Cytometry 8:461-467 (1987)

Visualization of DNA Loops in Nucleoids From HeLa Cells:

Assays for DNA Damage and Repair
J.L. Roti Roti and W.D. Wright
Section of Cancer Biology, Radiation Oncology Center, MIR, School of Medicine, Washington University,
St. Louis, Missouri 63108
Received for publication February 24, 1987; accepted May 8, 1987

An assay for visualization of DNA loops
undergoing supercoiling changes has been de-
veloped. The assay utilizes the fluorescent
dye, propidium iodide (PI),which intercalates
into the DNA and under the proper conditions
causes the supercoiling status of the DNA to
change. Thus, the DNA can be seen as a flu-
orescent halo that changes diameter with P I
concentration. At low PI concentrations (0-
7.5 pglml) the supercoils are relaxed with in-
creasing PI, while at higher PI concentra-
tions (7.50-50 pglml) supercoils in the opposite
winding sense are rewound with increasing
PI. When HeLa cells were irradiated with 1-
20 Gy of lS7Csy’rays, the ability to rewind the
DNA supercoils was inhibited in a dose-depen-
dent manner, presumably because of’ the
presence of radiation-induced DNA strand
breakage, which removed the topological con-
straints on the DNA loops. These lesions were
repaired rapidly during post-irradiation incu-
bation. The ability of the DNA loops to be
rewound was restored within 8 min after 10
Gy of y-irradiation, such that no difference
from control cells could be detected. The half-
time for repair of the radiation-induced le-
sions that inhibit DNA rewinding was similar
to that for repair of DNA single strand breaks.
The assay has certain advantages over cur-
rent methods for assaying DNA damage in
that it involves measurement of single cells
and it does not require the DNA to be labeled
with radioactive precursors.
Key terms: DNA supercoiling, propidium
iodide

Studies of DNA organization within the nucleus have (15). In this case the total particle appears to be the
shown that DNA-nuclear matrix attachment points de- DNA loops plus the residual nuclear matrix. Therefore,
fine domains of DNA for supercoiling changes (13). The we developed a method to measure DNA supercoiling
nuclear matrix-DNA attachment points are believed to changes by DNA loop extension directly. The assay,
be the sites of DNA replication (€9, RNA transcription which allows visualization of the DNA loops as a flu-
(2) and processing (51, and possibly DNA repair (12). orescent halo extruding from the nuclear matrix, in-
Thus, considerable interest has been directed toward volves the exposure of salt extracted (1-2 M) nuclei to
these anchor points and the DNA loops defined by them. varying concentrations of the intercalating dye, PI. The
One characteristic of these DNA loops is the manner in diameter of the fluorescent DNA “halo” is measured as
which they undergo supercoiling changes (16). One ap- a function of dye concentration using a n inverted fluo-
proach to the study of DNA supercoiling changes is to rescence microscope. This method is a n extension of that
use titration with intercalating dyes, e.g., ethidium bro- developed by Vogelstein and coworkers (17). In the pres-
mide or propidium iodide (PI), along with sedimentation ent study we have applied the method to study the
in neutral sucrose gradients (3,4). The distance sedi- effects of radiation-induced DNA damage and its repair
mented in the gradient is a measure of the degree of on the ability of DNA to undergo supercoiling changes.
supercoiling. That is, as the DNA supercoils relax and
the DNA loops become extended, nucleoid sedimenta- MATERIALS AND METHODS
tion becomes retarded conversely, when the loops be- Cell Maintenance and Irradiation
come more tightly supercoiled, sedimentation is HeLa cells were maintained in exponential growth in
increased. The sedimentation method has been used to suspension culture in spinner flasks (Bellco, Vineland,
investigate the role of supercoiling changes in DNA NJ) by daily subculture with Joklik MEM (GIBCO,
replication (9-11) and the effects of ionizing radiation (4)
and of hyperthermia (15) on DNA supercoiling.
