Correlation Between Barr Body and Overall Chromatin Decondensation in Vitro

J. Cell Set. 47, 187-195 (> 981) 187 Printed in Great Britain © Company of Biologist! Limited 1981

CORRELATION BETWEEN BARR BODY AND OVERALL CHROMATIN DECONDENSATION IN VITRO

M. GRATTAROLA.f A. BELMONT AND C. NICOLINI* Division of Biophysict, Department of Physiology/Biophysics, Temple University, Philadelphia, Pennsylvania, U.S.A.

SUMMARY Geometric and densitometric properties of the Barr body of early and late phase II confluent human fibroblasts are analysed by the automated image analyser Quantimet 720 D. In cells with the same 2C DNA content, the state of condensation of the Barr body varies proportionally with the state of condensation of the entire nucleus yielding a correlation between nuclear and Barr body area. In light of these results, indicating the participation of the Barr body in the overall process of chromatin condensation and decondensation, a definitive 'static' separation between dense ('heterochromatic') and dispersed (' euchromatic') regions of chromatin seems to be arbitrary. The implications of these results in terms of a possible attachment of interphase chromatin to the nuclear envelope are briefly discussed.

INTRODUCTION The division of interphase chromatin into two distinct regions, a dense 'heterochromatic' region and a dispersed 'euchromatic' region, and the often suggested correspondence of the dense region with transcriptionally inactive chromatin and the dispersed region with genetically active chromatin is still a matter of great controversy as well as confusion. For a review see Comings (1972), Maclean & Hilder (1977) or Back (1976). Since this generalization was made, large modulations in the morphology of the entire nucleus have been shown to occur as a function of the physiological state in a variety of cell systems, including changes in cell cycle phase (Nicolini, 1980; Kendall et al. 1977 a), virus transformation (Kendall, Beltrame & Nicolini, 1979), serum stimulation of human diploid fibroblasts (Kendall, Wu, Giaretti & Nicolini, 19776; Nicolini, Kendall, De Saive & Giaretti, 1977), 'activation' of hen erythrocytes (Bolund, Ringertz & Harris, 1969), PHA stimulation of human lymphocytes (Vonder- heid et al. 1980), and cell aging (Sawicki, 1979). In particular, nuclear condensation in these systems (as measured by the integrated optical density, proportional to DNA content, divided by nuclear area - i.e. the average optical density) was shown to relate

• To whom reprint requests should be addressed. f Present address: Sezione di Ingegneria Biofisica, Istituto di Elettrotecnica, Genoa University, Genoa, Italy. 188 M. Grattarola, A. Belmont and C. Nicolini closely with increased condensation of nuclear chromatin as revealed by optical and electron microscopy, as well as alterations in chromatin structure as measured by template activity, circular dichroism, and dye intercalation in these systems. Specifi- cally, increasing nuclear condensation was correlated with decreasing template activity, dye intercalation, and molar ellipticity at 272 nm (Nicolini, 1979). Exactly opposite results were observed for nuclear decondensation (relaxation). Such observations imply modulations in the chromatin structure of the entire genome occurring with changes in the physiological state of the cell. Meanwhile several independent investigations of the Barr body, which has long been considered a classic example of heterochromatin, have reported that in a fraction of the entire cellular population studied the Barr body is not detectable (Schwarzacher, 1963; Mittwoch, 1967), and that the appearance of Barr bodies correlated with an increase in cell density (Klinger, Daview, Goldhuber & Ditta, 1968; Mokherjee, Moser & Witowsky, 1972; Mokherjee & San Sebastian, 1978). These results suggested an interpretation in which the Barr body participated in the overall modulations in chromatin structure associated with cell growth, either relaxing to a density below the level of visual detection (Schwarzacher, 1963) during certain phases of the cell cycle or further condensing with cell confluency (Klinger et al. 1968). In this paper we have explored quantitatively such a possibility verifying that, indeed, in human diploid fibroblasts the state of condensation of the Barr body varies proportionally with the overall state of condensation of the entire nucleus.

