Characteristics of a Novel Deep RedInfrared

8/13/2019 Characteristics of a Novel Deep RedInfrared

1/12

Characteristics of a Novel Deep Red/InfraredFluorescent Cell-Permeant DNA Probe, DRAQ5, inIntact Human Cells Analyzed by Flow Cytometry,

Confocal and Multiphoton Microscopy

Paul J. Smith,1* Nicola Blunt,1 Marie Wiltshire,1 Terence Hoy,2 Paul Teesdale-Spittle,3

Michael R. Craven,3 James V. Watson,4 W. Brad Amos,4 Rachel J. Errington,5 andLaurence H. Patterson6

1Department of Pathology, University of Wales College of Medicine, Cardiff, United Kingdom2Department of Haematology, University of Wales College of Medicine, Cardiff, United Kingdom

3De Montfort University, Leicester, United Kingdom4MRC Centre, Cambridge, United Kingdom

5Department of Medical Biochemistry, University of Wales College of Medicine, Cardiff, United Kingdom6The School of Pharmacy, University of London, London, England

Received 24 March 2000; Accepted 20 April 2000

Background: The multiparameter fluorometric analysisof intact and fixed cells often requires the use of a nuclearDNA discrimination signal with spectral separation from

visible range fluorochromes. We have developed a noveldeep red fluorescing bisalkylaminoanthraquinone, DRAQ5(Exmax646 nm; Emmax681 nm; Emrange665800 nm),

with high affinity for DNA and a high capacity to enterliving cells. We describe here the spectral characteristicsand applications of this synthetic compound, particularlyin relation to cytometric analysis of the cell cycle.Methods: Cultured human tumor cells were examined forthe ability to nuclear locate DRAQ5 using single and mul-tiphoton laser scanning microscopy (LSM) and multipa-rameter flow cytometry.Results: Multiparameter flow cytometry shows that thedye can rapidly report the cellular DNA content of live andfixed cells at a resolution level adequate for cell cycleanalysis and the cycle-specific expression of cellular pro-teins (e.g., cyclin B1). The preferential excitation ofDRAQ5 by laser red lines (633/647 nm) was found to offera means of fluorescence signal discrimination by selectiveexcitation, with greatly reduced emission overlap withUV-excitable and visible range fluophors as compared

with propidium iodide. LSM reveals nuclear architectureand clearly defines chromosomal elements in live cells.DRAQ5 was found to permit multiphoton imaging of nu-clei using a 1,047-nm emitting mode-locked YLF laser. Theunusual spectral properties of DRAQ5 also permit live cellDNA analysis using conventional 488 nm excitation andthe single-photon imaging of nuclear fluorescence usinglaser excitation between 488 nm and low infrared (IR; 780nm) wavelengths. Single and multiphoton microscopystudies revealed the ability of DRAQ5 to report three-dimensional nuclear structure and location in live cellsexpressing endoplasmic reticulum targeted-GFP, Mito-Tracker-stained mitochondria, or a vital cell probe for freezinc (Zinquin).Conclusion: The fluorescence excitation and emissioncharacteristics of DRAQ5 in living and fixed cells permitthe incorporation of the measurement of cellular DNAcontent into a variety of multiparameter cytometricanalyses. Cytometry 40:280 291, 2000. 2000 Wiley-Liss, Inc.

Key terms: flow cytometry; infrared imaging; confocalmicroscopy; multicolor analysis; cell-permeant DNA dye

The ability to identify cell cycle position through DNAcontent analysis is fundamental to many studies on cellcycle-regulated protein expression and the assessment ofperturbations in cell cycle traverse. We sought to deter-mine whether the fluorescence excitation and emissioncharacteristics of DNA-binding anthraquinone derivativescould provide sufficient discrimination of cellular DNAcontent in multiparameter fluorescence studies. The an-thraquinones are a group of synthetic DNA-binding agents

(1) that are structurally related to the DNA intercalatinganthracycline antibiotics (2). As such, they can penetrateintact cellular membranes. The principal member of this

Grant sponsor: Association for International Cancer Research (AICR);Grant sponsor: UK Medical Research Council.

*Correspondence to: Prof. P.J. Smith, Department of Pathology, Univer-

sity of Wales College of Medicine, Heath Park, Cardiff, CF4 4XN, UK.E-mail: [email protected]

2000 Wiley-Liss, Inc. Cytometry 40:280 291 (2000)


2/12

group is the anticancer anthraquinone, mitoxantrone,which has been found to demonstrate pH-dependent butextremely weak fluorescence (3). Intact mitoxantrone-treated human and murine cells exhibit low levels offluorescence (4,5). Importantly, 647 nm wavelength laserlight can be used to optimally excite mitoxantrone andrelated anthraquinones with fluorescence detection intothe low infrared regions of the emission spectrum (6),providing a rationale for the use of such agents in mul-tiparameter analayses involving visible range fluoro-chromes.

The fluorochrome-based quantification of DNA contentin intact cells in order to ascribe cell cycle age is animportant component of multiparameter analytical proce-dures. There is a growing number of DNA-binding fluoro-chromes available with convenient Exmax/ Emmaxvaluesextending from the ultraviolet (UV) to the visible regionsof the spectrum, including Hoechst 33258 and Hoechst33342, DAPI, acridine orange, ethidium bromide, and pro-pidium iodide (PI; for reviews, see refs. 79). DNA-inter-calating cyanine fluorochromes, such as TOTO 3, havesignatures in the high red region of the visible spectrum(Exmax/Emmax of 642 nm/661 nm) and have been used inDNA content analysis (10). Other red-shifted DNA fluo-rochromes include aminoactinomycin D (7-AAD; Emmax610 nm wavelength; 11) and the cell-permeant nucleicacid stain, LDS-751(12). DNA-bound LDS-751 has anExmaxvalue of543 nm and exhibits a far-red emissionmaximum (13,14), but does not discriminate cellular DNAcontent at the resolution required for cell cycle analysis.The recent introduction of the cell-permeant SYTO nu-cleic acid stains (e.g., SYTO 17; Molecular Probes, Eugene,OR; 14,15) has attempted to provide reagents with con-

venient red-shifted spectral characteristics.Given the potential for side chains of anthraquinones

to modify DNA binding and intracellular disposition, wesought to identify agents that will provide high nuclearDNA discrimination (6,16,17) in intact and fixed cells.

