Prediction of in situ fluorescence of histochemical reagents using a structure-staining correlation procedure

Summary A total of ninety acid, basic, and non-ionic dyes were screened for fluorescent staining of various Carnoy fixed rat tissues. It was found that the fluorescence/nonfluorescence of a dye could be predicted using a conjugated bond number (CBN) cut-off value. Thus 90% of dyes with CBNs of 29 or less were fluorescent; whilst 70% of dyes whose CBNs exceeded 30 were nonfluorescent. The cut-off value was not significantly influenced by the charge, or the hydrophobic-hydrophilic character of the dye; though fluorescence was greatly influenced by the mode of fixation. The CBN cut-off value proved surprisingly robust. Thus most fluorochromes found in the histochemical literature have small conjugated systems, with CBNs less than the cut-off value. This includes labels of immunoglobulins, vital stains of neurones, and fluorescent Schiff reagents. Conversely several dyes used to quench background autofluorescence have large conjugated systems, with CBNs substantially above the cut-off value.In honour of Prof. P. van Duijn     

Selective labeling of lipid droplets in aldehyde fixed cell monolayers by lipophilic fluorochromes

Abstract

We evaluated a number of lipophilic dyes and fluorochromes, including oxazone and thiazone derivatives of oxazine and thiazine dyes, scintillator agents, a carotenoid and a metal-porphyrin complex for visualization of lipid droplets within aldehyde fixed cultured HeLa and BGC-1 cells. Observation under ultraviolet, blue or green exciting light revealed selective fluorescence of lipid droplets, particularly after treatment with aqueous solutions of Nile blue and brilliant cresyl blue oxazones, toluidine blue thiazone, or propylene glycol solutions of canthaxanthin, ethyl-BAO, and ZnTPyP. Mounting in water was required to maintain the fluorescence of lipids; the use of glycerol, Mowiol or Vectashield was not adequate. The effect of dye structure on staining intensity was assessed with the aid of numerical structure parameters modeling lipophilicity (HI and log P), overall size (MW) and the size of the conjugated system (conjugated bond number; CBN). The best stains for lipid droplets were relatively lipophilic (HI > 4.0, log P > 5.0), of small size overall (MW < 370), with small conjugated systems (CBN < 24), and not significantly amphiphilic. The two hydrophobic-hydrophilic parameters (the classic log P and the hydrophobic index, HI; values calculated by molecular modeling software) were highly correlated; however, HI was a more suitable hydrophobicity index for the dyes studied here.     

Quantitative comparison of long-wavelength Alexa Fluor dyes to Cy dyes: fluorescence of the dyes and their bioconjugates.

Amine-reactive N-hydroxysuccinimidyl esters of Alexa Fluor fluorescent dyes with principal absorption maxima at about 555 nm, 633 nm, 647 nm, 660 nm, 680 nm, 700 nm, and 750 nm were conjugated to antibodies and other selected proteins. These conjugates were compared with spectrally similar protein conjugates of the Cy3, Cy5, Cy5.5, Cy7, DY-630, DY-635, DY-680, and Atto 565 dyes. As N-hydroxysuccinimidyl ester dyes, the Alexa Fluor 555 dye was similar to the Cy3 dye, and the Alexa Fluor 647 dye was similar to the Cy5 dye with respect to absorption maxima, emission maxima, Stokes shifts, and extinction coefficients. However, both Alexa Fluor dyes were significantly more resistant to photobleaching than were their Cy dye counterparts. Absorption spectra of protein conjugates prepared from these dyes showed prominent blue-shifted shoulder peaks for conjugates of the Cy dyes but only minor shoulder peaks for conjugates of the Alexa Fluor dyes. The anomalous peaks, previously observed for protein conjugates of the Cy5 dye, are presumably due to the formation of dye aggregates. Absorption of light by the dye aggregates does not result in fluorescence, thereby diminishing the fluorescence of the conjugates. The Alexa Fluor 555 and the Alexa Fluor 647 dyes in protein conjugates exhibited significantly less of this self-quenching, and therefore the protein conjugates of Alexa Fluor dyes were significantly more fluorescent than those of the Cy dyes, especially at high degrees of labeling. The results from our flow cytometry, immunocytochemistry, and immunohistochemistry experiments demonstrate that protein-conjugated, long-wavelength Alexa Fluor dyes have advantages compared to the Cy dyes and other long-wavelength dyes in typical fluorescence-based cell labeling applications.     