~~~

Address reprint requests to Joseph L. Roti Roti, Ph.D., Washington

Nucleoid sedimentation depends not only on the DNA University School of Medicine, Section of Cancer Biology, 4511 Forest
loop extension but also on the mass of the total particle Park Blvd., St. Louis, MO 63108.
0 1987 Alan R. Liss, Inc.
462 ROTI ROTI AND WRIGHT

tor (Shepard Co., Mark I, Glendale, CA) was used to

FIG.1. Representative nucleoids showing core and halo structure.
This figure is a black and white print made from a color slide of typical
nucleoids. The magnification was ~ 4 0 for0 better visualization on the
print. For routine data collection photomicrographs were made at 200
power to allow more measurements to be made per field. Two perpen-
dicular diameters of the overall halo and central core were measured
and averaged for each nucleoid that was judged to be in good focus by
the ability to visualize a sharp edge on the nuclear lamina (the central
bright circle). The alpha numeric characters on the print are produced
by the camera’s “data back” to allow a coded identification of the slide
so that double-blind measurements of the halo and core diameters can
be made.
Grand Island, NY) supplemented with 3.5%each of calf
and fetal calf serum (GIBCO). Prior to irradiation, cells
were washed in Spinner’s salts (0.4 g KC1; 0.1 g MgS04;
6.8 g NaC1; 2.2 g NaHC03; 1.4 g NaHZP04; and 1.0 g
glucose per liter at pH 7.4) and resuspended in Spinner’s
salts at 0°C in a 50-ml centrifuge tube. A 137Csirradia-
deliver the radiation dose rate, 17 Gylmin. The unit for
dose used in this study was the Gy (International Sys-
tem of Units; see Radiat. Res. 103:471-472, 1985), which
is equivalent to 100 rads. Following irradiation the cell
samples were either assayed immediately or resus-
pended in medium, prewarmed in 37”C, for post-irradia-
tion incubation.
Fluorescent Halo Assay
To visualize the DNA loops, cells were resuspended to
a concentration of 1-2 x 105/mlin Spinner’s salts. Three
hundred microliters of this suspension were placed in
each well of a four-well LabTek slide (Miles Scientific,
Naperville, IL) that had been pretreated with poly-l-
lysine (Sigma, St. Louis) by coating the slide and pour-
ing off the excess solution (1.0 pg/ml, 47,000 number
average molecular weight a t pH 7.4). Equal volumes of
the dye-lysis solution (2.0 M NaC1, 10 mM ethylenedi-
amine tetraacetic acid [EDTA] [disodium salt], 2 mM
Tris [pH 8.01, 0.5% Triton X-100, with twice the desired
PI concentration) and cell suspension were added t o each
well and gently mixed. Following 10 min of lysis in the
dark at room temperature, the nucleoids were visualized
with a n inverted fluorescence microscope (Olympus Op-
tical Co., Tokyo) under excitation with green light (520-
570 nm, 546 nm peak transmission). Nucleoids were
observed and photographed through a 610-nm long-pass
filter. The objective was a x20 Achromat, numerical
aperture, 0.40, with a 5.44-mm working distance. Fluo-
rescence photomicroscope were taken of fields selected
for uniformity of focus and number of specimens with
a n Olympus OM-2 camera body containing Kodak Ek-
tachrome ASA 400 daylight film (Eastman Kodak Co.,
Rochester, NY). Photomicrographs were developed,
coded, and randomized prior to measurement of halo
diameters. An internal size standard was included with
each roll of film.

HALO

c I / I

CORE

U-LL-U-1
0 10 20 30 40 50
PI CONCENTRATION
FIG.2. The effects of increasing PI concentration on the DNA loop diameter in
HeLa cell nucleoids. HeLa cells were lysed in the presence of the indicated PI
concentration (abscissa)as described in the text. The diameter of the overall nucleoid
and the nuclear lamina (core) when visible (2-20 pg/ml PI) are plotted (ordinate).
These data are from a single typical experiment and the error bars represent 1 S.D.
