document

NATURE MEDICINE • VOLUME 5 • NUMBER 3 • MARCH 1999 351

NEW TECHNOLOGY

Advances in the development of fluorescent probe technologyhave greatly facilitated the analysis of single cellular functionor single-mode microscopy. Existing approaches monitorsingle probes or follow specific cellular events over time for thepurposes of dissecting complex molecular dynamics in livingcells1. In addition, fluorescent cellular probes are more fre-quently used with fixed samples, yielding static information onprocesses captured only at the time window when fixation ofcells or tissues had occurred. Fixation results in protein denatu-ration, cross-linking and other interactions, with the subse-quent loss of dynamic signals that can only be captured inliving cells. Techniques that would allow a tandem, coherentapproach using multiple probes in living cells would obviouslybe preferable. An ideal system of compatible probes wouldallow simultaneous monitoring of various dynamic functionsin a manner that was not feasible until recently. Here wedescribe a coherent system of fluorescent probes for trackingmultiple cellular functions in vitro closely spaced within a nar-row sampling period. The term ‘coherent’ applied to this sys-tem represents the systematic, coordinated and integralconnections between the probes used, their spectral character-istics and the instrumental mode of application to analyze si-multaneous target organelles or cellular functions withacquired images stacked on register. Observing different cellu-lar events simultaneously within a narrow time windowincreases the ability to discern interrelationships of processesaffecting cellular organelles from animals and humans. Wehave studied specific organelle functions in vitro using live livercells, a useful model system for observing cellular events in realtime. By analyzing the sequence of intracellular events closer to

their native state, this approach will reveal relevant new in-sights into normal, pathologic or toxic processes2.

Coherent multiprobe fluorescence probes and cellular functions The probes were chosen based on their compatibility, or inher-ent cell permeability, or whether they are readily internalized byliver cells. A similar approach could be applied to other celltypes in culture, with a multiple-probe coherent algorithm offluorophores constructed to monitor a group of cytoplasmic ornuclear events3. Transfection or other similar procedures to in-ternalize otherwise non-permeant probes cause substantial cel-lular artifacts and have not been explored in this context. Manyfluorophores are available. However, few probes can be usedconcurrently, and the organelle and cellular targets that can beexplored in liver cells limits the choice. Liver cells are ideal sub-jects of study because there is a variety of well-defined cytoplas-mic functions and organelle interactions. Many of thesefunctions or processes can be monitored by fluorophores withaffinity for organelle integrity and disposition, ion transport, cy-toskeleton assembly, and related enzyme/cofactor or transcrip-tional activities1. Intracellular functions affected during thedevelopment of toxicity provide convenient, useful targets forthe simultaneous monitoring of multiple organelles (Table). Inthe modality described here, a group of organelle responses canbe followed in real time using quantitative fluorescencemicroscopy and advanced digital imaging techniques. Thesimultaneous use of these fluorophores does not exclude the useof micro-injected or transfected non-permeant probes in orderto address specific functional hypotheses.

An example of a coherent multiprobe system that we testedsuccessfully included the parametersof mitochondrial function, calciumtransport and cytoskeletal integrity(Table, Group 1). Mitochondria havean important role in the regulatorypathways of energy supply, detoxifi-cation and cell survival. Manyprocesses leading to cell death are reg-ulated by cytoplasmic organelles,including mitochondria, which cangenerate superoxide anion radicalsand hydrogen peroxide4. Reductionof mitochondrial transmembranepotential (∆ψm) is an early event, pre-ceding irreversible apoptotic andnecrotic events. Decreased ∆ψm isassociated with loss of mitochondrialfunction, leading to disruption ofmetabolic activity and lack of cell via-bility5. Changes in ∆ψm can be mea-sured using cell-permeant, cationic,

Monitoring simultaneous subcellular events in vitroby means of coherent multiprobe fluorescence

DOUGLAS R. PLYMALE, JEFFREY R. HASKINS & FELIX A. DE LA IGLESIA

Pathology and Experimental Toxicology Department, Parke-Davis Pharmaceutical Research,2800 Plymouth Road, Ann Arbor, Michigan 48105, USA

Correspondence should be addressed to F.A. de la I.; email: [email protected]

