This article is subject to Copyright.
A "5-Minute Food Color Stain" for Microscopy
by C. Thorsten
Abstract: Lack of food dyes' wide histologic use is often cited as proof of their inferiority. To evaluate this claim the author tried making an acceptable stain using only food dyes, water, alcohol, and dilute acetic acid. All staining was done "supravitally", on unfixed, temporary wet-mount samples. The dyes tested were blue food color (Brilliant Blue FCF) and red food color (a mixture of Allura Red AC and Erythrosine B). Results were better than expected.
"Red" food coloring (FD&C Red 40 + Red 3)
"Blue" food coloring (FD&C Blue 1)
Spot Plate, white polystyrene
Acetic acid, 5% (vinegar)
Isopropyl Alcohol (IPA), 91%
Droppers or Pasteur pipettes
Compound Microscope (such as an Observer III)
Coverslips, glass or plastic
Various specimens to examine
Eyepiece Camera (e.g., Mini-VID; optional)
Methods & Observations:
The microscope was an Observer III, having a maximum of 400x magnification. The eyepiece camera was a Mini-VID USB.
Since the writer was concurrently doing research in the area of kidney physiology, one of the sample types studied was urine. It was provided by an apparently-healthy volunteer with no history of nephropathy. The sample was collected and examined using a slight variant of standard clinical technique: 12 mL collected first-morning, clean-catch, mid-stream; centrifuged in an Ultra-8V at about 3000 rpm for 10 minutes1; supernatant liquid removed down to ca. 0.5 mL; sediments resuspended by agitation.
Other microscopy samples examined in this experiment included swabs of nasal mucosa and scrapings of inner cheek cells.
The stain was made from "red" and "blue" food color mixed with other ingredients; Table 1 shows some of these combinations. No attempt was made to include every possible permutation of 0 to 5 drops of each ingredient; a representative cross-section was instead the goal. Several intermediate combinations were made by simply adding additional drops of vinegar to Version 1.1. (these formed versions 1.2 through 1.6, not shown in table). None of these was found to be any better than 1.0 or 1.1; thus, no photographs were taken.
Ambient lab temperature was 14 ºC. All specimens were presumably at the same temperature by the time they were photographed.
Table 1. Representative formulations tried
Stain in most cases was applied supravitally, i.e. directly to the wet samples with no fixing, dehydration, or destaining. Coverslips were applied prior to microscopic examination. A couple specimens were, however, mounted permanently (fixing, staining, dehydration, mounting in synthetic medium).
Stain version 1.0 (no alcohol) had a muddy, brown-gray and purple appearance in bulk, liquid form. The other versions contained alcohol and were more homogeneously purple; while these were much less silty, most of them were no better than Version 1.0 in practice. There was, however, some differential staining effect on various cloth fibers, debris, and dust that were tested. Cell nuclei were generally no more visible than the rest of the cytoplasm, although they did sometimes become darker on standing. As expected, dye performance also depended on the cell's surroundings (solitary in water, clumped with other cells, embedded in mucus; etc).
Too much vinegar seemed to ruin any differential staining effect, although certain fiber types received the same coloration through a wide pH range.
When first prepared, the stain did exactly what the food coloring box suggested: "blue" and "red" made either "purple" or "magenta", depending on proportions. In special circumstances this homogeneous color dissociated to give some useful differentiation. Sometimes there was even near-complete separation in a manner reminiscent of chromatography. For example, pure blue was observed to have concentrated in the inside of a few large areas of mucus; it also occurred in larger masses of cheek cells.
Figure 1 shows a transitional epithelial cell stained with the completely aqueous version of the "5-minute food color stain" (1.0). Organelle differentiation is remarkably good, considering the concoction was made with supermarket goods. The staining is a slightly dark, but this could be modified by using shorter staining time or less of the stain.
Figure 2 showcases the experiment's best differential staining. The large, purple-stained structures are of botanical origin2 and represent contaminants that fell into the sample. The small, round, magenta object (C.) is a white blood cell, perhaps a granulocyte. "B." is one of numerous bacteria that grew in the sample when it was allowed to stand, although some of the dark-stained spots in the photo may be dust or micro-crystals. "D." is a uric acid crystal; these consistently stained a pale pink-violet, while the few observed calcium oxalate crystals took no stain.
The results suggested no inherent inferiority of food dyes for microscopy, at least for temporary wet-mount use. In a few instances the Five-Minute Food Color Stain performed remarkably well. Even better performance seemed within reach of minor changes in recipe, specimen preparation, and so forth. (Update: It was found later that a certain fixative / mordant gave startling results with this stain; additional experiments may follow)
The dyes selected were certainly common, but they were chosen also because of structural similarities to well-known microscopy dyes.
The blue food color, made of Brilliant Blue FCF, is a dye with ponderous molecular structure and three sulfonate groups (q.v. the Merck Index or Wikipedia entries); two of these should be sterically accessible for protein binding.
The red food color provides two dyes, Allura Red AC and Erythrosin B; both also are "acidic" dyes, able to lose one or more H+ and become anions. The Allura Red is an azo dye, but like Acid Blue 9 it also has sulfonate groups. The Erythrosin has a carboxylate group and a phenolic OH, both of which can ionize negatively; this dye is chemically almost identical to Eosin Y, one of the most often-used dyes in microscopy.
Adjusting the pH of the stain had a noticeable effect on its behavior toward cells in general. While pH-dependence of dye affinity is common knowledge in science, the real-life outcome a particular pH adjustment on a stain's behavior can be hard to predict. Success often relies as much on trial-and-error or on past experience as on theory.