MATERIALS AND METHODS Cell culture Human diploid fibroblasts (WI-38) obtained from the American Type Cell Culture Collec- tion (Rockville, Maryland) were used in all experiments. Cells were routinely grown at 37 °C with Eagle Basal Medium supplemented with 10 % foetal calf serum (GIBCO) in 75 -cm1 Corning plastic culture flasks. Early phase II cells were seeded onto coverslips inside small Petri dishes and then fixed (3 :1 mixture of 80 % ethanol and acetic-acid) at 2, 5 and 10 days after seeding. In parallel, late phase II cells were grown under the same conditions and fixed 5 days after seeding. After fixation the cells were Feulgen stained directly on the coverslip, at the optimal hydrolysis conditions of 5 N HC1, at 37 °C, for 25 min (Linden, Fang, Zietz & Nicolini, 1979).

Image analysis Nuclear images were magnified by a Zeiss Ultraphot microscope equipped with a 100 x oil-immersion planar achromat of 1-25 NA. Illumination was via a 100-W tungsten halogen light source equipped with a 540-nm filter (half-band width of 40 ran) and a condenser of 13 NA. The image was registered on a plumbicon scanner by means of a Reichert high quality magnification changer. Total magnification was 1250 x . Image analysis was performed using a Quantimet 720 automated image analyser (Cambridge Corp., New York) equipped with a 720-D densitometer. The scanner area is divided into 880 x 588 elements whose optical density can be digitized into 64 grey levels. The linear dimension of the approximately square picture element was determined to be 0-08 /tm by means of a stage micrometer (American Optical). A blank area of each slide was used to set the shade corrector and to calibrate the densitometer by means of neutral density filters. The linearity of the densitometer was verified using a neutral density wedge filter (Esco Products, New Jersey). Two types of measurements were made - those where the feature was the entire nucleus and those where the feature was the Barr body. For nuclear measurements, data were acquired on Barr body and chromatin 189 line with a digital computer. A uniform threshold of 0-03 O.D. was used for all slides to define the nuclear border. Field uniformity tests on a single nucleus positioned in 9 locations around the field yielded for all slides coefficients of variations of less than 2-5 % for integrated optical density (I.O.D. - proportional to total DNA content due to the stoichiometry of the Feulgen reaction) and 1 % for area. Variation of either parameter was less than 0-5 % for 10 consecutive measurements of a single image in the centre of the field. Measurements of the Barr body were made manually. The nuclear image was positioned at the centre of the field and the Barr body was centred in a square frame (30 x 30 picture points in dimension, corresponding to 2-5 x 2-5 /tm2) electronically generated. Two methods of measurement were used. First the optical density threshold was adjusted until the border of the detected region corresponded to the border of the Barr body (as judged by eye). This threshold was recorded as the optical threshold, by eye identification. Using this threshold value, the area and I.O.D. values (proportional to the DNA content of the Barr body whose border is defined visually) of the detected feature inside the 30 x 30 picture point square frame were recorded. (As the Barr body was nearly always denser than the area in its immediate vicinity, these measurements actually corresponded to the area and I.O.D. of the actual Barr body - that is the only significant detected feature inside the square frame was that of the Barr body.) Subsequently the optical density threshold was readjusted until the I.O.D. of the detected feature inside the square frame most nearly matched 2-66 % of the integrated optical density of the entire nucleus; this fraction (266%) corresponds to the fractional DNA content of an X chromosome as reported in the literature (Mendelsohn, 1979). This threshold was then recorded as the optical density threshold by I.O.D. constraint, and the area of the detected feature inside the square frame was redetermined using this threshold. Essentially this second method defined the border of the Barr body through the constraint on I.O.D.