Table Examples of coherent fluorescent probes for simultaneous monitoring of cellular processes

Group Cellular function Probes Spectral characteristicsa

Excitation Emission

1 Intracellular free Ca2+ Fura-2 340/380b 510Mitochondrial transmembrane potential MitoTracker Green 490 516Plasma membrane permeability/ Texas-Red phalloidin 591 608

cytoskeletal Integrity2 Intracellular free Ca2+ Fura-2 340/380b 510

Ionic homeostasis BCECFc 450/505b 640/525b

3 Ca2+ compartmentalization/distribution Fluo-3 506 526(mitochondrial & cytoplasmic) Rhod-2 552 581

4 Detoxification (glutathione) Monochlorobimane 380 480Lysosomal activity FITC-Dextrand 494 518Mitochondrial transmembrane potential Rhodamine 123 507 529

5 Detoxification (glutathione) Monochlorobimane 380 480Plasma membrane permeability/ FITC-phalloidind 496 516

cytoskeletal integrityMitochondrial transmembrane potential Tetramethylrhodamine 548 573

aUnits in nanometers; bRatiometric probe; c2′,7′-bis-(2-carboxyethyl)-5-(and-6)-carboxyfluorescein (BCECF); dFluoresceinisothiocyanate (FITC) conjugated to either dextran or phalloidin.

© 1999 Nature America Inc. • http://medicine.nature.com©

199

9 N

atu

re A

mer

ica

Inc.

• h

ttp

://m

edic

ine.

nat

ure

.co

m

Fig. 2 Photobleaching andprobe stability. To evaluate probestability in response to maximalphotodosage conditions, we cap-tured sequential images from he-patocytes loaded individually withFura-2, MitoTracker Green, orTexas-Red phalloidin (a) or a co-herent combination of the probes(b). To evaluate photobleachingand probe stability over the in-tended period of image capture,we monitored cultures loadedwith the individual probes (c) orthe coherent combination (d)over 2 h, collecting image setsevery 5 min. Average field inten-sity values (± s.e.m.) of labeledhepatocytes are shown.

Fig. 1 Multimode microscopy. The robotic stage auto-matically scans the specimen in the x, y and z directions.For this four-chambered slide images are taken from thesequential fields in one pass (1–16) every 5 min for theduration of the experiment. At each time point, the filterwheel rotates through each filter optimally selected foreach probe and acquires an image with the CCD for eachprobe in sequence. Inset, Spectral basis for multiprobefluorescence assay. The spectral curves at emission (e) orexcitation (x) correspond to the probes used in the sys-tem. The overlap of MitoTracker Green (MTe,MTx) withFura-2 (F2e) is not relevant to this study because MTimage capture is ahead of F2x, thus eliminating secondaryexcitation artifacts. TR, Texas-Red phalloidin.

352 NATURE MEDICINE • VOLUME 5 • NUMBER 3 • MARCH 1999

NEW TECHNOLOGY

fluorescent probes that preferentially localize tomitochondria. Mitochondria maintain electro-chemical gradients across the inner membrane,permitting incorporation and retention of theseprobes. Thus, fluorescence decrements indicatereduced or negative ∆ψm and, therefore, are a sig-nal of mitochondrial dysfunction6,7. For our initialcoherent, multiprobe combination, we usedMitoTracker to label mitochondria and followed ∆ψm changes.This probe is similar to other ∆ψm-sensitive probes8,9, except thatMitoTracker has higher quantum efficiency, good photostabil-ity and uneventful internalization, as well as retention byhealthy cells10.

Calcium is a chief regulator of cellular homeostasis, intracel-lular messaging and cellular toxicity involving a variety of cellu-lar and signaling processes as free cytosolic calcium11,12. Theintracellular traffic of free calcium ions can be monitored usingcalcium-specific fluorescent probes. In a coherent multiprobereal-time capture of fluorescence approach, we used thecalcium-binding probe Fura-2 (ref. 13).