 VISUALIZATION OF DNA LOOPS 463
UNSORTED CELLS unit DNA causing the unwinding of the endogenous left-
handed superhelical domains. This process results in a n
increase in nucleoid halo diameter as the unwinding
DNA supercoils extrude from the nucleus. The halo di-
ameter continues to increase until the DNA loops are
fully extended (called the relaxation point). 2) The re-
winding phase begins at the relaxation point where
additional intercalation of PI causes the DNA helix to
form righthanded superhelical domains. This process
results in a decrease in the halo diameter as the rewind-

[/u,l,i
ing DNA loops pull back into the nuclear matrix. Re-
winding is dependent on the topological constraint of

i
20pq/m1

the DNA loop. This characteristic is the basis for use of

the method to assess DNA damage.
Covalently closed circular loops of DNA are known to
unwind and rewind in a symmetrical manner with re-
spect to PI concentration (18).Conversely, a loop contain-
ing a single-strand break will not rewind with increas-
5
ing PI concentration. The rewinding response illustrated
in Figure 2 is clearly intermediate between these two
0 situations, showing that the DNA nuclear matrix an-
chor points are not covalent nor are they unconstrained.
A reasonable expectation is that radiation-induced DNA
damage would change this response pattern.
Prior to testing the utility of the assay for measure-
ment of DNA damage, consideration has to be given to
the large uncertainty (10-20% of the mean value) in the
data points of Figure 2, which is typical of single exper-
iments. There are two reasonable possibilities for the
source of this uncertainty. 1)The uncertainty could be
due to random error in the assay. If this were true, then
the assay would lack sufficient sensitivity to be able to
t detect changes caused by low doses of radiation. 2 ) Since
the cell sample was obtained from exponentially grow-
ing cells, the uncertainty could arise from the size distri-
bution of the nucleoids, which were derived from cells
throughout the cell cycle. In this case, the problem could
10
be easily reduced either by averaging the results of
S repeated experiments and using the standard error of
the mean (S.E.M.) for determining significance or by
LL
' 5 20 35 50
HALO DIAMETER (p1
65 using sorted cells. That the second possibility is correct
is indicated by the distributions of halo diameter (Fig.
3) which are skewed toward larger- diameters consistent
with that expected for a cell-cycle-dependent size distri-
FIG. 3. Nucleoid halo diameter distributions as a function of PI bution. This conc~usionwas confirmed by twoadditional
concentration. Distributions of nucleoid diameters at the indicated PI
concentrations are plotted for experiments similar to that shown in
observations. 1)When cells were sorted from the G1 or
Figure 2. Four such experiments are pooled so that between 50 and G2 peak Of the DNA histogram Prior to n u c h i d PrePa-
130 measurements are included in each histogram; the histograms ration, the distribution of halo diameters (Fig. 4)became
have been normalized to 50 measurements. more nearly gaussian in shape and narrower (e.g., 86%
of the G1, 50 pg/ml PI sample have diameters within a
1.5-pm range from the mode, whereas in the unsorted
RESULTS sample 48% do within a similar range). 2) The S.E.M.
The appearance of representative nucleoids is shown values for repeated experiments tend to be small. For
in Figure 1. Characteristically, measurements of nu- example, the control curve presented in Figure 5, which
cleoids yield a curve (Fig. 2) that can be divided into two is the mean of five experiments, had S.E.M. values of
distinct phases. 1)The unwinding phase begins at low 1.5, 1.8, 1.5, 1.0, 1.0, 0.6, 0.7, and 0.6 pm, respectively,
PI concentration ( 2 0 . 5 pgiml) where the nucleoid diam- for the eight data points in the order of increasing PI
eter is = 16 pm, which is somewhat larger than the size concentration. These values are much smaller than the
of a n intact nucleus (11.6 0.6 pm) (1).As the PI concen- error values presented in Figure 2. Thus, repeated
tration increases, more dye molecules intercalate per experiments are needed to reduce the error that is
464 ROTI ROTI AND WRIGHT
SORTED CELLS
30 r
G,75pg/mI

25 1
t

I
2C&Y/Tl

;
I
20 35 50 65 35 50 65

FIG.4. Distribution of halo sizes from sorted GI, S, and Gz cells. Nucleoid halo size
distributions were obtained from HeLa cells sorted on the basis of DNA content
following lysis in the indicated PI concentrations. Cells were stained with H033342
prior to cell sorting. Staining with H033342 did not alter halo diameter (data not
shown). Data from two independent sorts are pooled, with the number of measure-
ments ranging from 50-90 per histogram. All histograms are normalized to 50. Note
the change in ordinate scale for the 50 pgiml histograms.

