The accurate measurement of DNA content in the cell is important both for clinical and basic research. However (or therefore), it seems that nowadays almost any reading of fluorescence or absorption of DNA-tropic dyes (beyond the domain of apoptosis) tends to be interpreted as genome size signal. This critique is especially relevant for small intraspecies variation and detection of chromosome abnormalities. At a low level of DNA content variation, variability in the structure of chromatin, which can affect the binding of DNA-tropic ligands, may be a significant component of the signal. In the extreme case of using cross-linking fixatives, fluorescence intensity of the same nuclei, fixed before and after the damage of cell membrane, can differ more than 2-fold (1).
As an exemplar case, intraspecies differences were studied for several mouse lines and their hybrids by means of absorption cytometry with Feulgen reagent and flow cytometry with propidium iodide (PI) staining (2). The authors of this otherwise very careful work made a conclusion that, notwithstanding the higher random error observed with the Feulgen technique, the results of both methods were in a generally good agreement (and interpreted them as genome size signal). However, regression analysis shows that there is a systematic (non-random) disagreement between the results obtained with these techniques, the degree of which depends on the signal intensity (Fig. 1A,B). The Feulgen data are consistently lower than the PI data, and this discrepancy is diminishing with increase of the signal. The intercept of the regression line is significantly negative and the slope is significantly greater than unity (Fig. 1A,B).
Which of the data give the accurate genome size values? It is plausible that the Feulgen data depend more strongly on the structure of chromatin since acid hydrolysis required for this technique is known to be especially sensitive to chromatin compactness (3-5), and the results of absorption cytometry depend on the opaqueness of the nuclei (6,7). In this case, chicken red blood cells were used as standard, and peripheral lymphocytes were taken as representatives of mouse cells. It was shown previously that for the chicken erythrocytes-mouse thymocytes pair, the greater opaqueness of the thymocytes (due to their higher DNA-content) can lead to the under-evaluation of DNA-content in these cells by absorption cytometry with Feulgen reagent (6). The negative intercept of the regression line (Fig. 1A,B) can be explained by this effect. However, since the difference between the Feulgen and PI data is gradually diminishing with the increase in signal, it could be suggested that this opaqueness (i.e. DNA con-