The integrity of the cytoskeleton plays an active part in chiefcellular events. Components of the cytoskeleton coordinate cellkinesis and cell shape, intracellular organelle trafficking, andprocesses leading to cell division and cell death. Actin is a com-ponent of the cytoskeleton, and the structure of the actin fila-

ment network can be visualized using fluorescent-conjugatedprobes in single-mode microscopy. For this purpose, we usedTexas Red-conjugated phalloidin, a cellular poison that binds topolymerized actin and prevents the cycling of actin subunits14,15.Phalloidins are normally cell-impermeant but injured hepato-cytes internalize sufficient dye conjugate to allow discretevisualization of the reticular architecture.

Multimode microscopy and coherent fluorescence analysisThe ability to monitor multiple cellular functions in real time ispredicated on a new generation of advanced light microscopesystems. The multimode microscope was developed initially atthe Center for Light Microscope Imaging and Biotechnology,Carnegie Mellon University16. Multimode microscopy allowsthe automated, rapid and sequential application of differentoptical microscopy modes, including epifluorescence and differ-ential interference contrast (DIC). The multimode microscope is

a c

b d

Exposure sequence

Single probes

Coherent probes Coherent probes

Single probes

Time (min)

Ave

rag

e fie

ld in

tens

ityA

vera

ge

field

inte

nsity

λ (nm)


199

9 N

atu

re A

mer

ica

Inc.

• h

ttp

://m

edic

ine.

nat

ure

.co

m

Fig. 3 Quantitation of multiprobe fluorescence in hepatocytes treatedwith tacrine. Timecourse of near-simultaneous changes in fluorescencefrom a multiprobe assay using rat hepatocytes with 1 mM tacrine in themedia. The image capture sequence at any time point is described in thetext. Plasma membrane and cytoskeleton are tracked by Texas-Red phal-loidin, mitochondrial transmembrane potential by MitoTracker Green andintracellular calcium by Fura-2.


NEW TECHNOLOGY

equipped with servos controlled by interfacing software linkedto a computer equipped with image capture and analysis pro-grams. The mechanical, electrical, digital and optical compo-nents of the multimode microscope have been optimized forbiological applications8 with further adaptations in our labora-tory, including environmental control of cell preparations nec-essary for the coherent multiprobe fluorescence system. Therobotic stage allows parallel tracking of multiple fields, bothwithin a single chamber and in multiple sample matrices suchas multiple-well chamber slides or plates (Fig. 1). Hence, thismultimode microscopy approach addresses inter- and intra-sample variability and coherent corrections needed while con-ducting these experiments.

Validation of coherent multiprobe fluorescence assaysTo use multiple probes simultaneously, we first had to define theresponses of each probe individually and under comparable ex-perimental conditions. The set of conditions thus obtained pro-vided the starting point for determining the effective sequentialparadigm for the coherent multiple fluorescent probe combina-tions. The first step defined optimization parameters such as dyeloading times, photobleaching and dye leakage over the timeand intensity of photoexposures for each individual probe.These parameters were evaluated by loading isolated rat orhuman hepatocytes with a single probe and applying the tools ofmultimode microscopy in the same way as with the coherentmultiple probe setup. Preliminary experiments identified effec-tive concentrations of Texas-Red phalloidin, MitoTracker Greenand Fura-2 that were not cytotoxic and gave adequate quantumfluorescent responses, generally greater than 75% over the linearrange of the charge-coupled device (CCD) but not sufficient togenerate interfering autofluoroscence. To evaluate photobleach-ing and probe stability over the intended period of image cap-ture, multiple fields within cultures loaded with individualprobes were monitored over 2 h, collecting image sets every 5 min. To evaluate probe stability in response to maximal photo-dosage conditions, sequential images were captured from hepa-tocytes loaded with individual probes. All three dyes were stableover the monitoring period and did not demonstrate significantdegradation based on the extent of photoexposure (Fig. 2). Therewas also no evidence of probe-induced phototoxicity in theexperimental conditions described here.