We have screened a range of substituted anthraqui-nones on the basis of their ability to bind to DNA insolution, retain high membrane penetrance, and inter-act with cellular DNA in intact cells. The synthesisprogram and in vitro screen identified an agent, DRAQ5(deep red-fluorescing bisalkylaminoanthraquinone numberfive), that met these criteria (18). DRAQ5 is a 1,5-bis{[2-(methylamino)ethyl]amino}-4,8-dihydroxy anthracene-9,10-dione. DRAQ5 appears to achieve nuclear discrimi-nation by its high affinity for DNA. Excitation at 647 nm

wavelength, close to the Exmax, produces a fluorescencespectrum extending from 665 to beyond 780 nm wave-lengths. Thus, the emission spectrum is beyond that offluorescein, phycoerythrin, Texas Red, and cyanine 3 (Cy3). We describe the spectral characteristics of DRAQ5 andexplore the potential applications of this novel DNAprobe for multiparameter analysis of living and fixed cellsusing flow cytometry and confocal laser scanning micros-copy (CLSM).

MATERIALS AND METHODSCell Culture

The human melanoma cell line, A375, was kindly pro-vided by Dr. R. Johnson (University of Cambridge, UK). Itwas grown as asynchronous cultures in Eagles minimumessential medium supplemented with 10% fetal calf serum

(FCS), 1 mM glutamine, and antibiotics and incubated at37C in an atmosphere of 5% CO2 in air. For imagingexperiments, cells were grown at a density of 5 104

cells per well as a monolayer on autoclaved glass cover-slips in six-well plates for 48 h prior to treatment. HL-60(human promyelocytic leukemia cell line) and SU-DHL-4(human B-cell lymphoma cell line) were grown as suspen-sion cultures in RPMI medium with 10% FCS, 1 mM glu-tamine, and antibiotics and incubated at 37C in an atmo-sphere of 5% CO2in air. For flow cytometry experiments,asynchronously growing suspension cultures were dilutedto 2.54 105 cells/ml at 2 h prior to drug treatment. Cellcycle-perturbed populations were obtained by treatingSU-DHL-4 cells with etoposide (VP-16-213) at 0.25 M for

18 h. Cells were treated with DRAQ5 and fluoresceindiacetate (FDA) as described above. Cell concentrations

were determined using a Coulter counter and cell cycledistribution determined using an algorithm for the normaldistribution of fluorescence intensity profiles for fluoro-chrome-stained G1 and G2 cells. GFPCRT (calreticulin ERsignal peptide/GFP [s65t]/calreticulin/Kdel) stable trans-fectants of HeLa cells (kind gift from D.H. Llewellyn,University of Wales College of Medicine [UWCM], UK)

were cultured under G418 selection as previously re-ported (19) in Dulbeccos modified Eagle medium(DMEM) supplemented with 10% FCS, 2 mM L-glutamine,and 100 g/ml penicillin/streptomycin. These cells were

seeded onto coverslips and mounted into a microscopeobservation chamber.

Reagent Preparation and Treatments

DRAQ5 was synthesized using the principles describedpreviously (19) and is a dark blue crystalline solid ofmolecular weight 412.54. The dye is stable at room tem-perature but was routinely stored at 4C, without freez-ing, as an acidified aqueous stock solution of 10 mM.DRAQ5 dilutions were prepared in phosphate- bufferedsaline (PBS) and added directly to cultures. FDA (KochLight Laboratories) was prepared as a stock solution of 12mM in acetone and stored at 20C. Cells were treated

with 0.2 M FDA for 10 min at 37C either alone or aftera 50-min exposure to DRAQ5. Likewise, DRAQ5-treatedcells were labeled with rhodamine 123 (laser grade;Kodak, Rochester, NY) at 2 g/ml culture medium for 10min, prior to analysis. Suspension cultures were analyzedby flow cytometry without washing, whereas attachedcells were mounted in fresh PBS for microscopy. Zinquinethyl ester [Zinquin E, [2-methyl-8-(4-methylphenylsulfo-nylamino) quinolinyl] oxyacetic acid ethyl ester] (AlexisCorporation, Nottingham, UK) was stored as a 5-mM stocksolution in ethanol at 4C. Where indicated, attached cells

were fixed with 70% methanol at -20C for 10 min prior to

281DEEP RED FLUORESCENT CELL-PERMEANT DNA DYE


3/12

rehydration and staining with ethidium bromide at 5g/ml for 10 min in the presence of 5 mg/ml RNase A(Sigma, Poole, UK). Where indicated, attached GFPCRTHeLa cells were incubated with 5 ng/ml Mitotracker or-ange CMTMRos (Molecular Probes) at room temperaturefor 5 min, followed by extensive washes in HEPES/Trisbuffer. These cells were then loaded with 10 M DRAQ5.

Spectral Analysis of DRAQ5

Absorbance spectra were obtained using a Perkin-Elmerlambda 16 UV spectrometer (Perkin-Elmer, Beaconsfield,UK) and a 10-M solution of agent dissolved in dichlo-romethane and measured in a 1-cm path length quartz-silica cuvette. Fluorescence spectra for 0.8 ml solution ofagent in a 1-cm path length semimicro quartz silica cu-

vette were determined using a Perkin Elmer LS50B spec-trofluorometer with slit widths set at 5 nm. The spec-trofluorometer was equipped with a red-sensitivephotomultiplier tube (PMT; type R928; Hamamatsu Pho-tonics KK, Japan). Data were accumulated for eight scans

for each condition. Samples were excited at 647 nm wave-length and emission scanned as indicated. Both agent andcalf thymus DNA were prepared in buffer (8 mM Tris base,0.05 M NaCl, pH 7.2, in distilled water). DNA-drug fluo-rescence was measured by the addition of microliter vol-umes of concentrated calf thymus DNA solutions to thecuvette with mixing. The spectra shown were correctedfor the buffer background and for the spectral sensitivityof the PMT where indicated. Rhodamine 123 spectra weregenerated in DNA-binding buffer using either 488/5 nmexcitation or monitoring emission at 530/5 nm. PI emis-sion spectra were generated using 488/5 nm excitation.Previously published excitation and emission spectra (20)

were obtained from original source files and normalized

for peak intensity and are reproduced for comparativepurposes. These spectra represent fluorochrome conju-gates: fluorescein isothiocyanate (FITC)-conjugated don-key anti-rabbit IgG, donkey anti-rabbit conjugated withTexas Red sulfonyl chloride, and donkey anti-rabbit IgGconjugated with Cy 5.18.

Loading Conditions for Zn2 and Zinquin

Cells were grown in asynchronous culture to approxi-mately 6 105/ml, resuspended in loading buffer (Hanksbalanced salt solution lacking Ca2 and Mg2 [HBSS- -]supplemented with 20 mM HEPES, pH 7.4), and held at37C. Cells were then exposed to 25 M Zn2 with

sodium pyrithione (1 M) and incubated for 30 min at37C. Following Zn2 or control loading, cells were

washed twice rapidly in loading buffer (centrifugation for30 s at 6,500 rpm using a MicroCentaur microfuge) andresuspended at 6 105/ml. Suspensions were exposed to25 M Zinquin ester and incubated for 30 min at 37C andmounted on microscope slides for imaging.