RESULTS Two hundred nuclei from each slide were chosen randomly for measurement of the integrated optical density (I.O.D., proportional to the total amount of DNA present) and area of each nucleus. The area histograms corresponding to the subpopulations from these slides with 2C DNA content are shown in Fig. 1 and confirm previous findings on the increased nuclear condensation occurring with time after seeding (Belmont, Kendall & Nicolini, 1980). Also confirmed are reports of increased nuclear condensation occurring with cell aging (Sawicki, 1979). Using these histograms as a reference, in each slide studied (5 days after seeding, early and late phase II, and 10 days after seeding, early phase II) 2 relatively narrow ranges of area (one large and one small) were chosen and from these ranges 30-35 nuclei with 2C DNA content (as measured by I.O.D.) were selected for measurement of Barr body parameters as described above in the Materials and methods section. Through such sampling data were acquired from representative nuclei over a wide variation of nuclear condensation. Only a few cells (mainly from the small area classes), which had a wide distribution of dense nuclear regions of comparable density to the Barr body in other cells prevent- ing definite Barr body identification, were not considered. In most cells the Barr body was located at the nuclear border although in a few cases it was found near the centre of the nucleus or at the nuclear periphery but nonadjacent to the nuclear border. Interestingly, such a procedure could not be used on the log phase (2 days after seeding) early phase II slide as many nuclei with 2C DNA content were greatly relaxed and had no Barr body visible while other smaller nuclei with 2C DNA content igo M. Grattarola, A. Belmont and C. Nicolini showed several dense regions similar in density and size to the Barr body. Therefore this slide was excluded from the study. The major results are summarized in Table i, and suggest the following interpreta- tions: (i) In the condensed nuclei (class I in Table i) the Barr body as defined by eye is well matched to the Barr body defined through the percentage constraint on I.O.D.

Fig. i. Two-dimensional histograms ot l.U.L). (arbitrary units) vs area Gum") of the chromatin of human fibroblasts, for young cells 2 days after seeding (top left), 5 days after seeding (top right), 10 days after seeding (bottom left) and old cells 5 days after seeding (bottom right). Only the nuclei in the 2C DNA window are shown. Note that the z axis corresponds to the number of cells with the given I.O.D. (x-axis) and area (y-axis) values.

As this constraint was based on the percentage of the total 2C DNA found in an X chromosome, this implies that what is therefore detected by eye as the Barr body in condensed nuclei may indeed be the entire X chromosome. In relaxed nuclei (class II in Table 1), however, the I.O.D. of the Barr body identified by eye is consistently lower than the I.O.D. fixed by the 2-66% condition. It is therefore unlikely that the Barr body as detected by eye in these nuclei is comprised of the entire X chromosome and based on the I.O.D. values recorded in Table 1 in fact cannot contain more than approximately 85 % of the total. Assuming that the Barr body does in fact consist of DNA almost entirely derived from the X chromosome, this implies that in the more relaxed nuclei from class II, a significant fraction of the DNA has dispersed to a

192 M. Gratia tola, A. Belmont and C. Nicolini density level comparable to the surrounding chromatin. (2) The optical density threshold for Barr body detection, both for the by eye and by the 2-66% I.O.D. constraint identification procedures, decreases consistently with the decrease of the nuclear average optical density. Not only does this mean that the Barr body is subject to a process of relaxation coordinated with the entire nuclear relaxation but it also means

0 100 Chromatin area, jim2 Fig. 2. Correlation between the mean values of nuclear (chromatin) area (x'-axis) and Barr body area (y axis) as, respectively, detected by eye (A) and by the 2'66% constraint (#). For details see the text. that the dark ' heterochromatic' region detected by eye is itself subject to a process of relaxation. A reasonable inference, therefore, is that the overall process of Barr body relaxation does not necessarily involve a transition between 2 distinct states of condensation whereby relaxation is accomplished by an increased proportion of the chromatin in the relaxed state, but instead may involve a continuous variation in condensation. Moreover, not only does this relation between Barr body and nuclear condensation appear in a comparison of the 2 different area classes of each slide but it is also present Barr body and chromatin 193 in a comparison of the same classes from different slides. In particular it is striking to note that just as the nuclei of the old fibroblasts are on the average smaller than in the younger fibroblasts, the respective Barr bodies are smaller and more condensed in the older population. This result agrees well with previous findings on the aging phenomenon of in vitro fibroblasts (Sawicki, 1979; Yanishensky, Mendelsohn, Mayall & Cristofalo, 1974). In all cases the area of the Barr body as detected by the 2-66 % constraint on I.O.D. correlates well with the total nuclear area and there still remains a clear correlation with the 'by eye' detection. Linear regressions between the 6 means of nuclear area, averaged for each class in each slide, and the corresponding Barr body areas, give correlation coefficients of 0-97 for the 2-66% constraint detection procedure and o-86 for the by eye detection procedure. The plot of mean nuclear area versus mean Barr body area is shown in Fig. 2.