We next evaluated multiple fluorophore interactions. Multiplefields in cultures loaded with the probe combination were moni-tored through acquisition of several multiprobe image sets. Amultiprobe image set is composed of a series of images capturedfrom each mode at each single time point. Parameters consideredwhen co-loading multiple probes included the loading sequence,the loading interval for each probe, probe leakage, photobleach-ing, emission-generated secondary excitation and residual phos-phorescence. The spectral emission characteristics were also

Fig. 4 Multiprobe fluo-rescence analysis. Thepower of the multimodemicroscope is demon-strated by analysis ofinformation from multipleacquisition modes within asingle time series of a sin-gle sample. For example,images from Texas-Redphalloidin (a and b),MitoTracker Green (c andd) and Fura-2 (e and f)acquisition modes caneach be displayed in asingle color channel andthen coherently super-imposed to create amulticolor complemen-tary image (g and h). Thisallows spatial resolutionof the information on register from multiple fluorescence acquisitionmodes. Analysis of images from the combined acquisition modes overtime demonstrates the interrelationships of multiple cellular functions.Control cells (a,c,e,g); tacrine-treated liver cells (b,d,f,h) were 1 h intothe 2-h exposure period. White arrows (right column) indicate a nor-mal cell with no uptake of Texas Red phalloidin (b), green mitochondr-

ial fluorescence (d) and moderate cytosolic calcium (f). White asterisksindicate a group of altered cells with heavy accumulation of Texas-Redphalloidin (b), little or no green (d) or blue (f), indicating significant cellinjury or death. A comparison of the complementary images (g and h)shows the extent of loss of activity in the treated hepatocytes (h) com-pared with the control (g).

Time (min.)

∆fr

om t

=0 m

in (

%)

a b c d

e f g h


199

9 N

atu

re A

mer

ica

Inc.

• h

ttp

://m

edic

ine.

nat

ure

.co

m

354 NATURE MEDICINE • VOLUME 5 • NUMBER 3 • MARCH 1999

NEW TECHNOLOGY

considered for compatibility. These parameters were evaluatedwith the multiprobe combination of Texas-Red phalloidin,MitoTracker Green and Fura-2. When loaded simultaneously,the three probes were photostable over 2 h and at maximal con-ditions of maximal photodosage (Fig. 2). As with the individualprobes, phototoxicity was not seen when the probes were evalu-ated in combination. This robust protocol showed noprobe–probe interactions that substantially affected the spectralproperties of the individual probes.

Coherent multiprobe fluorescence applications in hepatocytes After optimizing the multiple fluorophore combination ofTexas-Red phalloidin, MitoTracker Green and Fura-2, targetingactin filaments, mitochondrial transmembrane potential andintracellular calcium, respectively, we used this coherent multi-probe fluorescence approach to monitor cytotoxic effects in live,isolated hepatocytes.

To exemplify this approach, we used three agents to studyeffects caused by chemical intervention. The first was tacrine(1,2,3,4-tetrahydro-9-aminoacridine hydrochloride mono-hydrate), a cholinesterase inhibitor drug for the treatment ofAlzheimer's disease. This drug induces release of cytoplasmicenzymes in hepatocytes in vitro and causes mitochondrialdysfunction7. The second agent was melittin, a highly lyticpeptide derived from bee venom that binds to calmodulin17.The final agent was human recombinant transforming growthfactor β1 (TGF-β1), which induces apoptosis in cultured hepato-cytes18. Only results from trials with tacrine and melittin aredescribed here.

In hepatocytes labeled with multiple probes and analyzed inreal time, information from single acquisition modes wasexamined separately or collectively. Certain toxicants affectcellular events at different times after exposure and require par-ticipation of subcellular organelles either sequentially or some-times concomitantly19,20. The timecourse of functional changesin liver cells exposed to 1 mM tacrine for 2 h and using acoherent probe group (Group 1, Table) is shown in Fig. 3.

Mitochondrial transmembrane potential, asfollowed by MitoTracker Green fluorescence,increased over the first 35 min and graduallydecreased over the subsequent 60 min.Membrane permeability (Texas-Red phal-loidin) increased linearly within the same 35 min and remained essentially unchangedfor the subsequent 60 min. Calcium traffic,tagged by Fura-2, showed progressive accumu-lation in cells in parallel with the cell perme-ability changes. At the end of the 2-hexperiment, some cells showed prominentchanges (Fig. 4), with appreciable loss ofcytoskeletal integrity (Fig. 4b), loss of mito-chondrial membrane potential (Fig. 4d) andeventual loss of the excessively internalizedcalcium. Viable cells showed no accumulationof Texas-Red phalloidin, green mitochondrialfluorescence and moderate cytosolic calcium.These results agree with laser-confocalcytometry results obtained with rhodamine123 as a monoprobe for mitochondrial trans-membrane potential7.

