrogen abstraction on sugar backbone, resulting in terminal strand breaks with formation of abasic sites, aldehyde nucleotide, free bases, or oxidative lesions on DNA strands (Fig. 1) (10,33). In order to test the hypothesis that AA and copper ion mixtures could work as “chemical nucleases,” super-coiled plasmid ΦX174 DNA was incubated in different buffers containing high (50 mM) or low (0.01 mM) AA alone, or in the presence of 0.2 mM copper sulfate and analyzed by agarose gel electrophoresis at AA concentrations from 1 to 50 mM. In the absence of copper, increasing AA concentrations proportionally induce the formation of linear (“LIN”) versus super coiled (“SPC”) DNA plasmid (Fig. 2A). After addition of 0.2 mM Cu alone, linearized DNA was comparable to control sample. DNA incubation with AA/Cu mixtures caused dramatic changes in DNA integrity: few nicks were observed at AA concentrations (0.01–0.1 mM) plus 0.2 mM copper sulfate, while complete DNA degradation appeared at higher AA concentrations (1–50 mM) (Fig. 2A). In order to confirm this finding in intact cells, U-2 OS cells were fixed and exposed to AA from 50 to 1 mM ± 0.2 mM copper sulfate and analyzed by pulsed field electrophoresis [15]. Samples incubated with 10–50 mM AA and copper showed severe genomic DNA damage with appearance of low molecular weight fragments, confirming the results obtained with plasmid DNA (Fig. 2B).

image

Figure 1. A) AA in the presence of cupric ions acts as catalyst for 1,3-dipolar cycloaddition reaction between EdU and azide-BrdU or any azide derivatives (*) (figure modified from Ref. [13], and for Fenton reaction generating hydroxyl radicals and forming DNA strand breaks for subsequent anti-BrdU immunodetection. B) Putative Fenton reaction products generated by nucleotide hydrogen H-1 and H-5 abstraction after hydroxyl radical attack.

Download figure to PowerPoint

image

Figure 2. Effects of hydroxyl radicals in plasmidic and genomic DNA. A) Electrophoretic profile of DNA after exposure to sodium ascorbate ± 0.2 mM copper sulfate. M) Marker 10 Kb; NT) control ΦX174 DNA; Nuc) DNA with nuclease S1 plus exonuclease at RT; LIN) linear form; SPC) super coiled form; and DEG) degraded form. B) Electrophoretic profile of genomic DNA isolated from U-2 OS cells after exposure to ascorbate ± copper. C) DCFH assay. D) ARP-aldehyde reactive probe and E) 8-OH-dG. Green: U-2 OS cells not treated; black: 50 mM ascorbate; blue: 0.2 mM copper sulfate; purple: 50 mM ascorbate ± 0.2 mM copper; red: Reference standard (hydrogen peroxide for DCFH and EGCG solution for ARP and 8-OHdG). [Color figure can be viewed in the online issue which is available at wileyonlinelibrary.com.]

Download figure to PowerPoint

image

Figure 3. PI binding assay for DNA damage analysis. A) U-2 OS cells pre-stained with PI were treated with ascorbate alone or in the presence of 0.2 mM cupric ions (AA/Cu); B) U-2 OS cells stained with PI after AA/Cu Fenton reaction. C) Free radical protection by ROS scavenger: U-2 OS cells were exposed to ascorbate and treated with 0.2–2 mM copper ± 2 mM PP, 10 mM TPA, or 10 mM L-cysteine. Scavenger agents prevent hydroxyl radical DNA damage, as demonstrated by PI fluorescence shift. [Color figure can be viewed in the online issue which is available at wileyonlinelibrary.com]

Download figure to PowerPoint

image

Figure 4. A) FCM analysis of BrdU and EdU incorporation in U-2 OS cells. For BrdU detection, cells were stained after DNA denaturation by HCl treatment or by Fenton chemistry conditions using cupric ions (Cu) or ascorbate/cupric ions (AA/Cu) solutions. For EdU detection, cells were stained in the presence of AdU, BMA, or BDA using click chemistry conditions. B) U-2 OS cells were labeled with BrdU for 20 min and released in BrdU-free medium. At specific time points (8–18–27–40 h) cells were labeled with EdU for 20 min, then cells were processed for double staining with Click-IT™ reagents (AlexaFluor 488/EdU) and Fenton reagents (10 mM AA and 2 mM cupric solution). Gating of dot plots on the right represents different cell cycle phases: G1 (blue), G2 BrdU positive cells (green), G1 BrdU positive cells at the second cycle (red). [Color figure can be viewed in the online issue which is available at wileyonlinelibrary.com]

