STUDIES ON NUCLEIC ACID METACHROMASY I. The Effect of Certain Fixatives on the Dye Stacking Properties of Nucleic Acids in Solution

The stacking coefficients (K's) of nucleic acids have been thought to influence the color contrast between DNA and RNA in tissue sections stained with metachromatic dyes. This idea was tested by titrating toluidine blue (TB) and acridine orange (AO) in solution against DNA and RNA, native or treated with formaldehyde, acrolein, or Carnoy's fluid. Absorption spectra at varying polymer-dye ratios were used to compute K values by the methods of Bradley and colleagues. Results with both dyes fit Bradley's stacking equations. Fixatives did not block dye-binding sites but markedly altered K values. K of DNA was low, unaffected by aldehyde fixative, increased by Carnoy's fluid or heat denaturation. K of RNA was higher than that of DNA and was increased greatly by formaldehyde, almost as much by acrolein, considerably less by Carnoy's fluid. Aldehyde effects were partially reversed upon removal of aldehyde by dialysis. These observations accord with known effects of aldehydes and denaturation upon nucleic acid conformation. Differences between K's of DNA and RNA were greater after aldehyde treatment than after Carnoy's, and were greater with AO than with TB. This is generally consistent with the magnitude of the color contrasts observed in tissues. Additional factors must contribute to the intense color contrast observed in acrolein-fixed tissues stained with TB. I N T R O D U C T I O N When certain cationic dyes are used to stain tissue sections, some tissue polyanions acquire the same color shown by the dye in dilute solution, i.e. are stained orthochromatically, while other tissue polyanions acquire a different color, i.e. are stained metachromaticaily. Most studies of this phenomenon have been concerned with the large metachromatic color changes produced by acid mucopolysaccharides. However, similar though smaller color changes are produced in many of the same dyes by nucleic acids. The metachromatic color change has long been considered to result from reversible aggregation of dye molecules, which may occur when the concentration of dye is increased, the ionic strength is increased, the temperature is decreased, or when dye molecules are bound to neighboring sites on suitable polymers. For several representative metachromatic dyes, the formation of such aggregates has been proved (19, 30). The subject of metachromasy was comprehensively reviewed by Bergeron and Singer in 1958 (3). Some metachromatic dyes can be used to pro313 on O cber 9, 2017 jcb.rress.org D ow nladed fom duce a color contrast between DNA (deoxyribonucleic acid) and RNA (ribonucleic acid) 1 in tissue sections. Both nucleic acids will stain orthochromatically at one extreme and metachromatically at the opposite extreme of temperature, ionic strength, pH, or dye concentration, but intermediate conditions can be found under which DNA will stain orthochromatically, RNA metachromatically. This phenomenon is best recognized for the dye acridine orange (AO), which can be used to differentiate DNA from RNA in fluorescence microscopy (1, 2, 8, 9, 20, 28, 29). Analogous color contrasts have been observed with the more familiar thiazine dyes of ordinary light microscope histology (10, I 1, and references cited there) but have received less attention because they are less striking to the eye and more difficult to obtain consistently. Flax and Himes (11) showed that a consistent color contrast between DNA and RNA could be produced with the thiazine dye azure B by staining at high temperature (40°C), followed by prolonged differentiation. They stained sections at several dye concentrations, and showed that at each concentration the spectral shapes of stained DNA and RNA in the section were different. They suggested that "the greater metachromasy that occurs with RNA may be due to closer spacings of phosphate groups available for dye attachment in the RNA molecule," and discussed this suggestion in terms of the hypotheses current at that time regarding nucleic acid structure. In recent years, the usefulness of the aggregation concept has been extended by detailed studies (5-8, 25, 26) of the interaction of metachromatic dyes with polyelectrolytes in dilute solution at low ionic strength. Under these conditions, binding of dye to polymer is stoicbiometrically complete, and the concentration of dye, the concentration of polymer, and the molar extinction of the dye can all be measured with sufficient precision to allow physico-chemical analysis of the metachromatic color change. It has been shown that a group of polymers, on which a dye has the same metachro1Abbreviations employed: DNA, deoxyribonucleic acid; RNA, ribonucleic acid; AO, acridine orange; riB, toluidine blue; P/D, number of polymer dyebinding sites divided by number of bound dye naolecules; F1, that fraction of bound-dye molecules which are unstaeked; Era, molar extinction coefficient; kr~ax, wavelength of the peak in the absorption spectrum; K, stacking coefficient. matic spectrum when all binding sites are filled with dye molecules (P/D = 1) and the same orthochromatic spectrum when binding sites are present in great excess of the number of dye molecules (P/D ~ ~o), may induce quite different proportions of the two spectra at intermediate values of P/D. For example, at P/D = 5, AO has less than 50 per cent of the metachromatic spectral component when bound to DNA, about 70 per cent when bound to RNA, and nearly 100 per cent when bound to polyadenylic acid. These differences result from a non-random distribution of dyes among available sites, caused by an attractive interaction between those dye molecules which are bound to neighboring sites and therefore form a metachromatic aggregate. The decrease in free energy resulting from this interaction weights the distribution of dye in favor of such aggregates and is greater for some polymers (e.g. polyadenylic acid) than for others (e.g. DNA). These phenomena are collectively termed stacking in recognition of the shape of the aggregates thought to be formed by the planar, aromatic dye molecules. The changes of spectral shape and molar extinction with changing P/D are used to compute a number called the stacking coefficient, K, which is a measure of the strength of the attractive interaction, and bears direct thermodynamic relationship (5) to the loss of free energy on stack formation. K = 1 corresponds to zero free energy loss on stack formation and consequent totally random distribution of dyes among available sites. The greater the value of K, the more stacking is obtained at a given P/D, or, conversely, the higher P/D must be to obtain a given ratio of unstacked to stacked dye. Observed K values with AO range from 1.2 with DNA, on which 50 per cent of the dye is unstacked at P/D = 3.9, to several thousand for certain acid mucopolysaccharides on which P/D = 40,000 is required to obtain the same percentage of unstacking (26). Since the stacking coefficient has been found to be quite sensitive to known alterations of the secondary structure of polymers, quantitative analysis of dye stacking provides a tool for the investigation of polymer structure. The companion paper (10) describes a great enhancement of the color contrast between DNA and RNA in tissues fixed in acrolein, embedded in polyester wax, and stained with a metachromatic thiazine dye, toluidine blue (TB). If the action of acrolein on pure nucleic acids can be defined and 314 TIfF JOURNAL CF CELL BIOLOGY • Volume ~7, 1965 on O cber 9, 2017 jcb.rress.org D ow nladed fom compared wi th the act ion of other fixatives, i t might be possible to offer a quant i ta t ive interpreta t ion of this enhancem en t of the color contrast. In order to do this, the stacking experiments previously performed wi th A O must be matched by a parallel series of experiments using TB. W e have therefore compared the stacking of A O and TB by D N A and RNA, bo th in the nat ive state and after t r ea tment wi th formaldehyde, acrolein, or Carnoy 's fluid, the three fixatives whose effects at the histological level are compared in the compan ion paper. The two dyes show comparable , though not identical, stacking behavior. The effects of tile fixatives can all be expressed as changes in the stacking coefficient, K, resulting f rom alterations of nucleic acid secondary structure. T he differences between the K's of similarly t reated D N A and R N A with ei ther dye are largely consistent, in direct ion and magni tude , wi th the color contrasts obta ined histologically, suggesting tha t K is an impor t an t var iable influencing the color of bound metachromat ic dye in tissue sections.