Immunofluorescence

Cyclin B1 staining was achieved using the method out-lined previously (21). Briefly, attached cells were har-

vested and fixed in 80% ethanol for 2 h, washed in PBS,

and the pellet resuspended in 2 ml of 0.5% Triton X-100 inPBS (filtered) for 5 min followed by a wash in PBS. The cellpellet was resuspended in mouse anti-human cyclin B1antibody (GNS1; Santa Cruz Biotechnology, Heidelberg,Germany; 1:150 dilution) in 1% filtered bovine serumalbumin (BSA) or anti-p34Cdc-2 (H-297; rabbit polyclonalIgG) in PBS at 100 l/106 cells and incubated at room

temperature for 1 h. Cell suspensions were washed with 2ml of PBS containing 1% BSA and the pellet resuspendedin either AMCA- or FITC-conjugated goat anti-mouse IgGantibody (Caltag, San Francisco, CA) diluted 1:100 in 1%BSA in PBS (100 l/106 cells) and incubated for 30 min atroom temperature.

Flow Cytometry

Cells were analyzed using one of four flow cytometersaccording to the excitation requirements.

Cytometer A: single-beam high-power 647-nmkrypton laser excitation. The system was a custom-

built cytometer incorporating an Innova 3000K kryptonlaser (Coherent, Palo Alto, CA) tuned to the 647-nm line.Forward light scatter (FSC), 90 light scatter, and fluores-cence emissions were collected for 1 104 cells using the90 light scatter parameter as the master signal. The opti-cal system permitted the analysis of various fluorescenceemission wavelengths including 715 and 780 nm flu-orescence. Forward and 90 light scatter were analyzedfor the identification of cell debris. Laser power was set at200 mW and linear amplifiers were used for the fluores-cence signals. The analysis optics included a 675-nm colddichroic mirror. Filters were supplied by Melles Griot(Irvine, CA).

Cytometer B: dual-beam low-power 633-nm /high-

power 488-nm laser excitation. The system was aFACS 440 cell sorter (Becton Dickinson, Cowley, UK)incorporating a Spectra Physics argon ion laser (maximum500 mW output), tuned to the 488-nm line (100 mWoutput), and a secondary Spectra Physics 156 helium-neonlaser emitting at 633 nm (emitting 5 mW), with a tem-poral beam separation of about 30 s. FSC, 90 lightscatter, and fluorescence emissions were collected for 1104 cells using the FSC parameter as the master signalfrom the primary 488-nm beam. Side scatter was collectedthrough a 488/10-nm band-pass filter. The analysis opticsincluded a cold dichroic mirror (transmitting 675 nm);fluorescence from fluorescein excited by the 488-nm

beam detected at a PMT guarded by a 535/15-nm band-pass filter with no signal delay; and a red-sensitive PMTwith an appropriate delay, additionally guarded by a620-nm long-pass filter, to detect the transmitted beam ofDRAQ5-associated fluorescence at wavelengths beyond675 nm. Forward and 90 light scatter were analyzed toexclude any cell debris. All parameters were acquired at a256-channel resolution with Consort 30 software (BectonDickinson) and subsequently analyzed with WinMDI soft-

ware (J. Trotter, La Jolla, CA). The system employed thesame analysis optics when used in the single 488-nm beammode but with no signal delay for the red-sensitive PMT.

282 SMITH ET AL.


4/12

Cytometer C: Triple-beam medium-power 633-nm/ medium-power 488-nm/ medium-power multi-line UV laser excitation. The system was a FACS Van-tage cell sorter (Becton Dickinson) incorporating a Coher-ent Enterprise II laser simultaneously emitting at multilineUV (350360- nm range) and 488-nm wavelengths withthe beams made noncolinear using dichroic separators.

Beam-combining optics were used to align the UV beamwith that emitted by a Spectra Physics 127-35 helium-neonlaser (maximum 35 mW output) emitting at 633 nm witha temporal separation of about 25 s from that of theprimary 488- nm beam. FSC, 90 light scatter, and fluores-cence emissions were collected for 1 104 cells using theFSC parameter as the master signal from the primary488-nm beam. Side scatter was collected through a 488/10-nm band-pass filter. The analysis optics were (1) pri-mary beam-originating signals analyzed at FL1 (FITC filter;barrier filter of 530/30 nm) after transmission at SP610 andSP560 dichroics, at FL2 (barrier filters of 585/42 or 575/26nm) after transmission at SP610 and reflection at SP560

dichroics, or at FL3 (barrier filter of LP715 nm) afterreflection at a SP610 dichroic; (2) delayed beam-originat-ing signals analyzed at FL4 (barrier filter of LP695 nm) orat FL5 (barrier filter of DF424/44 nm) after transmission orreflection at a LP640 dichroic, respectively. Forward and90 light scatter were analyzed to exclude any cell debris.

All parameters were analyzed using CellQuest software(Becton Dickinson).

Cytometer D: single-beam, low-power 488-nm ex-citation. A FACScan system (Becton Dickinson) incorpo-rating an argon ion laser (maximum 15 mW output) tunedto the 488-nm line was used. FSC, 90 light scatter, andfluorescence emissions were collected for 1 104 cellsusing the FSC parameter, where appropriate, as the mas-

ter signal. The standard analysis optics provided the FL1(green), FL2 (orange), and FL3 (red; 650 nm) PMT pa-rameters, with pulse analysis performed on the FL3 orig-inating signals. Pulse height measurements are displayedfor FL1 and FL2 parameters, with pulse area for the FL3parameter. All parameters were analyzed using CellQuestsoftware.

CLSM and Multiphoton Laser Scanning Microscopy

Three imaging systems were used.Comparison of 488, 568, and 647-nm excitation of

DRAQ5. The equipment used was a Leica confocal im-aging system (TCS 4D; LaserTechnik Gmbh, Germany)

scanner coupled to a Leitz DM R microscope and operat-ing with an Ominchrome argon/krypton laser. The laserprovided emission lines at 488, 568, and 647 nm with

variable power. Coverslip cultures were washed briefly inPBS, mounted in inverted positions on glass slides, thecoverslips being supported at the edges by a piping ofpetroleum jelly to prevent the cells from being com-pressed. The slides were examined immediately using100 or 40 oil immersion objective lenses withmidrange pinhole and PMT gain settings. Emission detec-tion for DRAQ5 was 665 nm wavelength. Gain settings

were adjusted such that the most fluorescent drug-treated

sample gave pixel intensities just below saturation. Theblack level/offset was adjusted to give effectively zerobackground (less than four for pixel value) after 16 linenoise filtration of images for untreated controls. Using thisapproach, the untreated controls showed minimalautofluorescence and gave no discernible image, obviatingthe need for a background correction. Saved images were

analyzed using IP Lab Spectrum Image analysis software(Signal Analytics, Vienna, VA).