Download figure to PowerPoint

image

Figure 5. Effects of ascorbate and copper concentrations on Fenton BrdU- or click EdU-based analysis of DNA synthesis. A) HCA of U-2 OS cells stained with anti-BrdU antibody following hydroxyl radicals treatment; B) percentage of U-2 OS cells stained with anti-BrdU antibody following click reaction (CuAAC) with BrdU azide; C) chemiluminescence enhancement of luminol in the presence of hydroxyl radicals and copper; D) cuprous ions (Cu+) determination by spectrophotometric assay with BSCA. E) Similarity clustering of AA/copper effects observed in cell-free assays (luminol-enhanced chemiluminescence and cuprous determination), FCM (PI MFI reduction, CV% of DNA content, 8-hydroxy-deoxyguanosine and ARP increase) and HCA (BrdU positive cells fold increase and relative MFI). Condition “0” (untreated reference samples) is highlighted with a yellow square; AA/Cu combinations are indicated in Supporting Information (Fig. 12). [Color figure can be viewed in the online issue which is available at wileyonlinelibrary.com]

Download figure to PowerPoint

We performed DCFH assay for ROS determination, ARP assay for detection of aldehyde DNA modifications and induction of 8-hydroxy-deoxyguanosine (8-OH-dG) following cell exposure to 50 mM AA and 0.02 mM copper, alone or in combination, in comparison with reference standard treatments (hydrogen peroxide and EGCG). AA/Cu combination induced a paradoxical fluorescence decrease in DCFH assay due to cell death (see Supporting Information), which impaired detection of ROS activity (Fig. 2C). For this reason, fixed U-2 OS cells were also assayed for ARP and 8-OH-dG. In both cases, an increase of positive cells was observed for the AA/Cu combination as compared to untreated cells or AA and Cu alone (Figs. 2D and 2E), in agreement with previous observations with isolated DNA.

Oxidative Damage Leads to Decreased DNA Stainability with Fluorescent Dyes

In order to evaluate DNA damage after exposure to free radicals generated by AA decomposition in the presence of copper ions, U-2 OS cells pre-stained with PI were exposed to AA and copper mixture. As shown in Figures 3A and 3B, DNA histograms of cells exposed to AA alone were similar to untreated cells (CV 7% and 4%, respectively), with a DNA content close to that of peripheral blood mononuclear cells (PBMCs), used as reference standard (ratio: 1.3). After copper addition to AA, the resolution of DNA histograms decreased (CV% 17%), with a significant fluorescence bleaching observed in both untreated and AA-treated samples, as compared to PBMCs (ratio: 0.05). The same experimental procedure, but using 7-AAD and Hoechst 33342 instead of PI, resulted in four- and 16-fold decrease of mean fluorescence intensity, respectively, as compared with untreated stained cells. A2780 (ovarian cancer), HCT-116 (colon cancer) and MOLM-13 (leukemia) cell lines were stained and treated in same conditions described for U-2 OS cells to demonstrate that these effects are cell line independent: similar effects were observed on DNA histograms resolution and fluorescence intensity (Supporting Information Fig. 7). When AA/Cu reaction occurred before PI staining, as in the case of counterstaining for multiparametric FCM assay, we observed that AA alone had no effects on DNA degradation (CV 4%), while AA/Cu mixtures significantly impaired DNA stainability by PI (CV 10%), even though PI staining was performed at saturation condition (Supporting Information Fig. 8).

We investigated further DNA dyes, including TO-PRO-3, DRAQ5 and Vybrant Dye Cycle Green: fluorescence reduction appeared more pronounced for Hoechst 33342, PI, TO-PRO-3, and DRAQ5 (20-fold decrease as compared with control) and less pronounced for 7AAD and Vybrant Dye Cycle™ Green (five fold decrease as compared with control) (Supporting Information Fig. 9). These findings are in agreement with results obtained with plasmidic DNA, showing that AA used at high concentrations induces DNA damage also in the absence of copper ions.