Studies of polypeptide structure by fluorescence techniques. 3. Interaction between dye and macromolecule in fluorescent conjugates

The effects of pH on the polarization of fluorescence of dyes dissolved in media of high viscosity or conjugated to polypeptides that undergo no structural transitions indicate that DNS is useful for studying pH-dependent molecular transition over the range pH 2.5–14, whereas fluorescein is useful only over the range pH 6–8. Heating and cooling in aqueous solutions cause no change in the polarization of fluorescein or of DNS; therefore, the dyes themselves do not introduce artifacts into heating studies of the dye conjugates. The interaction between fluorescein or DNS and the molecule to which it is conjugated varies and thus may affect the measurements made with the conjugates: the rotational relaxation times of polylysine, of a copolymer of glutamic acid and lysine, and of lysozyme are approximately twice as long when measured with DNS-conjugates as when measured with fluorescein-conjugates. The explanation for this observation is postulated to lie in the tighter binding between fluorescein and the molecule to which it is conjugated, presumably around the point of its covalent attachment, which makes it a better indicator of the behavior of the rotational kinetic unit of the polypeptide chain. The stronger binding of fluorescein is inferred from two lines of evidence: (1) the fluorescent intensity and ultraviolet spectra of a fluorescein–polylysine conjugate are less susceptible to changes in solvent than those of the DNS conjugate, and (2) the net charge of the polypeptide affects the ionization of fluorescein much less than it affects the ionization of DNS. Additional evidence from previous studies corroborates this conclusion. Thus, it is important to establish the relationship between the fluorescent dye and the molecule to which it is conjugated before using the fluorescence data to calculate rotational relaxation times and other molecular parameters.     

Synthesis of thiol-reactive, long-wavelength fluorescent phenoxazine derivatives for biosensor applications

Two environmentally sensitive, long-wavelength fluorescent phenoxazine derivatives, INR and IANR, were synthesized with linkers for conjugation to the thiol group of cysteine in binding proteins. The linkers were designed based on the attachment sites at two different positions on the phenoxazine, which were chosen in order to study the orientation of the dye with respect to the binding protein. Conjugation of the dyes to the S337C maltose binding protein (MBP) mutant provided conjugates of these dyes that are capable of detecting maltose with different sensitivities. The dye INR gave a 3-fold (+200%) change in fluorescence intensity upon maltose binding when conjugated to S337C MBP with a binding constant (K(d)) of 435 microM. The fluorescence change for IANR was only 20% and the K(d) was 1.4 mM. Conformational analysis of the dyes by molecular modeling suggested that the linker in IANR imparted greater conformational freedom to the dye, resulting in little change in environment between the open and the closed-form conformations. The linker in INR, on the other hand, showed restricted motion, which placed the dye in different environments in the open and closed forms of the protein. Thus, design and placement of the linker play a critical role in the performance of these dyes as environmentally sensitive probes.     

Membrane potential can be determined in individual cells from the nernstian distribution of cationic dyes.

The distribution of a selection of cationic fluorescent dyes can be used to measure the membrane potential of individual cells with a microfluorometer. The essential attributes of these dyes include membrane permeability, low membrane binding, spectral properties which are insensitive to environment, and, of course, strong fluorescence. A series of dyes were screened on HeLa cells for their ability to meet these criteria and several commercially available dyes were found to be satisfactory. In addition, two new dyes were synthesized for this work by esterification of tetramethyl rhodamine. The analysis of the measured fluorescent intensities requires correction for fluorescence collected from outside the plane of focus of the cell and for nonpotentiometric binding of the dye. The measurements and analysis were performed on three different cell types for which there exists a body of literature on membrane potential; the potentials determined in this work were always within the range of literature values. The rhodamine esters are nontoxic, highly fluorescent dyes which do not form aggregates or display binding-dependent changes in fluorescence efficiency. Thus, their reversible accumulation is quantitatively related to the contrast between intracellular and extracellular fluorescence and allows membrane potentials in individual cells to be continuously monitored.     