In this case, too-acidic pH appeared to ruin the stain's affinity for cell nuclei and other structures. This habit was consistent with theory; at low enough pH, one might expect over-protonation, leading to poor dye binding. In other words, the basic side chains of proteins become protonated (e.g., --NH2 giving --NH3+), but so do the acidic groups on the dye molecules (e.g., --SO3- becoming --SO3H). This interferes with the charge attraction between dye and substrate.
At some ideal pH, the basic side chains on target proteins would be protonated as much as possible, while the acidic groups on the dyes would still be mostly unprotonated and free to bind to the proteins. While it is impossible to make every side chain ionize according to our wishes, the hope is to manipulate pH to achieve as much of a charge difference as real-life conditions permit.
Generally, one can arrive at a target pH by knowing or at least estimating the pKa values of the dye(s) used and of the protein side chains to be colored. Naturally, the real-life results will differ from theory due to a number of factors; for example, the side chain pKa of an amino acid incorporated in a protein will be different from that of the free amino acid. Steric hindrance and other variables can also have considerable effect.
Since the combination of "red" and "blue" create a homogeneously purple solution that colors many objects purple as well, it seems natural to assume some kind of interaction between dye species. While there is no reason expect any kind of permanent dye-dye adduct3, there exists a common phenomenon called "dye stacking", in which multiple dye molecules form temporary aggregates (Daruwalla, 1974). Van der Waals forces and hydrogen bonding play an important role (1974), although characterizing transient dye complexes is difficult. J. M. Williams (1992) points out that H-bond proton transfer is one of the fastest of all chemical reactions known; even with sophisticated equipment at one's disposal, the structures obtained are still just averages or composites of multiple states (1992).
Lacking the most sophisticated instruments costing tens of thousands (and upward), the investigator might use a spectrophotometer to look for absorption maxima that are different from any of the component dyes. One could also look for multiple crystal habits and colors during slow evaporation. Certain dyes are difficult to crystallize and will come out of solution as amorphous-granular masses, but this fact itself can be a minor clue to what is happening.
The formulation tried here was macroscopically purple and stained many objects purple, yet there was sporadic separation of blue dye in certain areas. There were also fibers that stained only pink. Thus, it seems likely that transient "super-molecules" were indeed floating around in the experiment, coming apart under the right conditions.
Metachromasy -- a phenomenon in which, for example, a blue dye stains certain objects pink or some other, unexpected color-- is of course an alternate (or simultaneous) possibility.4 For example, this writer has found aqueous eosin Y to show apparent metachromatic behavior in certain, binucleated "umbrella cells": the cytoplasm and nuclei in general stain pink, chromatin stains purple, and nucleoli stain either red or purple.
Aside from opening many avenues of further study, this mini-experiment shows the potential for biological stains using little more than food coloring. Fine adjustments of the pH of the dye solution and/or the use of standard histo recipe components (such as glycerol, methanol, lactic acid, etc.5) might be useful. Starting with nothing more than some colored liquid obtained from the shelves of a supermarket, one can open the door to a host of complex and poorly-understood phenomena.
Daruwalla, E.H. "Physical Chemistry of Dyeing: State of Dye in Dyebath and in Substrate". In The Chemistry of Synthetic Dyes, ed. K. Venkataraman. New York: Academic Press, 1974.
Williams, J.M. "Hydrogen bond". In McGraw-Hill Encyclopedia of Chemistry, ed. Sybil B. Parker. New York: McGraw-Hill, 1992. p. 506.
1 CLIA-recommended setting for urine is 400 x g, which in the Ultra 8 series of centrifuge translates to 1,800 rpm. Other sources recommend anywhere from 1,000 to 3,000 rpm for urine sample spun in tabletop units of comparable size. In the Ultra 8 series, this upper range would be approaching the CLIA speed for blood (3,400 rpm). There are technical pros and cons for various rotor speeds as applied to urinalysis, but here it doesn't matter anyway. This is a demonstration experiment, not someone's clinical lab workup.
Please note: The choice of sample depends on the experimenter's lab competence; readers lacking sufficient experience in clinical hygiene may want to stick to cheek cells, plants, etc. There is of course the possibility that the experimenter could use his or her own urine, obviating the need for worry about infection from others. Urine, though technically not sterile, is widely considered to present a much lower risk of disease transmission than blood. This lab writeup does not constitute a recommendation that the reader work with samples of urine, blood, etc., if he / she is not technically competent.
Back to article
2 This writer is no botanist, but the fibers appear to be tracheids from a conifer of some type. If a reader can definitively identify the plant species before the author tracks it down in a textbook, CR Scientific will send them $10 in free microscope accessories (slides, coverslips, vials for storing stains). Free shipping. Email your answers to us.
Back to article
3 Such is unlikely to happen at room temperature with just food colors and some vinegar in a spot plate well. Still, not all organic chemistry needs hours of reflux at 90 ºC to happen. In the present example we might look for reactive or comparatively weak bonds such as the azo linkage (--N=N--). Chemical changes could take place if azo dyes were left in sunlight for a while or allowed to stand in the presence of certain catalysts.
There do exist reactive dyes that form covalent adducts with various substrates, but the dyes in this experiment are not of that type. The noncovalent adduct "methylene blue eosinate", on the other hand, is formed by charge attraction of two dyes and is analogous to a salt; the anionic eosin molecule combines with the cationic methylene blue.
Back to article
4 The term "metachromasy" or "metachromasia" normally applies to a single-component dye. The literature contains numerous reviews of metachromasy, especially in the histological and histochemical journals. Back to article
5 With attention paid to any possible chemical incompatibilities, of course. Review the MSDS or Merck Index entry for any new reagents tried. Be careful with methanol; don't forget that it's usually present in denatured alcohol as well.
Back to article