AA/Cu effects on DNA stainability can be limited using hydroxyl radical scavengers such as PP, TPA, or cysteine solutions (Cys). As shown in Figure 3C, PP 2 mM, TPA 10 mM, and Cys 10 mM added together with AA/Cu mixture to PI pre-stained U-2 OS cells were able to prevent fluorescence bleaching, more efficiently at AA concentration of 10 mM and less at 50 mM AA. Similarly, addition of TPA or Cys to the reaction mixture improved the resolution of DNA content profiles, but BrdU or EdU detection by CuAAC was no more possible in these conditions (data not shown).

BrdU and BrdU/EdU Double Staining Following Ascorbate/Copper-Induced Oxidative Damage

Figure 4A shows a comparative experiment for BrdU detection by means of hydrochloride acid (HCl) DNA denaturation, Fenton reaction or EdU detection by CuAAC using BMA and BDA probes. U-2 OS cells pulsed with BrdU were denatured with HCl 3 N or treated with copper (2 mM) alone or together with AA 10 mM. EdU-pulsed cells were treated with 0.01 mM BMA/BDA or AdU (as negative control) in the presence of AA/Cu mixture. AA-treated cells BrdU/PI dot plots were similar to untreated samples (not shown), while treatment with cupric ions alone induced unspecific increase of fluorescence (autofluorescence). When BrdU pulsed cells were treated with AA/Cu mixture, the BrdU/PI profile was similar to that of cells processed by acidic denaturation, showing both about 40% BrdU positive cells. In these same conditions, pulse with EdU followed by click reaction with BrdU azides [13] gave a similar value of BrdU positive cells.

To evaluate the possibility to perform BrdU/EdU multiplexing, U-2 OS cells were labeled with BrdU for 20 min and released in BrdU free medium. At specific time points (8–18–27–40 h) cells were labeled with EdU for 20 min and fixed. BrdU was detected by anti-BrdU mAb, whereas EdU cells were stained using the Click-IT AlexaFluor 488 reagent. As shown in Figure 4B, BrdU/EdU incorporation can be simultaneously detected in the presence of AA/Cu mixture: cuprous ions (Cu+) catalyze CuAAC reaction with EdU and BrdU azide as previously described, and hydroxyl radicals cleave DNA and allow BrdU detection by specific antibodies.

Study of the Molecular Mechanism Underlying Fenton Reaction for BrdU Detection

In order to study the kinetics of BrdU detection, BrdU or EdU-pulsed U-2 OS cells grown in 96 well plates were incubated with increasing concentrations of AA and cupric ions, in the presence of BMA (CuAAC conditions). HCA results are shown in Figure 5A (BrdU) and Figure 5B (EdU). As expected, CuAAC labeling of cell pulsed with EdU cells depends on copper concentration: staining signal reached a maximum at 0.2–2 mM copper sulfate, even at low AA concentrations. A completely different effect was observed using AA/Cu mixtures as DNA damaging agent for BrdU detection. In these conditions, BrdU signal appeared at 1 mM AA and peaked at 10–50 mM AA, regardless of copper concentration. This demonstrates that AA is able to initiate the reaction even if in the presence of copper traces.

Hydroxyl radicals were detected in cell-free assay by enhanced luminol chemiluminescence (Fig. 5C) and cuprous ions by colorimetric assay with BCSA (Fig. 5D). According to the model depicted in Figure 1A, high amounts of cuprous ions, as detected by BCSA colorimetric assay, are produced from the reaction between cupric ions and AA, leading in turn to the production of hydroxyl radicals (revealed by luminol reaction). This reaction chain eventually generates massive DNA oxidative damage (Fig. 3), allowing BrdU detection by specific antibodies.

The degree of BrdU detection by HCA was correlated to parameters associated with Fenton reaction through hierarchical cluster analysis and heat map visualization. These parameters included: luminol chemiluminescence, BC-cuprous ions, PI fluorescence bleaching and CV% increasing, anti-8-hydroxy-deoxy-guanosine detection, and ARP signal induction (Fig. 5E). Results show that at high AA concentrations (10 mM, “G” and “M” conditions, or 50 mM, “F” and “L” conditions) BrdU positive cells display high fluorescence intensity, independently of copper concentration, whereas little to no effect was observed on BrdU signal using copper without reducing agent (“A,” “E,” and “I” conditions), as compared to negative controls (“O” condition). Experimental details and raw data are reported in Supporting Information (Fig. 12).