Optimization of staining conditions for microalgae with three lipophilic dyes to reduce precipitation and fluorescence variability

When the fluorescence signal of a dye is being quantified, the staining protocol is an important factor in ensuring accuracy and reproducibility. Increasingly, lipophilic dyes are being used to quantify cellular lipids in microalgae. However, there is little discussion about the sensitivity of these dyes to staining conditions. To address this, microalgae were stained with either the lipophilic dyes often used for lipid quantification (Nile Red and BODIPY) or a lipophilic dye commonly used to stain neuronal cell membranes (DiO), and fluorescence was measured using flow cytometry. The concentration of the cells being stained was found not to affect the fluorescence. Conversely, the concentration of dye significantly affected the fluorescence intensity from either insufficient saturation of the cellular lipids or formation of dye precipitate. Precipitates of all three dyes were detected as events by flow cytometry and fluoresced at a similar intensity as the chlorophyll in the microalgae. Prevention of precipitate formation is, therefore, critical to ensure accurate fluorescence measurement with these dyes. It was also observed that the presence of organic solvents, such as acetone and dimethyl sulfoxide (DMSO), were not required to increase penetration of the dyes into cells and that the presence of these solvents resulted in increased cellular debris. Thus, staining conditions affected the fluorescence of all three lipophilic dyes, but Nile Red was found to have a stable fluorescence intensity that was unaffected by the broadest range of conditions and could be correlated to cellular lipid content.     

Dye with low specificity to nucleotide sequences of DNA: use for assessing the quantity of oligonucleotides, immobilized in cells of biological microchips

To assess the DNA amount in samples (e.g., in biological microchip gel pads) by means of fluorescent dyes, one should use the dyes whose fluorescence weakly depends on DNA composition and structure. With the ImD-310 dye created for this purpose, we have analyzed the staining of single- and double-stranded oligo- and polynucleotides of different nucleotide composition, length, and concentration both in solution and being immobilized in biological microchip gel pads. It turned out that ImD-310 has no pronounced specificity to the single- and double-stranded nucleotide sequences, while the intensity of fluorescence for the dye complexes with d(A)8, d(T)8, d(C)8, and d(G)8 at high temperatures (50 degrees C) differs by less than 25%. A linear correlation has been established between the intensity of fluorescence and the amount of oligonucleotides immobilized on a biological microchip. The plots of the intensity of fluorescence against the concentration of NaCl and the temperature were obtained. By using a generic microchip containing all 4096 hexamer oligonucleotides, it has been determined that the dye has no distinct specificity to any certain motifs of the nucleotide sequence. Thus, ImD-310 may serve as an efficient fluorescent probe to quickly estimate the amount of oligonucleotides immobilized in a microchip, in an electrophoretic gel, etc.     

A Fluorescent Dye with Low Specificity to DNA Nucleotide Sequences: Quantitative Assessment of Oligonucleotides Immobilized in Microchip Gel Pads

To assess the DNA amount in samples (e.g., in biological microchip gel pads) by means of fluorescent dyes, one should use the dyes whose fluorescence weakly depends on DNA composition and structure. With the ImD-310 dye created for this purpose, we have analyzed the staining of single- and double-stranded oligo- and polynucleotides of different nucleotide composition, length, and concentration both in solution and being immobilized in biological microchip gel pads. It turned out that ImD-310 has no pronounced specificity to the single- and double-stranded nucleotide sequences, while the intensity of fluorescence for the dye complexes with d(A)₈, d(T)₈, d(C)₈, and d(G)₈ at high temperatures (50°C) differs by less than 25%. A linear correlation has been established between the intensity of fluorescence and the amount of oligonucleotides immobilized on a biological microchip. The plots of the intensity of fluorescence against the concentration of NaCl and the temperature were obtained. By using a generic microchip containing all 4096 hexamer oligonucleotides, it has been determined that the dye has no distinct specificity to any certain motifs of the nucleotide sequence. Thus, ImD-310 may serve as an efficient fluorescent probe to quickly estimate the amount of oligonucleotides immobilized in a microchip, in an electrophoretic gel, etc.     

Differences in the red fluorescence of acridine orange bound to single-stranded RNA and DNA.