due to the cell-cycle-dependent distribution of nucleoid
diameters.
Ionizing radiation is known to cause a wide variety of
DNA damage (141, which should change the topological
constraints on the DNA loops. When cells were exposed
to 10 Gy of lS7Cs y-rays prior to nucleoid preparation,
the rewinding response was greatly reduced (Fig. 51,
suggesting that radiation-induced DNA damage (pre-
sumably strand breaks) removes topological constraints
that are needed for rewinding of the DNA supercoils.
The loss of the ability of the DNA supercoils to rewind
was dependent upon the radiation dose up to a t least 20
Gy (Fig. 5) At doses above 10 Gy the maximum halo
diameter increased above that for control. The dose de-
pendency of the radiation effects on the fluorescent halo
showed that the assay can be used to detect radiation
damage to DNA and that it is adequately sensitive to
detect damage resulting from a dose of more than 1Gy.
After 1Gy of radiation the difference between irradiated
and control was the limit of resolution (i.e., the statisti-
cal significance was marginal). However, the 2-Gy data
was significantly different from control (at and above 10
pg/ml PI) to the 95% confidence level by Student’s t-test.
Cells are known to have the ability to repair many
types of DNA damage. To determine if the fluorescent
halo assay would reflect the repair of DNA damage, cells
were exposed to either 10 or 20 Gy of y-rays and incu-
bated a t 37°C for time intervals ranging from 7.5 min
to 240 min. Following irradiation with 10 Gy, the ability
of the DNA to rewind was completely restored after 7.5
min post-irradiation incubation (data not shown). At 10
min following the 20-Gy dose the ability of the DNA to
 VISUALIZATION OF DNA LOOPS 465
Dose (GY)

I I I I 1
0 10 20 30 40 50
PI CONCENTRATION (pq/mI)

FIG.5. The effects of y-irradiation on the DNA supercoiling changes in HeLa

nucleoids. HeLa cells were irradiated with various doses of 137Csy-rays (indicated
on the Figure), and the DNA available for supercoiling changes was assayed by the
fluorescent halo technique. The overall halo diameter was measured and plotted
against PI concentration. Each data point represents the mean of at least three
experiments (except for the 2-Gy data, which is from one experiment). For clarity,
errors bars are omitted (see text).

rewind was approximately equivalent to that obtained greater. We can then define the "excess halo diameter"
immediately after a 5-Gy dose (Fig. 6). However, even above control at PI concentrations of 10-50 pgiml as a
after 40 min of repair, the rewinding had not completely measure of DNA damage, which can be plotted as a
returned to control levels. This result implies that the function of radiation dose andor repair time (Fig. 7).
repair of the DNA damage responsible for inhibiting This method of plotting allows the expression of the
DNA supercoiling is multiphasic. DNA damage remaining after repair as Gy equivalent
The unwinding-rewinding curves (Figs. 5, 6) are diffi- damage, independent of biochemical identification of the
cult to quantify in terms of the dose-effect relationship. damage. For example, following the 20-Gy dose, we can
Therefore, we developed the following as a means to see from this figure that 4 Gy equivalents remain after
quantify radiation-induced damage repair in terms of 10 min of repair while about 2 Gy remain after 40 min
Gy equivalents remaining. The DNA rewinding process of repair. The multiphasic nature of repair is clearly
is clearly operating a t PI concentrations of 10 pgiml and evident in this plot.