In hepatocytes exposed to melittin, theplasma membrane lost integrity and the cy-

toskeleton underwent extensive changes as the cells detachedfrom the plate. From the DIC images acquired over the 2-h timeperiod, melittin caused substantial and extensive blebbing of theplasma membrane in most cells examined (Fig. 5). Themultimode microscope facilitated the visualization of eventswithin the visible range in real time using DIC coupled with asingle channel of fluorescence, which added a complementarylevel of image information. Analysis of coherent images from theFura-2 acquisition mode corresponding to the same samplingtimes as the DIC image mode demonstrated that blebs containedcytoplasmic components, including calcium, and were releasedinto the media. This observation indicates that high anionconcentrations were actively compartmentalized and reduced bythe cell in order to maintain homeostasis. The pinching-offprocess provides evidence of a mechanism whereby extempora-neous membrane-bound cytoplasmic fragments containing en-zymes are released into the media, an event that could explainenzyme elevations in vivo. From the corresponding imagesacquired with the Texas-Red phalloidin acquisition mode, sub-stantial amounts of the dye conjugate were taken up and cells be-came considerably brighter, indicating profoundly increased cellpermeability (Fig. 5). Therefore, the dynamic range of the cap-tured fluorescence represented an expression of cells progressingtowards death.

Coherent fluorescent multiprobes: Future directions We have demonstrated the feasibility of coherent multiprobefluorescence microscopy with a system of three cellular functionstracked simultaneously in live liver cells. This approach can beextended by using other fluorophore combinations, increasingthe number of probes or, ultimately, by applying more sensitive,rapid and higher-throughput capacity systems for monitoringseveral more cell functions. It is possible that with improved sta-bility of primary cell cultures and longer observation periods thatirreversible pathways or even compensatory or proliferativeresponses will be recognized. Further probe combinations or alarger number of coherent probes can be evaluated in a similar

Fig. 5 Time-lapse images of hepatocytes treated with melittin. Melittin (final concentration,2 µM) was added to primary hepatocyte cultures loaded with Texas-Red phalloidin,MitoTracker Green and Fura-2. Each timed set (5′, 25′ and 50′) shows paired images of hepa-tocytes corresponding to transmitted DIC (top) and calcium-sensitive (bottom) wavelengths.Arrows indicate cells for which substantial probe uptake precedes loss of viability.

DIC

Ca2+


199

9 N

atu

re A

mer

ica

Inc.

• h

ttp

://m

edic

ine.

nat

ure

.co

m


NEW TECHNOLOGY

manner and used to study interrelationships of different cellularconstituents and processes. The feasibility of these improvementsrelies not only on an increased number of probes or a larger set offluorophores but also on advances in the hardware components.We are now incorporating spectral analysis into the coherentmultiprobe approach by means of a Sagnac interferometer. Theapplication of spectral analysis in this setup will increase theresolution of emission peaks separated by as little as 10–20 nm,which will avoid the spectral overlap caused by wide bandpasseffects21. This approach will also allow the development of a largernumber of multiple coherent probes. A hardware limitation ofthe multimode microscope is the time delay caused by themechanical positioning of the filters in the microscope beampath. This limitation can be obviated by replacing the glass filtersand carriers with acousto-optical tunable filters with resultingincreased speed of the system22. This continuous solid-statetunability will permit switching between multiple fluorescencemodes within milliseconds or less, allowing examination ofevents at closer time intervals than the current system couldpossibly attain. These increased efficiencies will facilitate moni-toring functional probes for enzyme inhibition or induction,membrane or organelle synthesis and cytoplasmic motion.

MethodsMultimode microscopy. We used a microscope system modified fromthat previously described8 (Fig. 1). The multimode microscope used hereconsists of a robotic stage fitted with a controlled environmental chamber,a series of dichroic filter sets housed within a motorized carrier, supple-mental high-speed fluorescence and neutral density filter wheels, a high-sensitivity CCD camera (Photometrics 12-bit Series 300 CH250, Tucson,Arizona) and a video-rate camera (C2400 CCD; Hamamatsu, HamamatsuCity, Japan), all attached to an inverted microscope (Zeiss Axiovert 135)equipped with 100-W mercury and halogen light sources and high-numerical aperture objective lenses.