Fluorescence of acridine orange bound to RNA or DNA in the single-stranded form including single-stranded synthetic polyribo- or polydeoxyribonucleotides was measured in the expectation that some distinct structural characteristic between single-stranded RNA and DNA might be reflected by a specific fluorescent behaviour of bound dyes. It was found that the complex of the dye with single-stranded RNA emits a weaker red fluorescence around 650 nm than the complex with single-stranded DNA at low phosphate-to-dye ratios. The fact could be explained neither by a direct interaction of bound dyes with the 2′-hydroxyl group of ribose in RNA nor by the difference in the G-C content of the nucleic acids. On the basis of the character of dye molecules emitting the red fluorescence, it was suggested that the bases in single-stranded RNA might be buried in some hydrophobic environment that would make the dyes less likely to interact with them, compared with the bases in single-stranded DNA. It was further inferred that some conformational rigidity of single-stranded RNA may partially be responsible for the weaker red fluorescence.     

Ligands with affinity to specific sequence of DNA base pairs. XII. The synthesis and cytological studies of dimeric Hoechst 33258 molecules

We synthesized dimeric Hoechst dye molecules composed of two moieties of the Hoechst 33258 fluorescent dye phenolic hydroxy groups of which were tethered via pentamethylene, heptamethylene, or triethylene oxide linkers. A characteristic pattern of differential staining of chromosome preparations from human premonocytic leukemia HL60 cells was observed for all the three fluorescent dyes. The most contrast pattern was obtained for the bis-Hoechst analogue with the heptamethylene linker; its quality was comparable with the picture obtained in the case of chromosome staining with 4',6-diamidino-2-phenylindole. The ability to penetrate into the live human fibroblasts was studied for the three bis-Hoechst compounds. The fluorescence intensity of nuclei of live and fixed cells stained with the penta- and heptamethylene-linked bis-Hoechst analogues was found to differ only slightly, whereas the fluorescence of the nuclei of live cells stained with triethylene oxide-linked bis-Hoechst was considerably weaker than that of the fixed cells. The bis-Hoechst molecules are new promising fluorescent dyes that can both differentially stain chromosome preparations and penetrate through cell and nuclear membranes and effectively stain cell nuclei.     

Suitable labels for molecular imaging--influence of dye structure and hydrophilicity on the spectroscopic properties of IgG conjugates

Aiming at the design of highly brilliant NIR emissive optical probes, e.g., for in vivo near-infrared fluorescence imaging (NIRF), we studied the absorption and fluorescence properties of the asymmetric cyanines Dy678, Dy681, Dy682, and Dy676 conjugated to the model antibody IgG. The ultimate goal was here to derive general structure-property relationships for suitable NIR fluorescent labels. These Dy dyes that spectrally match Cy5 and Cy5.5, respectively, were chosen to differ in chromophore structure, i.e., in the substitution pattern of the benzopyrylium end group and in the number of sulfonic acid groups. Spectroscopic studies of the free and IgG-bound fluorophores revealed a dependence of the obtained dye-to-protein ratios on dye hydrophilicity and control of the fluorescence quantum yields (Φ(f)) of the IgG conjugates by the interplay of different fluorescence reduction pathways like dye aggregation and fluorescence resonance energy transfer (FRET). Based upon aggregation studies with these dyes, the amount of dye dimers in the IgG conjugates was determined pointing to dye hydrophilicity as major parameter controlling aggregation. To gain further insight into the exact mechanism of dye dimerization at the protein, labeling experiments at different reaction conditions but constant dye-to-protein ratios in the reaction solution were performed. With Dy682 that displays a Φ(f) of 0.20 in PBS and 0.10 for moderate dye-to-protein ratio of 2.5, a low aggregation tendency, and a superior reactivity in IgG labeling, we identified a promising diagnostic tool for the design of NIR fluorescent probes and protein conjugates.     

Cyanine dye-protein interactions: looking for fluorescent probes for amyloid structures

We ascertained the ability to detect fibrillar beta-lactoglobulin (BLG) of a series of mono-, tri-, penta-, and heptamethinecyanines based on benzothiazole and benzimidazole heterocycles, and of benzothiazole squaraine. Fluorescence properties of these cyanine dyes were measured in the unbound state and in the presence of monomeric and fibrillar BLG and compared with those for the commercially available benzothiazole dye Thioflavin T. The correlation between the chemical nature of the dye molecules and the ability of dyes to bind aggregated proteins was established. We found that meso-substituted cyanines with amino substituents in heterocycle in contrast to the corresponding unsubstituted dyes have a binding preference to fibrillar BLG and a noticeable fluorescence response in the presence of the aggregated protein. For the squaraines and benzimidazole penthamethinecyanines studied, fluorescence emission increased both in the presence of native and fibrillar protein. The trimethinecyanines T-49 and SH-516 exhibit specifically increased fluorescence in the presence of fibrillar BLG. These dyes demonstrated the same or higher emission intensity and selectivity to aggregated BLG as Thioflavin T, and are proposed for application in selective fluorescent detection of aggregated proteins.     