50 r
20 Gy

2% - 10'
y
20Gy+20'
20 G y f 4 0 '

"c
0

= 10

0 I J
0 10 20 30 40 50
P I C9NCENTRATION (,uq/ml)

FIG.6 . The effects of y-irradiation and repair on the DNA supercoiling changes in
HeLa cell nucleoids. HeLa cells were irradiated with 20 Gy of 137Cs y-rays and
incubated at 37°C for 10,20, and 40 rnin after irradiation. The effect of 5 Gy without
repair is included for reference. After the appropriate repair interval the DNA
available for supercoiling changes was assayed as described. The mean halo diame-
ters from three experiments are plotted against PI concentration. Error bars and
nuclear lamina diameters are omitted for clarity.
466 ROTI ROTI AND WRIGHT
0 5 10 15 20
DOSE (Gy)
FIG.7. Radiation-induced DNA damage and its repair as measured
by the ability of DNA loops to rewind. The left panel shows the
radiation-induced loss in the ability of DNA loops to rewind as a
function of radiation dose. The excess halo diameter at 10+ pgiml PI
(ordinate) is the sum of halo diameters at 10, 20, 35, and 50 pgiml PI
for the nucleoids from irradiated cells minus the same sum for nu-
cleoids from control cells. The plotted points represent the mean of
these differences from at least three experiments (except for the 2-Gy
points, which is from one experiment). The error bars are f 1 S.D. The
DISCUSSION
Two considerations arise from the results presented in
this paper. First, the data show the utility of the fluores-
cence halo method in detecting and measuring the pro-
duction and repair of radiation-induced DNA damage.
Presumably, single-strand DNA breaks cause the inabil-
ity of the DNA to rewind, while double-strand breaks
may cause the maximum halo diameter to increase at
doses above 10 Gy. Second, these studies address the
repair of radiation damage which affects the ability of
DNA to supercoil and show that restoration of this abil-
ity occurs rapidly in HeLa cells. More than 80% of the
ability to rewind DNA is recovered in 10 min after 20
Gy (Fig. 7); in contrast, this process in rat cerebellar
neurons requires -8 h. (6). In HeLa cells the rate of
recovery of the ability to rewind DNA appears to be
similar to that of single-strand break rejoining both in
view of the mukiphasic characteristic of the kinetics
and the similarity of the initial rates. The initial TI,* is
=4.5 min for the first portion of the rewinding compo-
nent, which is similar to that reported for single-strand
break rejoining (7). The slower component(s) could be
due to the rejoining of double-strand breaks. The data
available suggest that restoration of the ability of the
DNA to supercoil is dependent upon the rejoining of
single- and double-strand breaks. However, a n extensive
study would be required to establish if, in fact, the loss
of ability of DNA supercoils to rewind is due to the
nresence of single- and double-strand DNA breaks.
0 10 20 30 40 120 240
REPAIR TIME (min)
dose-response function was represented as two straight lines fit visu-
ally to the data points, which is the simplest model consistent with the
data. The right panel shows the restoration of the ability of DNA loops
to rewind as a function of post-irradiation repair time. HeLa cells were
irradiated with either 10 Gy (squares) or 20 Gy (circles) of lR7Csy-rays
and incubated at 37°C to culture medium for the time interval indi-
cated (abscissa). The ordinate is the same as for the left panel. Each
plotted point represents the mean of three separate experiments. Error
bars have been omitted for clarity.
The fluorescence halo method has two advantages over
the sucrose gradient and alkaline elution methods: 1)
the potential to make measurements of individual cells
which allows one to detect undamaged cells in the pres-
ence of damaged cells; 2) the ability to obtain data from
a modest number of cells ( - lo4 are used per LabTek
well) that are not radioactively labeled allows the
method to be used for in vivo systems, such as cerebellar
neurons (6). Another aspect of our assay that represents
an improvement over previously published halo assays
(16) is that cells are not required to be grown on cover-
slips for the assay. Furthermore, the method can detect
damage at radiation doses on the order of 2 Gy. It is
possible that with a n automated analysis method the
techniques should be able to detect damage a t 1 Gy or
less. Thus, the method provides a convenient assay hav-
ing several advantages over current methods.