In order to verify “Fenton method” applicability to BrdU detection using multiplexed image-based HCA, U-2 OS cells were treated with PTX in 96 well plates for 24 h. Cells were pulsed with 30 μM BrdU or 10μM EdU 30 min before fixation. BrdU-incorporating cells were treated with nuclease and processed for BrdU immunofluorescence using anti-BrdU mAb and AlexaFluor 488 secondary antibodies as already described [16]. EdU-pulsed cells were processed for BMA/CuAAC reaction. Phospho-histone H3 was immunostained in these same samples using AlexaFluor 647-conjugated secondary antibody (red channel) and acquired together with BrdU/EdU (green channel) (Fig. 6).

image

Figure 6. ArrayScan biparametric analysis of phospho-histone H3 and BrdU incorporation using hydroxyl radicals method, nuclease treatment, or EdU coupled to click reaction with BMA. A) Representative images of U-2 OS cells asynchronously growing (NT) or treated with the indicated concentrations of PTX and stained with DAPI (blue), BrdU/EdU (AlexaFluor 488, green), and phospho-histone H3 immunofluorescence (AlexaFluor 647, red). Distinct dataset are reported in Supporting Information. B) Percentages of BrdU and EdU positive cells and C) phospho-histone H3 positive cells obtained from image analysis of samples shown in A). [Color figure can be viewed in the online issue which is available at wileyonlinelibrary.com.]

Download figure to PowerPoint

Results for untreated cells showed that the Fenton (BrdU + AA/Cu) and the click chemistry (EdU + BMA) approaches provide values of phospho-histone H3 positive and BrdU/EdU positive cells nearly identical to those obtained with the nuclease approach. As expected, PTX treatment induced a dose-dependent decrease in S-phase cells (Fig. 6B), as revealed by BrdU/EdU incorporation, and an increase of mitotic cells detected by phospho-histone H3 mAb (Fig. 6C). Hydroxyl radicals via Fenton reaction did not alter phospho-histone H3 detectability, just like the click chemistry approach, and thus the method here presented can be used to combine BrdU detection with immunofluorescence of labile epitopes, such as phosphoproteins.

Discussion

  1. Top of page
  2. Abstract
  3. Materials
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgment
  8. Literature Cited
  9. Supporting Information

The use of cuprous ions and AA has been described in FCM and HCA for the detection of cells that incorporated EdU or BrdU azides through a click chemistry approach [12, 13]. Nevertheless, the use of EdU has been so far limited to few applications in cellular assays due to some drawbacks: (i) EdU inhibits cell proliferation at lower doses than BrdU [13, 16, 34]; (ii) the CV of DNA profiles using Cu/AA catalysis is higher than those obtained with BrdU [35]; (iii) presence of some aspecific staining using alkyne labeling and CuAAC [36]. Cupric ions (Cu++) in the presence of AA or other reducing agents was demonstrated to cause DNA strand scission following site-specific formation of hydroxyl radicals (Fenton reaction), with production of cuprous ions (Cu+) and oxidized AA [5], as shown in Figure 1. The hydroxyl radicals may react by hydrogen abstraction, electron transfer, and addition reactions.

Here, we show that AA at high concentrations in the presence of copper salts is able to induce damage of both plasmidic DNA [7, 33] or genomic DNA extracted from cell lysates, generating small DNA fragments in both cases. Generation of hydroxyl radicals in proximity to DNA causes base damages and DNA strand breaks; this effect is proportional to the accessible surface areas of the hydrogen atoms of the DNA backbone [9]. To note, alternative bioconjugation methods can be employed to avoid undesirable effects on DNA damage, such as difluorocyclooctyne coupling using strain-promoted alkyne-azide cycloaddition chemistry [37] or redox-protected reactions [38, 39].

We verified the appearance of DNA base modification after exposure to AA/Cu mixtures of live cells, by means of 8-hydroxy-guanosine and ARP assays [22]. Experiments with pre-stained DNA using dyes including PI, Hoechst 33342, and 7AAD showed fluorescent bleaching using dyes with affinity for T and G sequences [40-42]. In order to explain the effect of Cu/AA treatment on DNA stainability (Fig. 3), we explored different experimental conditions for CuAAC reaction.