Symmetric and asymmetric squarylium dyes as noncovalent protein labels: a study by fluorimetry and capillary electrophoresis

Noncovalent interactions between two squarylium dyes and various model proteins have been explored. NN127 and SQ-3 are symmetric and asymmetric squarylium dyes, respectively, the fluorescence emissions of which have been shown to be enhanced upon complexation with proteins such as bovine serum albumin (BSA), human serum albumin (HSA), beta-lactoglobulin A, and trypsinogen. Although these dyes are poorly soluble in aqueous solution, they can be dissolved first in methanol followed by dilution with aqueous buffer without precipitation, and are then suitable for use as fluorescent labels in protein determination studies. The nature of interactions between these dyes and proteins was studied using a variety of buffer systems, and it was found that electrostatic interactions are involved but not dominant. Dye/protein stoichiometries in the noncovalent complexes were found to be 1:1 for SQ-3, although various possible stoichiometries were found for NN127 depending upon pH and protein. Association constants on the order of 10(5) and 10(7) were found for noncovalent complexes of SQ-3 and NN127, respectively, with HSA, indicating stronger interactions of the symmetric dye with proteins. Finally, HSA complexes with NN127 were determined by capillary electrophoresis with laser-induced fluorescence detection (CE-LIF). In particular, NN127 shows promise as a reagent capable of fluorescently labeling analyte proteins for analysis by CE-LIF without itself being significantly fluorescent under the aqueous solution conditions studied herein.     

Symmetric cyanine dyes for detecting nucleic acids

A novel approach to the design of sensitive fluorescent probes for nucleic acids detection is proposed. Suitable modifications of tri- and pentamethine cyanine dyes in the polymethine chain and/or in the heterocyclic residues can result in a significant decrease in unbound dye fluorescence intensity and an increase in dye emission intensity in the presence of DNA compared to the unsubstituted dye. The sharp enhancement in the fluorescence intensity upon dye interaction with double-stranded DNA permits the application of the modified tri- and pentamethine dyes as fluorescent probes in double-stranded DNA detection in homogeneous assays.     

Polymerizable fluorescein derivatives: synthesis of fluorescent particles and their cellular uptake

Fluorescent particles are used for a diverse number of biochemical assays including intracellular imaging, cellular tracking, as well as detection of a variety of biomolecules. They are typically prepared by postpolymerization conjugations of dyes onto preformed particles. Herein we report the synthesis of aminomethyl-functionalized fluorescent particles via the synthesis and application of polymerizable fluorescein monomers. These monomers allowed high and controllable fluorophore loading into the particles, resulting in enhanced fluorescence properties in comparison with more commonly used carboxyfluorescein conjugated particles. Furthermore, the particles were rapidly taken up by cells with enhanced fluorescence. The herein presented results demonstrate the advantages of dye polymerization in contrast with more conventional conjugation strategies for fluorescent particle generation with applications in the life sciences.     

Masked red-emitting carbopyronine dyes with photosensitive 2-diazo-1-indanone caging group

Caged near-IR emitting fluorescent dyes are in high demand in optical microscopy but up to now were unavailable. We discovered that the combination of a carbopyronine dye core and a photosensitive 2-diazo-1-indanone residue leads to masked near-IR emitting fluorescent dyes. Illumination of these caged dyes with either UV or visible light (λ < 420 nm) efficiently generates fluorescent compounds with absorption and emission at 635 nm and 660 nm, respectively. A high-yielding synthetic route with attractive possibilities for further dye design is described in detail. Good photostability, high contrast, and a large fluorescence quantum yield after uncaging are the most important features of the new compounds for non-invasive imaging in high-resolution optical microscopy. For use in immunolabelling the caged dyes were decorated with a (hydrophilic) linker and an (activated) carboxyl group.     