Multiparameter analysis of live cells. The scanningunit was a BioRad 1024MP system (BioRad Microscience,Hemel Hempstead, UK) linked to a Zeiss Axiovert 135inverted microscope. The system was used in two modes.First, three-dimensional (x,y,z) images were collected us-ing a confocal configuration. All three fluors (GFPCRT,Mitotracker orange, and DRAQ5) were excited simulta-neously with 488, 568, and 647-nm lines of a krypton-argon laser, respectively, and the corresponding emissionscollected at 522/35, 598/40, and 680/32 nm. Second, thesystem was operated in combined multiphoton and single-

photon mode. This was achieved using a 100-fs pulsedtitanium-sapphire laser (Coherent) at 780 nm for two-photon excitation of the Zn2-sensitive probe, Zinquin E,and one-photon excitation of DRAQ5.

Two-photon excitation of DRAQ5. The scanningunit was a BioRad 1024MP system incorporating a YLFmode-locked femtosecond-pulsed laser providing two-photon excitation of DRAQ5 at 1,047-nm wavelength ex-citation at 15 mW, with detection of fluorescence in thefar red. Images of ethanol-fixed SU-DHL-4 cells in equilib-rium with 20 M DRAQ5 were gained using a 60 N.A.1.4 oil objective, a zoom of 1.9, and a Kalman filteringaverage of 37 scans.

RESULTSSpectral Characteristics and Interaction of DRAQ5

With DNA

Figure 1 shows the UV-visible absorbance spectrum forDRAQ5 in phosphate buffer at pH 7.4. The spectrum gavemaxima at 622 and 676 nm. In addition, other maximaoccurred at 240 and 314 nm. The extinction coefficientat the 676-nm wavelength was determined as 20,949cm-1mol-1.

Figure 2a shows the excitation and emission spectra forDRAQ5 and another vital cell dye, rhodamine 123, forcomparison. The excitation spectrum for DRAQ5 corre-sponds to the absorbance peak, with fluorescence emis-

sion extending from 665 to beyond 780-nm wavelengths.There is a clear preferential excitation of DRAQ5 at wave-lengths in the red region of the spectrum (Exmax 646nm). Despite the relative insensitivity of the fluorometerdetector, it was possible to reveal excitation of DRAQ5 at

wavelengths extending into the 485520-nm region, witha20-fold reduction in peak fluorescence. Comparing theExmax/Emmaxvalues (Fig. 2b) for DRAQ5 (646/681 nm)

with those of other fluorochromes (20) indicates that thisdye is similar to Cy 5.18 (649/666 nm), but is distinct fromTexas Red (589/610 nm), rhodamine 123 (498.5/524.5nm), and FITC (494/517 nm).



5/12

The shape of the emission spectrum of DRAQ5 can bereduced to the wavelengths (681, 660.5, 762.5 nm) asso-ciated with the maximum fluorescence (Emmax) and the50% maximum fluorescence on the low (Em50%low) orhigh (Em50%high) wavelength side of peak emission, re-spectively. The presence of calf thymus DNA (697, 674.5,778 nm) results in a red-shift of 16 nm, whereas thespectrum of DRAQ5 is red-shifted by up to 58 nm com-pared with that of PI (665, 602.5, 732.5 nm).

Thus, DRAQ5 represents a dye that is optimally excitedat the 647-nm wavelength, an emission line of a krypton-argon laser. Its far red/low infrared emission spectrumoverlaps with that of PI, but is significantly red-shifted inthe visible region. A further red shift occurs in the pres-ence of DNA, although there is no indication of fluores-cence enhancement.

CLSM Analysis of Nuclear-DRAQ5 Fluoresence inViable Cells

Visual examination of cells excited with green light,under normal epifluorescence microscopy conditions, re-

vealed an extremely faint dark-red fluorescence associated

with DRAQ5-treated intact cells. Using CLSM with optimal647-nm excitation, there was clear demonstration of nu-clear-located fluorescence. This was quite different fromother anthraquinone- and anthracycline-based agentsscreened, which produced both nuclear and cytoplasmicsignals. However, DRAQ5 is present in the cytoplasm ofliving cells because we have occasionally observed smallintracellular blue deposits of DRAQ5 that were not fluo-rescent. Such deposits were not observed in fixed mate-rial. We suggest that intracellular drug deposits involvestacked nonfluorescent molecules effectively not in freesolution. DRAQ5-treated viable cells display clear nuclear

architecture and the definition of the edges of nucleolarand nuclear membrane regions.

To gain some insight into the dependence of DRAQ5-nuclear fluorescence on the excitation wavelength andthe spectral separation of its fluoresence signal from thatof a reference DNA probe, we have compared live cellsstained with DRAQ5 and ethanol-fixed cells stained with

ethidium bromide. The results are shown in Figures 3af.The dynamic range of the CLSM at either 488 or 568-nmexcitation was optimized with respect to the fluorescenceof ethidium bromide, whereas imaging at 647-nm excita-tion was optimized for DRAQ5 fluorescence. The resultsindicate that 488 and 647-nm excitation conditions can beused to exclusively image either ethidium bromide orDRAQ5 staining, respectively. Extension studies (data notshown) have indicated that DRAQ5 could also be used toimage fixed cells, with retention of much of the nucleararchitecture observable in intact, viable cells. Dual imag-ing was also extended to a vital dye capable of definingcytoplasmic organelles (rhodamine 123) in DRAQ5-cola-

beled cells with evidence that mitochondrial form andfunction were not affected significantly by DRAQ5 vitalstaining of nuclei when compared with controls (data notshown).

Single and Multiphoton Excitation Applications ofDRAQ5 in Laser Scanning Microscopy

We sought to confirm the ability of DRAQ5 to mark thenuclear compartment of a living cell and aid the identifi-cation of the subcellular location of other intracellularfluorescent molecules. Figures 4af show the multiparam-eter analysis of live cells using simultaneous combined488/568/647-nm single-photon excitation of the intracel-

lular reporter molecules MitoTracker (a mitochondrialprobe), GFP (linked to calreticulin and providing a markerfor the endoplasmic reticulum; 19), and DRAQ5. There isclear separation of the respective signals, permiting thenonnuclear location of labeled mitochrondria to be con-firmed together with the occasional intrusion of GFP-labeled endoplasmic reticulum as tube-like structures intothe nuclear mass. Figures 4df show the clear delineationof mitotic figures in cells that have disassembled theirnuclear membranes.

