DNA double-strand breaks induced by high NaCl occur
predominantly in gene deserts
Natalia I. Dmitrievaa,1, Kairong Cuib, Daniil A. Kitchaevc, Keji Zhaob, and Maurice B. Burga,1
aSystems Biology Center andbLaboratory of Molecular Immunology, National Heart, Lung, and Blood Institute, National Institutes of Health, Bethesda,
MD 20892; andcDepartment of Chemical Engineering, California Institute of Technology, Pasadena, CA 91125
Contributed by Maurice B. Burg, October 14, 2011 (sent for review November 5, 2010)
High concentration of NaCl increases DNA breaks both in cell
culture and in vivo. The breaks remain elevated as long as NaCl
concentration remains high and are rapidly repaired when the
concentration is lowered. The exact nature of the breaks, and their
location, has not been entirely clear, and it has not been evident
how cells survive, replicate, and maintain genome integrity in
environments like the renal inner medulla in which cells are
constantly exposed to high NaCl concentration. Repair of the
breaks after NaCl is reduced is accompanied by formation of foci
containing phosphorylated H2AX (γH2AX), which occurs around
DNA double-strand breaks and contributes to their repair. Here,
we confirm by specific comet assay and pulsed-field electrophore-
sis that cells adapted to high NaCl have increased levels of double-
strand breaks. Importantly, γH2AX foci that occur during repair of
the breaks are nonrandomly distributed in the mouse genome. By
chromatin immunoprecipitation using anti-γH2AX antibody, fol-
lowed by massive parallel sequencing (ChIP-Seq), we find that
during repair of double-strand breaks induced by high NaCl,
γH2AX is predominantly localized to regions of the genome de-
void of genes (“gene deserts”), indicating that the high NaCl-in-
duced double-strand breaks are located there. Localization to gene
deserts helps explain why the DNA breaks are less harmful than
are the random breaks induced by genotoxic agents such as UV
radiation, ionizing radiation, and oxidants. We propose that the
universal presence of NaCl around animal cells has directly influ-
enced the evolution of the structure of their genomes.
hypertonicity|salt|DNA damage|kidney|mIMCD3 cells
medullary cells in vivo (1), cells of the soil nematode Caeno-
rhabditis elegans (3), and marine invertebrates (4). Acute eleva-
tion of NaCl in cell culture increases the number of DNA breaks
(2, 5) and transiently arrests cells in all phases of the cell cycle (6,
7). After several hours, the cells begin proliferating again, despite
the continued presence of high NaCl (7). However, even after
cells adapt to high NaCl and reenter the cell cycle, numerous
DNA breaks persist (1). Excessive elevation of NaCl causes ap-
optosis (7). However, the increased DNA breaks that occur at
levels of NaCl that cells survive and to which they adapt differs
from the chromatin fragmentation that occurs during apoptotic
cell death. Thus, high NaCl increases DNA breaks in viable cells
without the activation of caspases, nuclear condensation, or for-
mation of apoptotic bodies characteristic of apoptosis (8, 9).
The increase of DNA breaks caused by high NaCl is not
limited to proliferating cells in culture. High NaCl also induces
DNA breaks in normal cells in animal tissues in vivo. Thus,
numerous DNA breaks are normally present in the mouse renal
inner medulla (1), where high interstitial NaCl provides the
driving force for concentration of the urine (10). The excess
breaks in the inner medulla disappear quickly when the high
intercellular NaCl concentration in the renal medulla is lowered
by the diuretic furosemide (1). The soil nematode, C. elegans is
able to adapt to and live in a high NaCl environment (11), and
adaptation of C. elegans to high NaCl is accompanied by increased
igh extracellular NaCl increases the number of DNA breaks
in mammalian cells in tissue culture (1, 2), mouse renal inner
DNA breaks (3). Finally, according to some estimates, ≈80% of
all Earth’s life lives in the ocean, which has a high osmolality of
≈1,000 mosmol/kg, the dominant solute being NaCl. Many ma-
rine invertebrates are osmoconformers, i.e., the NaCl in their
extracellular fluids is as high as in seawater (12). Cells in tissues of
osmoconforming marine invertebrates have many DNA breaks
that disappear if the seawater in which they are immersed is grad-
cells exposed to high NaCl is an evolutionarily conserved phe-
nomenon. However, the nature of the DNA breaks, their location,
and the mechanism of their induction has not been entirely clear.
A striking feature of adaptation to high NaCl is that despite
increased DNA breaks, the cells do not activate the DNA dam-
age response (1, 5, 8). However, the DNA damage response is
activated quickly when NaCl is lowered. Thus, reducing NaCl to
total osmolality of 300 mosmol/kg (the level normally maintained
in mammalian blood and body fluids by osmoregulatory mecha-
nisms) results in rapid repair of the DNA breaks (1, 5). This
repair is accompanied by rapid phosphorylation of histone H2AX
(called formation of γH2AX) (1, 5), the histone modification that
normally accompanies repair of double-strand breaks (13, 14).