Forskolin-Induced Clearance of the Fluorescent Dye Sulforhodamine from Rat Parotid Intralobular Duct Lumen: Visualization of the Secretory Function under a Confocal Laser Scanning Microscope

Cyclic AMP evokes fluid secretion with bicarbonate in exocrine ducts. Clearance of fluorescent dyes from rat parotid intralobular ducts by forskolin was visualized as a fluorescence change in the duct luminal space by optical sectioning under a confocal laser scanning microscope to clarify the secretory function in the ducts. When the isolated rat parotid intralobular duct segments were superfused with membrane-impermeable fluorescent dyes during the experimental period, fluorescent dyes were passively moved into the duct space. Forskolin and isobutylmethylxanthine decreased the fluorescence of anionic dye, sulforhodamine B, and neutral dye, dextran tetramethyl-rhodamine, in the duct space, suggesting that the forskolin-induced clearance of fluorescent dyes might be the result of fluid secretion in the ducts. Methazolamide inhibited a forskolin-induced sustained decrease in duct fluorescence and intracellular acidification. Low concentrations of external Cl?, DIDS, bumetanide and amiloride did not markedly inhibit a forskolin-induced decrease in duct fluorescence. These findings suggest that a major portion of the steady decrease in duct fluorescence by forskolin was related to intracellular HCO3? production, not the uptake mechanism of external Cl?. Glibenclamide, NPPB, DPC and DMA inhibited the forskolin-induced decrease. Forskolin evokes the clearance of fluorescent dyes from duct space possibly due to fluid secretion in rat parotid ducts, associated with secretion through CFTR and DPC-sensitive anion channels of carbonic anhydrase-dependent bicarbonate linked with the Na+/H+ exchange mechanism.     

Cyanine dye–protein interactions: Looking for fluorescent probes for amyloid structures

We ascertained the ability to detect fibrillar β-lactoglobulin (BLG) of a series of mono-, tri-, penta-, and heptamethinecyanines based on benzothiazole and benzimidazole heterocycles, and of benzothiazole squaraine. Fluorescence properties of these cyanine dyes were measured in the unbound state and in the presence of monomeric and fibrillar BLG and compared with those for the commercially available benzothiazole dye Thioflavin T. The correlation between the chemical nature of the dye molecules and the ability of dyes to bind aggregated proteins was established. We found that meso-substituted cyanines with amino substituents in heterocycle in contrast to the corresponding unsubstituted dyes have a binding preference to fibrillar BLG and a noticeable fluorescence response in the presence of the aggregated protein. For the squaraines and benzimidazole penthamethinecyanines studied, fluorescence emission increased both in the presence of native and fibrillar protein. The trimethinecyanines T-49 and SH-516 exhibit specifically increased fluorescence in the presence of fibrillar BLG. These dyes demonstrated the same or higher emission intensity and selectivity to aggregated BLG as Thioflavin T, and are proposed for application in selective fluorescent detection of aggregated proteins.     

Multi-photon microscopy of cell types in the viable taste disk of the frog

The morphology of viable taste disks of the frog was explored with multi-photon microscopy. In order to identify single sensory or supporting cells within the tissue, we searched for fluorescent dyes that stained subsets of the cell population or possibly cell types. Some cell types indeed stained preferentially with certain fluorescent dyes. A subset of glia-like cells (type Ic) stained with BCECF, a H⁺-sensitive dye, and indo-1, a Ca²⁺-sensitive dye, both presented in the membrane-permeant ester form. BCECF-ester also stained the dendrites of type III receptor cells, but indo-1 ester did not. Receptor cells of type II stained with MQAE, a positively charged Cl^–-sensitive dye. A subset of type II cells accumulated amiloride, a positively charged fluorescent diuretic. Certain supporting cells, i.e., wing cells (type Ib) and glia-like cells (type Ic), were labeled by negatively charged dyes, e.g., calcium green-1 dextran. Mucus cells (type Ia) were stained with only two of the 19 dyes examined, and Merkel-like basal cells (type IV) were stained only with a membrane-labeling voltage-sensitive dye, presumably by endocytosis. No dye was found which would stain all types of cells or all receptor cells. This finding reveals a potential problem for future functional imaging aiming at population responses, as the responses of unstained cells then would remain unobserved. Specificity of dyes with respect to cell types was sufficient to identify supporting cells and receptor cells. Cell shape could then be reconstructed, using optical slicing and rendering techniques. Thus populations of dye-loaded elongated cells, especially types Ic, II and III, could for the first time be visualized in three dimensions.This work was supported by the Deutsche Forschungsgemeinschaft (SFB 530, project B2)