DISCUSSION In conclusion, while the state of condensation has been used until now as a means of characterizing the Barr body this paper shows that at least under our experimental conditions it is impossible to associate the Barr body with any fixed level of condensation. Moreover, the variations in Barr body condensation are not random but rather are closely linked to variations in nuclear condensation. Small, dense Barr bodies correspond to small and dense nuclei and the relationship between Barr body and nuclear condensation yields a linear relationship between Barr body and nuclear area. As a consequence, the often proposed link between condensation and genetic inactivity of the sex chromatin must be reexamined. Indeed our data support the state- ment that 'condensation, leading to sex chromatin formation, is not the cause of (genetic) inactivity, but a secondary phenomenon' (Lyon, 1968). In particular, this secondary phenomena would appear to be related to the overall modulation of chromatin structure occurring in a number of well characterized systems, including serum stimulation of diploid fibroblasts, PHA stimulation of human lymphocytes, and erythrocyte activation which has led to the concept of increasing metabolic activity (in terms of transcription or translation) being associated with increased chromatin relaxation (Comings, 1972; Nicolini, 1979, 1980). Our results show that such a phenomenon extends even to the genetically inactive chromatin of the Barr body and thus probably involves the entire genome, active or inactive (although possibly to varying degrees). Finally, all the findings described above are at least consistent with the hypothesis that the Barr body is physically attached to the nuclear membrane, so that its uncoiling could be related to the membrane expansion caused by an increase in nuclear volume associated with both nuclear and chromatin relaxation. Moreover, if what is involved with the Barr body is the extreme case of a general phenomenon, it is possible to join the speculation of many authors that physical attachment of chromatin to the nuclear membrane in interphase nuclei occurs for all chromosomes (DuPraw, 1965; Comings, 1968; Nicolini, 1979, 1980; Murray & Davies, 1979; Nicolini et al. 1977) 194 M. Grattarola, A. Belmont and C. Nicolini but is masked in these chromosomes due to their lower relative level of condensation, and that the same type of phenomenon is involved in the overall process of chromatin relaxation occurring as one of the early events of chromatin activation.

This research was supported by NIH Grant CA20034. M. Grattarola was the recipient of a North Atlantic Treaty Organization (N.A.T.O.) fellowship from the National Research Council of Italy (C.N.R.).