The use of intracellular probes to detect transientchanges in the internal environment requires scanningmethods that reduce the level of beam damage to a sam-

ple, particularly if a UV-excitable probe is to be used. Wehave assessed the ability of DRAQ5 to be used in combi-nation with a live cell UV-excitable probe, employing atwo-photon excitation of that probe with a femotosecond-pulsed IR laser beam. The probe Zinquin [(2-methyl-8-p-toluenesulphonamido-6-quinolyloxy)acetic acid] has pre-

viously been used to detect free Zn2 in intact cells (22).Cells treated with Zinquin-ester can be loaded with theprobe through intracellular esterase conversion. The Zin-quin-acid form can detect free or loosely bound intracel-lular Zn2 through the formation of highly fluorescentcytoplasmic Zinquin-Zn2 complexes (350360 nm

FIG. 1. UV-visible absorbance spectrum for DRAQ5 (10 M in dichlo-romethane) with inset of DRAQ5 chemical structure.

284 SMITH ET AL.


6/12

Exrange; 485 nm Emmax). Using ionophore-mediatedZn2 loading of live cells to enhance the Zinquin-Zn2

signal in SU-DHL-4 cells, brief exposure to DRAQ5 permitsthe identification of the nuclear compartment and con-firms the cytoplasmic location of the Zinquin-Zn2 com-

plexes (Fig. 5ac). This was determined by the dual im-aging of two-photon excited Zinquin-Zn2 complexes atthe 780-nm wavelength (i.e., approximately equivalent toexcitation 390 nm) and the suboptimal single-photon ac-tivation of DRAQ5.

FIG. 2. Spectral characteristics ofDRAQ5. a: Corrected fluorescenceexcitation (closed symbols) and emis-sion (open symbols) spectra of rhoda-mine 123 (open and closed triangles,0.5 g/ml, 488-nm excitation) andDRAQ5 (open and closed circles, 20M, 647-nm excitation).b:Corrected

fluorescence emission spectra DRAQ5and other fluorochromes analyzed inthe current study or using publisheddata obtained from original sourcefiles (20). Open triangle, FITC (ref.20); closed square, Cy 5 (ref. 20);closed triangle, Texas Red (ref. 20);open circle, DRAQ5 (20 M, 647-nmexcitation); closed circle, DRAQ5plus 1,280 M calf thymus DNA(647-nm excitation); open square, PI(5 g/ml, 488-nm excitation).

FIG. 3. Spectral characteristics of DRAQ5-associated DNA fluorescence detected by CLSM. ac: Excitation at 488, 568, and 647-nm wavelengths,respectively, for viable attached human A375 melanoma cells. df:Excitation at 488, 568, and 647-nm wavelengths, respectively, for ethanol-fixed cellsstained with ethidium bromide. Dynamic range for imaging fixed for a given excitation wavelength. Images are 100 100 m.



7/12

We have also observed that it was possible to obtainlow-intensity fluorescence images of DRAQ5-treated nu-

clei by single-photon 488-nm excitation by increasing theamplifier gain settings used to generate the image in Fig-ure 3a. This suggested that IR wavelengths 1,000 nm

may be capable of generating two-photon excitation ofDRAQ5. Figure 5d shows that DRAQ5-nuclear fluores-

cence can be generated by excitation at 1,047 nm using amode-locked femtosecond-pulsed YLF laser.

Discrimination of Nucleated Cells by FlowCytometry

Despite DRAQ5 excitation being optimal at the 647-nmwavelength, preliminary studies indicated that the probe

FIG. 4. Multiparameter imaging by CSLM of GFPCRT HeLa cells (b,e) labeled with MitoTracker (a,d) and DRAQ5 (c,f). Scale bars 10 m for panels acand df.

FIG. 5. Multiphoton imaging for DRAQ5-stained SU-DHL-4 cells. a,b: Two-photon excitation of the intracellular Zn2 probe Zinquin (a) and simultaneous

single-photon excitation of DRAQ5 (b). c: DRAQ5-nuclear fluorescence in ethanol-fixed cytocentrifuge preparations in equilibrium with 20 M DRAQ5generated by two-photon excitation at 1,047-nm wavelength. Scale bars 10 m

286 SMITH ET AL.


8/12

could be suboptimally excited at lower visible wave-lengths, including 488, 514, and 633 nm. We have soughtto assess DRAQ5 as a DNA probe for use in flow cytometrycomparing various laser configurations.

Figure 6 shows that when a low-power HeNe laser(cytometer B) is used, complete separation does not occurfor autofluorescence and DRAQ5 signals for viable HL-60cells. However, even under these limiting excitation con-ditions, the 633-nmderived DRAQ5 signal shows a clearlinear dose response (Fig. 2) down to approximately 2.5M, which is comparable to that obtained for optimal647-nm excitation (using cytometer A) and detection at

wavelengths 780 nm.Figure 7a,b shows that a low-power HeNe laser (cytom-

eter B) can be used to identify DRAQ5-associated fluores-cence in fluorescein-loaded cells analyzed in a dual-beamconfiguration. Figures 7c,d show that coexcitation ofDRAQ5 and fluorescein is also possible using a singlebeam of 488 nm wavelength (cytometer B). There is clearseparation of signals, due to the distinct, nonoverlappingspectra, despite the low intensity signal derived fromsuboptimal excitation of DRAQ5.

DNA Content Analysis by Flow Cytometry

Given the imaging results, it was important to deter-

mine whether the nuclear-DRAQ5 fluorescence could re-port cellular DNA content. Our approach was to obtainreference distribution profiles of ethidium bromide-stained asynchronous and cell cycle-perturbed cell popu-lations using dual-beam flow cytometry (stream-in-air sys-tem) and compare these results with the correspondingdistribution profiles obtained for DRAQ5-stained live SU-DHL-4 cells. We used the 488-nm laser line as the primaryreference beam for both agents. Figures 8af show thecomparison of DNA content analysis by dual-beam flowcytometry for ethidium bromide-stained (488 nm versusUV excitation) and DRAQ5-stained (488 nm versus medi-

um-power 633-nm excitation) nuclei. DNA content distri-butions for ethidium bromide-stained nuclei are refined byRNase A digestion and the ability to equilibrate DNA withthe dye. The dot plot (Fig. 8a) and distribution profiles(Figs. 8b,c) show the ability of 488 nm and UV excitationto report DNA content. Viable SU-DHL-4 cells exposed to20 M DRAQ5 for 5 min generated a similar pattern, but

less refined distribution profiles (Figs. 8df) to that forethidium bromide for both the 488 and 633-nmgener-ated signals. Shifting cellular DNA content by exposure toa G2/M arresting agent was clearly detectable by bothagents, the corresponding data being shown in Figures9af.