ACKNOWLEDGMENTS
T h i s work w a s supported by grant CA-41102 from t h e Na-
tional Cancer Institute of the National Institutes of Health.
The a u t h o r s would like t o thank Ms. K. McDonald for help in
the preparation of this m a n u s c r i p t and Dr. L.J. Tolmach for
critical comments.
LITERATURE CITED
1. Blair OC, Winward RT, Roti Roti JL:The effect of hyperthermia
on the protein content of HeLa cell nuclei: A flow cytometric
analysis. Radiat Res 78:474-484, 1979.
 VISUALIZATION OF DNA LOOPS
2. Cijek EM, Tsai M-J, O’Malley, BW: Actively transcribed genes are
associated with the nuclear matrix. Nature 306607-609, 1983.
3. Cook, PR, Brazell IA: Supercoils in human DNA. J Cell Sci 19:261-
279, 1975.
4. Cook PR, Brazell IA: Conformational constraints in nuclear DNA
J Cell Sci 22287-302, 1976.
5. Gallinaro H, Puvion E, Kister L, Jacob M: Nuclear matrix and
hnRNP share a common structural constituent associated with
premessenger RNA. EMBO J 2:953-960, 1963.
6. Jaberaboansari A, Nelson GB, Roti Roti JL, Wheeler KG: Post
irradiation alterations in neuronal chromatin structure. Radiat
Res, 1987 (submitted for publication).
7. Jorritsma JBM, Konings AWT: The occurrence of DNA strand
breaks after hyperthermic treatments of mammalian cells with
and without radiation. Radiat Res 98:198-208, 1984.
8. McCready SJ, Godwin J, Mason DW, Brazell IA, Cook PR: DNA is
replicated at the nuclear cage. J Cell Sci 46:365-380, 1980.
9. Mattern MR, Painter RB: Dependence of mammalian DNA repli-
cation on DNA supercoiling. I. Effects of ethidium bromide on
DNA synthesis in permeable Chinese hamster ovary cells. Biochim
Biophys Acta 563: 273-305,1979a.
10. Mattern MR, Painter RB: Dependence of mammalian DNA repli-
cation on DNA supercoiling. 11. Effects of novobiocin on DNA
synthesis in Chinese hamster ovary cells. Biochim Biophys Acta
563:306-312, 197913.
467
11. Mattern MR, Scudiero DA: Dependence of mammalian DNA syn-
thesis on DNA supercoiling. 111. Characterization of the inhibition
of replicative and repair-type DNA synthesis by Novobiocin and
nalidixic acid. Biochim Biophys Acta 653:248-258, 1981.
12. Mullenders LHF, van Kesteren AC, Bussman CJM, van Zeeland
AA, Natarajan A T Preferential repair of nuclear matrix associ-
ated DNA in xeroderma pigmentosum complementation group C.
Mutat Res 141:75-82, 1984.
13. Nelkin B, Pardoll D, Robinson S, Small D, Vogelstein B: Nuclear
structure and DNA organization. In: Tumor Cell Heterogeneity:
Origins and Applications, Owens AH, Coffey DS, Baylin SB, (eds).
Academic Press, New York, 1982, pp 441-457.
14. Okada S: Radiation Biochemistry, Academic Press, New York,
1970, pp 109-120.
15. Roti Roti JL,,Painter RB: Effects of hyperthermia on the sedimen-
tation of nucleoids from heLa cells in sucrose gradients. Radiat
Res 89:166-175, 1982.
16. Vinograd J, Lebowitz J, Radloff R, Watson R, Laipis P The twisted
circular form of polyoma viral DNA. Proc Natl Acad Sci USA
53:1104-1111, 1965.
17. Vogelstein B, Pardoll DM, Coffey DS: Supercoiled loops and eucar-
yotic DNA replication. Cell 22:79-85, 1980.
18. Wang JC: Variation of the average rotation angle of the DNA
helix and the superhelical turns of convalently closed cyclic DNA.
J Mol Biol43:25-39, 1969.