REFERENCES BACK, F. (1976). The variable condition of euchromatin and heterochromatin. Int. Rev. Cytol. 45. 25-64- BELMONT, A., KENDALL, F. & NICOLINI, C. (1980). Coupling of nuclear morphometry to cell geometry and growth in human fibroblasts. Cell Biophys. (In Press.) BOLUND, L., RINGERTZ, N. R. & HARRIS, H. (1969). Changes in the cytochemical properties of erythrocyte nuclei reactivated by cell fusion..7. Cell Sci. 4, 71-87. COMINGS, D. E. (1967). Sex chromatin, nuclear size, and the cell cycle. Cytogenetics 6, 120-144. COMINGS, D. E. (1968). The rationale for an ordered arrangement of chromatin in the interphase nucleus. Am .J. Hum. Genet. 20,440-460. COMINGS, D. E. (1972). The structure and function of chromatin. Adv. Hum. Genet. 3, 323-431. DUPRAW, J. (1965). Macromolecular organization of nuclei and chromosomes - a folded fibre model based on whole mount electron microscopy. Nature, Lond. 206, 338-343. KENDALL, F., SWENSON, R., BORUN, T., ROWINSKI, J. & NICOLINI, C. (1977a). Nuclear morphometry during the cell cycle. Science, N. Y. 196, 1106-1109. KENDALL, F., WU, C, GIARETTI, W. & NICOLINI, C. (19776). Multiparameter geometric and densitometric analyses of the G0-G1 transition of WI-38 cells. J. Histochem. Cytochein. 35, 724-729. KENDALL, F., BELTRAME, F. & NICOLINI, C. (1979). Nuclear morphometry in normal and SV-40 transformed human diploid fibroblasts. IEEE Trans. Biomed. Eng. BME-26, 3, I72-I73- KLINGER, H. P., DAVIEW, J., GOLDHUBER, P. & DITTA, T. (1968). Factors influencing mammalian X chromosome condensation and sex chromatin formation. Cytogenetics 7, 39-57. LINDEN, W., FANG, S., ZIETZ, S. & NICOLINI, C. (1979). Differential effects of feulgen hydrolysis. In Chromatin Structure and Function (ed. C. Nicolini), pp. 323-340. New York and Lon- don: Plenum Press. LYON, M. F. (1968). Chromosomal and subchromosomal inactivation. A. Rev. Genet. 2, 31-52. MACLEAN, N. & HILDKR, V. A. (1977). Mechanisms of chromosome activation and repression. Int. Rev. Cytol. 47, 1-54. MENDELSOHN, M. G. (1979). Scanning and flow photometry of chromosomes. In Chromatin Structure and Function (ed. C. Nicolini), pp. 341-356. New York and London: Plenum Press. MlTTWOCH, U. (1967). Barr bodies in relation to DNA values and nuclear size in culture human cells. Cytogenetics 6, 38-50. MOKHERJEE, A. B., MOSER, G. C. & WITOWSKY, H. M. (1972). Fluorescence of X and Y chromatin in interphase cells. Cytogenetics 11, 216-227. MOKHERJEE, A. B. & SAN SEBASTIAN, J. R. (1978). Differential frequency of human X and Y chromatin as related to cell density in vitro. Cytogenet. Cell Genet. 21, 139-144. MURRAY, A. B. & DAVTJSS, H. G. (1979). Three dimensional reconstruction of the chromatin bodies in the nuclei of mature erythrocytes from the newt Triturus cristatus: the number of nuclear envelope-attachment sites. J. Cell Sci. 35, 59-66. NICOLINI, C, KENDALL, F., DE SAIVE, C. & GIARETTI, W. (1977). The G0-G1 transition of WI-38 cells: geometric and densitometric texture analyses. Expl Cell Res. 106, 119-125. NICOLINI, C. (1979). Chromatin structure, from angstrom to micron level, and its relationship to mammalian cell proliferation. In Chromatin Structure and Function (ed. C. Nicolini), pp. 613-666. New York and London: Plenum Press. Barr body and chromatin 195 NlCOLINl, C. (1980). Nuclear morphometry, quinternary chromatin structure and cell growth. Jl submicrosc. Cytol. 12, 475-505. SAWICKI, W. (1979). Chromatin pattern in situ: dependence upon cell cycle, preimplantation and development, and cellular aging in vitro. In Chromatin Structure and Function (ed. C. Nicolini), pp. 667-682. New York and London: Plenum Press. SCHWARZACHER, H. G. (1963). Sex chromatin in living human cells in vitro. Cytogenetics 2, 117-128. VONDERHEID, E. C, FANG, S., HELFRICH, M. B., ABRAHAM, S. & NICOLINI, C. (1980). Bio- physical characterization of sezary cells. J. invest. Derm. (In Press.) YANISHENSKY, R., MENDELSOHN, M., MAYALL, B. & CRISTOFALO, V. (1974). Proliferative capacity and DNA content of aging human diploid cells in culture: a cytophotometric and autoradiographic analysis.,7. cell. Physiol. 84, 165-170.

(Received 31 January 1980 - Revised 22 April 1980)