The data presented in Figures 8 and 9 indicate thatsuboptimal 488-nm excitation of DRAQ5-nuclear fluores-cence can report changes in cell cycle distribution. Wehave sought to confirm the ability of a low-power 488-nmlaser incorporated into a standard benchtop flow cytom-eter to discriminate DNA content in the same cells ana-lyzed in Figures 8 and 9. Figures 10ad show the FACScan

analyses verifying the similarity between the patterns ob-tained by DRAQ5 and ethidium bromide. Figure 10 alsoshows the results of analyzing the channel positions of themean value for the G1 and G2/M peaks (23) together withtheir corresponding coefficient of variation (CV) values.The DRAQ5-stained cells show wider CV values, but cor-rectly distinguish the difference in DNA content betweenG1 and G2/M cells.

Multiparameter Cell Cycle Analysis

Given the ability of 488 or 633-nm laser lines to provideDNA content analysis, we sought to incorporate this flex-ibility into typical multiparameter formats employed inthe analysis of cell cycle events in fixed cell populations.

Figures 11ad demonstrate the ability of DRAQ5 (633-nmexcitation) to identify the G2/M location of the increase incyclin B1 (21) in drug-arrested populations of SU-DHL-4.

Antibodies labeled with either AMCA (UV excitation) orFITC (488-nm excitation) gave similar results.

Figures 12af show the results of a triple-laser analysisof the relatively constant expression of p34Cdc-2 through-out the cell cycle and the presence of a low-frequencysubpopulation of cells with elevated cyclin B1 represent-ing asynchronous cells approaching late G2/M. The triple-beam configuration used in this study permitted the anal-

ysis of UV and visible range fluorochromes using spatiallyseparated laser beams together with the simultaneous

collection of DNA content signals also on spatially sepa-rated 488/633-nm beams. We have noted that any differ-ences between DRAQ5-DNA G1 peak CV values for 488

versus 633-nm excitation appear to relate to the accuracyof the coalignment of the secondary beam.

CONCLUSIONS

We report on the spectral characteristics of a novelcell-permeant fluorescent dye for DNA showing an Exmaxof 646 nm, an Emmax of 681 nm, and an Emrange of665800 nm. The agent is based on a modified anthra-quinone structure and has been prepared as a pure, stable,

FIG. 6. Flow cytometric analyses of DRAQ5 accumulation, for a 1-hexposure period, in viable HL-60 cells. Frequency distribution histogramsare for low-power 633-nm wavelength excitation using cytometer B.Closed circles, triangles, and squares represent 0, 5, and 10 M DRAQ5,respectively. Inset: Linearity of DRAQ5 dose response, using two differ-ent cytometers (namely, B and A with correlation coefficients of 0.96 and0.97 for 633 and 647-nm excitations, respectively). Open circles, cytom-eter A; closed circles, cytometer B.



9/12

and fully synthetic derivative that is soluble in biologicallycompatible solvents (18). This vital stain for DNA, evenunder nonstoichiometric binding conditions, has been

used for the fluorescence imaging of nuclear architectureand the disposition of nuclei within cell populations. Theability of DRAQ5 to reflect cellular DNA content has beenconfirmed, providing a flexible means of integrating thisparameter into the flow cytometric analyses of cell cycle-related events in fixed or live cells. Flexibility arises fromthe laser configurations that can be used, ranging fromsingle 488 argon lasers to triple-laser UV/488/red-line com-binations. We have also demonstrated the incorporationof DRAQ5 identification of nuclear location in CLSM anal-

yses of live cell reporter probes including GFP. The spec-tral characteristics of DRAQ5 permit the simultaneous useof multiphoton activation of intracellular reporter mole-

cules by a pulsed IR laser, exemplified here using theUV-excitable Zn2 probe Zinquin, together with a simul-taneous single-photon excitation of DRAQ5. Furthermore,direct two-photon activation of DRAQ5 for nuclear imag-ing can be achieved at wavelengths 1,000 nm.

Preliminary studies have revealed only a minimal signalderived from RNA-associated DRAQ5 fluorescence infixed cells, consistent with the ability of the dye to reportcellular DNA content in intact cells. The impact of differ-ent levels of RNA within live cells in confounding thereporting of DNA content would need to be assessed in arange of cell types. Our imaging studies indicate that

melanoma and HeLa cells show little evidence of cytoplas-mic fluorescence when exposed to DRAQ5. We havefound no evidence of fluorescence enhancement when

DRAQ5 associates with DNA in free solution. Modelingstudies indicate that DRAQ5 interaction through interca-lation may contribute to the red-shift observed in theemission spectrum and potentially involve a change inbase pair binding preference compared with other anthra-quinones (PT-S, PJS, and LP, unpublished data). The spec-tral separation of DRAQ5 from several visible range fluo-rochromes in common use enhances its utility as a nuclearprobe, given the reduced requirement for electronic sig-nal compensation in flow cytometry and threshholding forsignal contamination in imaging. This has been exploitedfor the detection of nucleated cells in immunophenotyp-ing applications using a standard FACScan system (24).

Under the excitation conditions used for imaging, pho-tobleaching and photoablation were not apparent andDRAQ5 was found to be persistent within cells. Nuclearstaining of live cells could be achieved within seconds incomplete culture medium at room temperature and canbe enhanced by incubation at 37C. The agent belongs toa class of compounds with defined cytotoxic mechanisms(2,17,25, 26), permitting evaluation of the agent for haz-ard monitoring purposes. In keeping with several cell-permeant anthraquinones, DRAQ5 appears to be cyto-toxic to cells in a time-dependent manner. Parallel toxicitystudies (PJS, unpublished observations) have indicated

FIG. 7. Dual-beam flow cytometric analysis(cytometer B) for the detection of DRAQ5-associated fluorescence in fluorescein-la-beled viable HL-60 cells. Representative flowcytometric bivariate plots of green (FL2-height; fluorescein) versus deep red/low in-frared (FL1-height; DRAQ5) whole cell fluo-rescence signals. a,b: Dual-beam excitation offluorescein (488 nm) and DRAQ5 (low-pow-er 633 nm). a: FDA alone (0.2 M for 10min); b: cell pretreated with 5 M DRAQ5for 1 h prior to FDA treatment. c,d: Repeatthe same cell treatment conditions except forthe use of single-beam excitation at 488 nmfor fluorescein and DRAQ5. Numbers indi-cate the percentage of gated events withinthe quadrant regions.