Primary liver cell isolation and culture. Primary hepatocyte cultures wereprepared from rat livers perfused with collagenase as described7.Hepatocytes were purified by differential centrifugation through a Percollgradient, washed and suspended in hepatocyte culture medium (HCM):Leibovitz-15 with L-glutamine (Life Technologies) supplemented with7.5% BSA (Life Technologies), 1% penicillin/streptomycin, 3 mg/ml pro-line, 50 mg/ml galactose, 0.1% insulin-transferrin-selectin (ITS;Collaborative Biomedical Products, Bedford, Massachusetts), 0.4 mg/mldexamethasone (Sigma), 8.4% sodium bicarbonate, and 0.1% trace ele-ments (CuSO4, Fe(NO3)3, ZnSO4 and MnCl2). Hepatocyte viability was as-sessed by Trypan Blue exclusion. Hepatocytes were plated intofour-chambered glass slides to a density of approximately 75,000 cells percm2, and incubated overnight at 37 °C in an environment of 5% CO2. Freshmedia was added to each chamber before the cells were exposed to thefluorescence probe coherent combination.

Fluorescent probe loading. The fluorescent probes were loaded by ‘spiking’ each chamber with an appropriate volume of probe stocksolution. A 1-mM stock solution of Fura-2 and a 1-µM stock solution ofMitoTracker Green were prepared in DMSO. A stock solution of Texas-Red-conjugated phalloidin (200 unit per ml) was prepared in methanol.For single-probe experiments, media with the final probe concentrationwas added to the culture chamber. After incubation at 37 °C, the cells werewashed once with fresh media and 0.5 ml was added to each chamber.Preliminary experiments identified optimal probe concentrations andloading times (not shown). For Fig. 4, Fura-2 was loaded for 30 min at afinal concentration of 5 µM, MitoTracker Green was loaded for 15 min at a final concentration of 5 nM, and 1 unit of Texas-Red phalloidinper ml of media was loaded for 15 min. For the coherent probe combina-tions, MitoTracker Green and Texas-Red phalloidin were added 15 minafter the addition of Fura-2. All other parameters and steps were identicalto those just described for single-probe loading.

Image acquisition and analysis. Under multimode microscopy, digital im-ages were captured with sequential exposures of 0.05 s, 0.2 s, 1.0 s and 1.0s for Texas-Red phalloidin, MitoTracker Green, Fura-2 (380 nm) and Fura-2(340 nm), respectively. Illumination was provided by a 100-W mercurylamp and 560/645, 480/530, 340 or 380/510 excitation/emission filtersfor the same sequence of probes. These exposures consistently saturatedgreater than 75% of the linear range of the 12-bit CCD detector. In coher-ent labeling experiments, the order of acquisition for discrete field imageswas transmitted/DIC, Texas-Red phalloidin, MitoTracker Green and Fura-2at 340 nm and then at 380 nm (Fig. 1). A series of images using this filtercombination was sequentially acquired at each stage location every 5 minover 2 h using the high-resolution, liquid-cooled CCD camera mounted onthe multimode microscope. Four fields from each chamber of a four-cham-ber slide were photographed a total of 25 times in a single run, generating400 images over the 2-h sampling period. For digital image analysis, theaverage pixel intensities from the discrete fields were calculated using theraw 12-bit images as sources and applying image analysis software (MediaCybernetics, Silver Spring, Maryland).

1. Giuliano, K.A., DeBinsio, R., Feineigle, P. & Taylor, D.L. in Motion of Living Cells (eds.Soll, D.R. & Wessels, D.) 53–66 (Wiley, New York, 1998).

2. Taylor, D.L. et al. Potential of machine vision light microscopy in toxicologicalpathology. Toxicol. Pathol. 22, 145–159 (1994).

3. Waggoner, A., Taylor, D.L., Seadler, A. & Dunlay, T. Multiparameter fluorescenceimaging microscopy: Reagents and instruments. Hum. Pathol. 27, 494–502 (1996).

4. Sohal, R.S., & Dubey, A. Mitochondrial oxidative damage, hydrogen peroxiderelease, and aging. Free Rad. Biol. Med. 16, 621–626 (1994).