Several forms of DNA damage, including strand scission, base oxidation, and base liberation, can contribute to the loss of fluorescence. DNA damage after exposure to ROS generated by AA/Cu mixture resulted in decreased mean fluorescence because of physical dissociation of the dye from its binding site on DNA and fluorescence bleaching is an indicator of DNA damage [43]. The extent of DNA damage determined through FCM analysis of PI staining corroborated DNA gel electrophoresis data, which showed massive DNA fragmentation [32]. No base specificity for DNA cleavage was found, since DNA-binding fluorochromes, Hoechst 33342 (which preferentially binds A-T sequences), and PI (a DNA intercalating agent pairs without base specificity) were equally affected by the fluorescence decrease [32]. A similar effect on fluorescence bleaching was observed for TO-PRO-3 and DRAQ5, and to a lesser extent for 7-AAD and Vybrant Dye Cycle Green. Since anthraquinone and diazo dyes can be easily oxidized by hydroxyl radicals, it can be postulated that the observed fluorescence quenching represents the result of DNA damage (loss of stainability) and dye alteration (fluorescence bleaching) [44].

Most Fenton reactions generate non specific DNA cleavage; however, there is evidence that cuts preferentially occur at thymidine residues [33].

Cleavage effects by copper chemistry, however, are well known: copper complexes, such as bis-(1,10-phenanthroline)copper(I) complex, are responsible for direct DNA strand scission [45, 46] through hydrogen atom H-1′ abstraction from the deoxyribose moiety. Moreover, hydroxyl radicals are employed for RNA and DNA footprinting assays [47-49], where the [Fe(EDTA)]2− complex is oxidized by hydrogen peroxide in the presence of AA, to abstract hydrogen from each deoxyribose carbon of B-form DNA. Hydroxyl radicals are highly reactive, with a half-life in aqueous solution of less than 1 ns: when produced in vivo, they react very close to the site of formation such [50]; cupric ions were found to interact with phosphate groups of DNA backbone [51], and specific binding at N-7 guanine bases over adenine bases was demonstrated in double-stranded DNA [52]. Traces of transitions metals such as copper and iron are found in buffer solutions [53], cell culture media, syringes and laboratory tools, making possible that Fenton reactions might occur in intact cells in the presence of AA. In addition to copper and iron, the list of metals potentially involved in Fenton reactions include chromium, titanium, cobaltum, rhodio, nickel, manganese, and palladium [33]. Rare earth from the lanthanide group, such as cerium and terbium, possess similar redox proprieties and have already been described to cause DNA cleavage in live cells [54, 55].

Additional methods to introduce DNA strand breaks for BrdU incorporation detection include ultraviolet light [56, 57] (DNA photolysis). Two disadvantages of DNA photolysis are the need of transilluminators or sensitizing agents such as Hoechst 33342 and the poor reliability of UV lamps, as compared to approaches like Fenton or click chemistry. On the other hand, Fenton reaction has the drawback of causing massive DNA degradation: this implies that experimental conditions should carefully set to allow simultaneous DNA content analysis using fluorescent dyes.

Our results are in agreement with a those obtained by Koberna and coworkers [58], who also demonstrated oxidative attack at the deoxyribose moiety by copper(I) in the presence of oxygen and AA and exploited Fenton reactions to track BrdU incorporation. In this article, we examine in depth the molecular mechanism of Fenton reactions in intact cells and show that they can occur as side reaction of click CuAAC cycloaddictions. Moreover, we demonstrate that hydroxyl radicals' reaction can be combined with click chemistry for BrdU/EdU pulse-and-chase experiments and can be employed in multiplexed immunofluorescence assays both for HCA and for FCM.

It is known that transition metals play a biological role in ROS signal transduction [8], but there is increasing evidence that they might represent a valuable tool in bioinorganic research and many bioanalytical applications [59]. Additional analytical approaches involving the use of transition metals to generate hydroxyl radicals include nucleic acid-protein interaction studies [49] and pure proteomics [60]. Here, we demonstrate that Fenton reactions occur under certain circumstances alongside click chemistry-based BrdU analysis, providing a novel tool for BrdU/EdU pulse-and-chase experiments and for multiplexed analysis of DNA replication and immunofluorescence of labile epitopes.

Literature Cited

  1. Top of page
  2. Abstract
  3. Materials
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgment
  8. Literature Cited
  9. Supporting Information

Supporting Information

  1. Top of page
  2. Abstract
  3. Materials
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgment
  8. Literature Cited
  9. Supporting Information

Additional Supporting Information may be found in the online version of this article.

FilenameFormatSizeDescription
cytoa22343-sup-0001-suppinfo.doc55KSupporting Information
cytoa22343-sup-0002-suppinfo.doc2195KSupporting Information

Please note: Wiley Blackwell is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.