288 SMITH ET AL.


10/12

that the dose yielding 10% survival of clonogenic potentialvalue is 10 nM 24 h for various human tumor celllines. Given the persistent nature of DRAQ5, the probedoes not appear to be appropriate for viable cell sorting

where long-term cell survival is sought. However, currentfindings (Figs. 4af) and our observations on DRAQ5-labeled cells exposed to the mitochondrial function re-porter probe, rhodamine 123 (PJS, unpublished data),suggest no major disruption of cellular function in shorter-term viable cell experiments. Importantly, the rapidity

with which DRAQ5 enters cells permits the analysis ofcellular DNA content or nuclear location in short timeframes after the probing of other characteristics of viablecells. A derivative of DRAQ5 (i.e., DRAQ5NO; 17,27) hasbeen found to have slower uptake, permitting discrimina-tion of intact and membrane-damaged cells by flow cytom-etry and is therefore applicable to studies on apoptoticcell discrimination.

The principle of two-photon excited fluorescence mi-croscopy was first demonstrated by Webb and coworkersat Cornell University (28). In essence, this involves theabsorption of two long-wavelength photons by a fluoro-phore to the first excited singlet state. Importantly, mul-tiphoton excitation avoids the need for potentially dam-aging UV wavelengths and lends itself to the tracking of

events with time in live tissues where illumination pene-tration and bleaching above and below the plane of focusare significant issues. The laser power required for two-photon absorption is much greater than for single-photonexcitation. Compressing the light energy into short pulses,

with a repeat rate that does not increase the time averagepower, enables multiphoton excitation to occur. In thecurrent study, the Ti-sapphire laser produces pulsesaround 100 fs (or less) in length repeated at 80100 MHz.Fluorescence intensity is proportional to the square of theillumination intensity and hence only occurs in the focalplane generating an optical section (29, review). The highpenetration of red-line laser beams into tissues and the

permeant properties of DRAQ5 may provide a useful com-bination that allows three-dimensional orientation and lo-cation of nuclei within living tissues. Preliminary compar-

FIG. 8. Comparison of DNA content analysis by dual-beam flow cytom-etry (cytometer C) for ethidium bromide-stained SU-DHL-4 nuclei andDRAQ5-stained viable SU-DHL-4 cells. DNA content distributions forethidium bromide-stained nuclei refined by RNase A digestion and pulsearea analysis. Viable cells were incubated with 20 M DRAQ5 for 5 minat room temperature and pulse area analyzed. ad: Correlation for fluo-

rescence emissions for first (488 nm) and second laser (UV or 633 nm)excitation of the DNA dyes. c,d: DNA analyses using 488 nm and UVexcitation, respectively, for ethidium bromide-stained SU-DHL-4 nuclei.e,f: DNA analysis using 488 and 633-nm excitation, respectively, forDRAQ5-stained viable SU-DHL-4 cells.

FIG. 9. af: Comparison of perturbed DNA content analysis by dual-beam flow cytometry (cytometer C) for ethidium bromide-stained SU-DHL-4 nuclei and DRAQ5-stained viable SU-DHL-4 cells. Samples werederived from cells exposed to VP-16 (0.25 M 18 h) to generate a latecell cycle accumulation. Panel information given in legend to Figure 8.



11/12

ative studies have been carried out using the two-photonIR activation of the lipophilic cell-permeant dye, Hoechst33342, versus the single-photon 647-nm activation of

DRAQ5 in whole tissue mounts of mouse seminiferoustubules (PJS, unpublished data). Our results indicate anability of DRAQ5 to penetrate rapidly into multilayer tis-sue structures, to an extent achievable by the UV fluoro-chrome Hoechst 33342 and to permit the imaging of cellnuclei and individual chromosomes using CLSM.

We conclude that DRAQ5 represents a useful additionto the present range of DNA fluorochromes available forthe assessment of cellular DNA content, cell cycle posi-tion, nuclear architecture, and location. Its primary advan-tages for flow cytometry arise from its cell-permeantproperties and spectral separation from visible range fluo-

FIG. 10. DNA content resolution of asynchronous (a,c) and cell cycleperturbed (b,d; 0.25 M VP-16 18 h) cells using a benchtop FACScancytometer (cytometer D). Residual G1 and G2/M channel locations weredetermined by analyzing cell cycle distributions as described previously(23).a,b: Ethidium bromide-stained cells. c,d: DRAQ5-stained live cells.

FIG. 11. Analysis of DNA content and cyclin B1 expression by dual-beam (488 nm/UV; cytometer C) flow cytometry. Fixed, RNaseA-digestedand DRAQ5-stained (FL3; 488-nm excitation) SU-DHL-4 cells were pre-pared from an asynchronous culture exposed to 0.25 M VP-16 for 18 hto accumulate cells in G2/M. G2/M-phase expression of cyclin B1 protein

was monitored by indirect immunofluorescence either using an FITC-labeled second antibody (FL1-height; 488-nm excitation) or an AMCA-labeled second antibody (FL5-height; multiline UV excitation) to detectthe binding of anti-cyclin B1 (GNS1) mouse monoclonal IgG. a,c: Anti-body controls (nonspecific IgG plus second antibody). b,d: Results forspecific antibody plus second antibody. G1, S, and G2/M represent cellcycle stages; unlettered arrow, expected position of cells expressing highlevels of cyclin B1 and located in G2/M of the cell cycle.

FIG. 12. Analysis of DNA content, cyclin B1, and p34Cdc-2 expression bytriple-beam (488 nm/UV/633 nm; cytometer C) flow cytometric analysisof DNA, cyclin B1, and p34Cdc-2 in fixed, RNase A digested, and DRAQ5-stained asynchronous SU-DHL-4 lymphoma cells. DRAQ5 fluorescence

was monitored by 488-nm excitation (FL3 pulse height) and 633-nmexcitation (FL4 pulse height). Cell cycle-independent p34Cdc-2 protein andG2/M-specific cyclin B1 protein were monitored by indirect immunoflu-orescence using FITC-labeled second antibody (FL1-height; 488-nm exci-tation) and AMCA-labeled second antibody (FL5-height; multiline UV ex-citation) respectively. a,b: DNA versus p34Cdc-2. c,d: DNA versus cyclinB1.e,f:DNA histograms for blue and red excitation wavelengths, respec-tively. Arrowed subpopulations: G1, S, and G2/M represent cell cyclephases; H CyB, small subpopulation of high cyclin B1-expressing cellslocated in G2/M of the cell cycle.

290 SMITH ET AL.


12/12

rochromes (e.g., FITC and Texas Red). We have alsoexploited the unusual spectral properties of DRAQ5 toshow how a range of laser-based systems can be used toprovide excitation of intracellular DRAQ5-DNA using laserlines available from argon ion, krypton, krypton-argon,and helium-neon lasers. This flexibility is particularly im-portant in multiparameter analyses. DRAQ5 would also

appear to have significant applications in combining thesingle- and two-photon excitation of intracellular reportermolecules, including GFP, in live cells with the simulta-neous imaging of the nuclear compartment. We suggestthat real-time imaging studies, e.g., involving the live celltracking of GFP-tagged proteins in intra- and extranuclearcompartments, would benefit from the ability of DRAQ5to simultaneously report nuclear architecture and loca-tion.