5. Zamzami, N., Hirsch, T., Dallaporta, B., Petit, P.X. & Kroemer, G. Mitochondrial im-plication in accidental and programmed cell death: Apoptosis and necrosis. J. Bioenerg. Biomembr. 29, 185–193 (1997).

6. Johnson, L.V., Walsh, M.L., Bockus, B.J. & Chen, L.B. Monitoring of relative mito-chondrial membrane potential in living cells by fluorescence microscopy. J. Cell.Biol. 88, 526–535 (1981).

7. Monteith, D.K., Theiss, J.C., Haskins, J.R. & de la Iglesia, F.A. Functional and subcel-lular organelle changes in isolated rat and human hepatocytes induced by tetrahy-droaminoacridine. Arch. Toxicol. 72, 147–156 (1998).

8. Taylor, D.L. et al. Automated light microscopy for the study of the brain: Cellularand molecular dynamics, development and tumorigenesis. Ann. NY Acad. Sci. 820,208–228 (1997).

9. Garner, D.L., Thomas, C.A., Joerg, H.W., DeJarnette, J.M. & Marshall, C.E.Fluorometric assessments of mitochondrial function and viability in cryopreservedbovine spermatozoa. Biol. Reprod. 57, 1401–1406 (1997).

10. Poot, M. et al. Analysis of mitochondrial morphology and function with novelfixable fluorescent stains. J. Histochem. Cytochem. 44, 1363–1372 (1996).

11. Berridge, M.J. Elementary and global aspects of calcium signaling. J. Physiol. 499,297–306 (1997).

12. Trump, B.F. & Berezesky, I.K. Calcium-mediated cell injury and cell death. FASEB J.9, 219–228 (1995).

13. Grynkiewicz, G., Poenie, M. & Tsien, R.Y. A new generation of Ca2+ indicators withgreatly improved fluorescence properties. J. Biol. Chem. 260, 3440–3450 (1985).

14. Cooper, J.A. Effects of cytochalasin and phalloidin on actin. J. Cell. Biol. 105,1473–1478 (1987).

15. Miki, M., Barden, J.A., dos-Remedios, C.G., Phillips, L. & Hambly, B.D. Interaction ofphalloidin with chemically modified actin. Eur. J. Biochem. 165, 125–130 (1987).

16. Farkas, D.L. et al. Multimode light microscopy and the dynamics of molecules, cells,and tissues. Annu. Rev. Physiol. 55, 785–817 (1993).

17. Cornut, I., Thiaudiere, E. & Dufourcq, J. in The Amphipathic Helix (ed. Epand, R.M.)174–219 (CRC, Boca Raton, 1993).

18. Schulte-Hermann, R., Bursch, W., Low-Baselli, A., Wagner, A. & Grasi-Kraupp, B.Apoptosis in the liver and its role in hepatocarcinogensis. Cell Biol. Toxicol. 13,339–348 (1997).

19. Plymale, D.R. & de la Iglesia, F.A. Acridine-induced subcellular and functionalchanges in isolated human hepatocytes in vitro. J. Appl. Toxicol. (in the press).

20. Feuer, G. & de la Iglesia, F.A. in Drug-Induced Hepatotoxicity Vol. 121 (eds.Cameron, R., Feuer, G. & de la Iglesia, F.A.) 46–73 (Springer-Verlag, Heidelberg,1996).

21. Garini, Y. et al. in Fluorescence Imaging Spectroscopy and Microscopy, ChemicalAnalysis Series Vol. 137 (eds. Wang, X.F. & Herman, B.) 87–124 (Wiley, New York,1996).

22. Wachman, E.S., Niu, W. & Farkas, D.L. AOTF microscope for imaging withincreased speed and spectral versality. Biophys. J. 73, 1215–1222 (1997).

CORRECTIONIn the New Technology article “Gene expression profiles of laser-captured adjacent neuronal subtypes,” which appeared in theJanuary issue (Nature Medicine 5, 120; 1999), a company namewas misspelled. The correct spelling is Arcturus Engineering.


199

9 N

atu

re A

mer

ica

Inc.

• h

ttp

://m

edic

ine.

nat

ure

.co

m

Documents

document