LITERATURE CITED

1. Lown JW, Morgan AR, Yen S-F, Wang YH, Wilson WD. Characteristicsof the binding of the anticancer agents mitoxantrone and amet-antrone and related structures to deoxyribonucleic acids. Biochemis-

try 1985;24:40284035.2. Cheng CC, Zbinden G, Zee-Cheng RK-Y. Comparison of antineoplas-

tic activity of aminoethylaminoanthraquinones and anthracycline an-tibiotics. J Pharm Sci 1979;68:393396.

3. Bell DH. Characterization of the fluorescence of the antitumouragent, mitoxantrone. Biochim Biophys Acta 1988;949:132137.

4. Fox ME, Smith PJ. Subcellular localisation of the antitumour drugmitoxantrone and the induction of DNA damage in resistant andsensitive human colon carcinoma cells. Cancer Chemother Pharma-col 1995;35:403410.

5. Smith PJ, Sykes HR, Fox ME, Furlong IJ. Subcellular distribution of theanticancer drug mitoxantrone in human and drug-resistant murinecells analyzed by flow cytometry and confocal microscopy and itsrelationship to the induction of DNA damage. Cancer Res 1992;52:19.

6. Smith PJ, Desnoyers R, Blunt NE, Giles Y, Patterson LH, Watson JV.Flow cytometric analysis and confocal imaging of anticancer alkylami-noanthraquinones and their n-oxides in intact human cells using 647nm krypton laser excitation. Cytometry 1997;27:4353.

7. Waggoner AS. Fluorescent probes for cytometry. In: Melamed MR,Lindmo T, Mendelsohn ML, editors. Flow cytometry and sorting, 2nd

edition. New York: John Wiley & Sons; 1990. p 209225.8. Darzynkiewicz Z, Kapuscinski J. Acridine orange: a versatile probe of

nucleic acids and other cellular constituents. In: Melamed MR,Lindmo T, Mendelsohn ML, editors. Flow cytometry and sorting, 2nd

edition. New York: John Wiley & Sons; 1990. p 291314.9. Latt SA, Langlois RG. Fluorescent probes of DNA microstructure and

DNA synthesis. In: Melamed MR, Lindmo T, Mendelsohn ML, editors.Flow cytometry and sorting, 2nd edition. New York: John Wiley &Sons; 1990. p 249290.

10. Hirons GT, Fawcett JJ, Crissman HA. TOTO and YOYO: new verybright fluorochromes for DNA content analyses by flow cytometry.Cytometry 1994;15:129140.

11. Rabinovitch PS, Torres RM, Engel D. Simultaneous cell cycle analysisand two-color surface immunofluorescence using 7-amino-actinomy-cin D and single laser excitation: applications to study of cell activa-tion and the cell cycle of murine Ly-1 B cells. J. Immunol 1986;136:27692775.

12. Terstappen LW, Shah VO, Conrad MP, Recktenwald D, Loken MR.Discriminating between damaged and intact cells in fixed flow cyto-metric samples. Cytometry 1988;9:477484.

13. Terstappen LW, Meiners H, Loken MR. A rapid sample preparation

technique for flow cytometric analysis of immunofluorescence allow-ing absolute enumeration of cell subpopulations. J Immunol Methods1989;123:103112.

14. Frey T. Nucleic acid dyes for detection of apoptosis in live cells.Cytometry 1995;21:265274.

15. Frey T, Yue S, Haugland RP. Dyes providing increased sensitivity inflow-cytometric dye-efflux assays for multidrug resistance. Cytometry1995;20:218227.

16. Patterson LH. Anthraquinone anticancer compounds with (disubsti-tuted amino-N-oxide)alkylamino substituent. 1989; UK Patent GB 2237 283.

17. Smith PJ, Blunt NJ, Desnoyers R, Giles Y, Patterson LH. DNA topo-isomerase II dependent cytotoxicity of alkylaminoanthraquinonesand their N-oxides. Cancer Chemother Pharmacol 1997;39:455461.

18. Smith PJ, Patterson LH. An anthraquinone and its derivatives. 1998;British Patent Application No. 9813062.8.

19. Roderick HL, Llewellyn DH, Campbell AK, Kendall TM. Role ofcalreticulin in regulating intracellular Ca2 storage and capacitativeCa2 entry in HeLa cells. Cell Calcium 1998;24:253262.

20. Wessendorf MW, Brelje TC. Multicolor fluorescence microscopy us-ing the laser-scanning confocal microscope. Neuroprotocols 1993;2:121140.

21. Gong J, Traganos F, Darzynkiewicz Z. Simultaneous analysis of cellcycle kinetics at two different DNA ploidy levels based on DNAcontent and cyclin B measurements. Cancer Res 1993;53:50965099.

22. Zalewski PD, Forbes IJ, Betts WH. Correlation of apoptosis withchange in intracellular labile Zn(II) using Zinquin [(2-methyl-8-p-toluenesulphonamido-6-quinolyloxy)acetic acid], a new specific flu-orescent probe for Zn(II). Biochem J 1993;296:403408.

23. Watson JV, Chambers SH, Smith PJ. A pragmatic approach to theanalysis of flow cytometric DNA histogram data. Cytometry 1987;8:18.

24. Smith PJ, Wiltshire M, Davies S, Patterson LH, Hoy T. A novel farred-fluorescing DNA probe, DRAQ5, for intact blood cell discrimina-tion by flow cytometry. J Immunol Methods 1999;229:131139.

25. Fox ME, Smith PJ. Long-term inhibition of DNA synthesis and thepersistence of trapped topoisomerase II complexes in determining

the toxicity of the antitumour DNA intercalators mAMSA and mitox-antrone. Cancer Res 1990;50:58135818.

26. Shenkenberg TD, Von Hoff DD. Mitoxantrone: a new anti-cancer drugwith significant clinical activity. Ann Intern Med 1986;105:67 81.

27. Wiltshire M, Patterson LH Smith PJ. A novel deep red/low infraredfluorescent flow cytometric probe, DRAQ5NO, for the discriminationof intact nucleated cells in apoptotic cell populations. Cytometry2000;39:217223.

28. Denk W, Strickler JH, Webb WW. Two photon laser scanning fluo-rescence in microscopy. Science 1990;248:7376.

29. White NS, Errington RJ. Improved fluorescence laser scanning mi-croscopy by multi-photon excitation. In: Hawkes P, editor. Advancesin imaging and electron physics. San Diego: Academic; 2000. p249277.


Documents

Characteristics of a Novel Deep RedInfrared