In the present studies, we find that the high NaCl-induced
double-strand DNA breaks are not randomly distributed in the
mouse genome, but are predominantly located in gene deserts,
which are regions of the genome devoid of genes. Our findings
are summarized on Fig. S1.
High NaCl Induces Double-Strand Breaks That Are Rapidly Repaired
When the NaCl Is Lowered. We added 100 mM NaCl (which ele-
vates the osmolality to ≈500 mosmol/kg) for 22 h to the medium
bathing mIMCD3 cells. This addition of NaCl causes immediate
G2/M arrest that lasts ≈6 h (6). Then, the cells begin prolifer-
ating again. By 22 h, the cell cycle distribution, appearance of the
cells, and their rate of proliferation return to the condition be-
fore salt was elevated (ref. 1 and Fig. S2).
It has remained an open question what kind of DNA breaks
are present in cells exposed to high NaCl. The breaks induced
immediately by acute elevation of high NaCl for 1 h were orig-
inally characterized as double-strand DNA breaks (DSBs) by
pulsed field gel electrophoresis (PFGE), which detects DSBs but
not single-strand breaks (SSBs) (2). However, the methods used
to study DNA breaks after cells adapt to high NaCl and resume
proliferation were not specific to DSBs. Those methods included
Author contributions: N.I.D., K.Z., and M.B.B. designed research; N.I.D. and K.C. per-
formed research; N.I.D., D.A.K., and M.B.B. analyzed data; and N.I.D. and M.B.B. wrote
The authors declare no conflict of interest.
Data deposition: The data reported in this paper have been deposited in the Gene Ex-
pression Omnibus (GEO) database, www.ncbi.nlm.nih.gov/geo (accession no. GSE32882).
See Commentary on page 20281.
1To whom correspondence may be addressed. E-mail: email@example.com or dmitrien@
This article contains supporting information online at www.pnas.org/lookup/suppl/doi:10.
| December 20, 2011
| vol. 108
| no. 51www.pnas.org/cgi/doi/10.1073/pnas.1114677108
a version of the neutral comet assay originally developed by
Ostling and Johanson (15), alkaline comet assay (1), and in vitro
labeling of their 3′-OH ends with biotinylated deoxynucleotides
in a reaction catalyzed by terminal deoxynucleotidyl transferase
(1, 3, 4). All these methods do not distinguish between DSBs and
SSBs (16–18). In the present studies, to clarify whether DNA
breaks that remain induced by high NaCl in adapted cells include
DSBs, we used a modification of the neutral comet assay that is
specific for DSBs as opposed to SSBs (18, 19).
To confirm specificity of the neutral comet assay (19) that we
used here for DSBs, we exposed cells to hydrogen peroxide
(H2O2), known to induce only single-strand breaks (20). DNA
breaks after H2O2are not detected by the modified neutral comet
assay but are detected by the alkaline comet assay, consistent with
their single-stranded nature (Fig. 1A). We also tested mIMCD3
cells exposed to the topoisomerase inhibitor Etoposide, known to
induce both double-strand and single-strand breaks in a pro-
portion similar to ionizing radiation (20). The DNA breaks in-
alkaline comet assay, consistent with presence of both double-
and single-strand breaks (Fig. 1A). The breaks induced by high
NaCl are detected by the modified neutral comet assay (Fig. 1A),
indicating that they are double-stranded. The DSBs induced
by high NaCl are repaired within several hours after the NaCl
concentration is reduced to a total osmolality of 300 mosmol/kg
(Fig. 1 B and C). Thus, an increased level of DSBs persists after
cells have adapted to high NaCl and appear otherwise normal. In
addition, there may also be more SSBs, but the effect of high
NaCl on SSBs has not been specifically tested.
in mIMCD3 cells that are repaired within several hours after lowering NaCl
to control level (300 mosmol/kg). (A) Verification of specificity for DSBs of
the Neutral Comet method used to detect DNA breaks. DNA breaks were
induced by H2O2and Etoposide. NaCl was elevated for 22 h. H2O2induces
only SSBs and oxidative damage to nucleotides. Etoposide induces both DSBs
and SSBs. The % DNA in Comet Tails was used as a measure of the DNA
breaks. Data are plotted as mean ± SEM (n = 5–7; *P < 0.05, t test relative to
control). Neutral Comet Assay detects only the DSBs induced by Etoposide
and NaCl, but not the SSBs induced by H2O2. Alkaline Comet Assay detects
DNA breaks in cells treated with Etoposide, NaCl, and H2O2. Conclusion: high
NaCl induces DSBs. (B and C) DSBs that are induced by high NaCl are repaired
after NaCl is lowered. Cells were treated with medium in which NaCl was
elevated to total osmolality 500 mosmol/kg for 22 h, and then the high NaCl
medium was replaced with medium at 300 mosmol/kg. DSBs were measured
by the neutral comet assay. (B) Representative distributions of % DNA in
comet tails. (C) Analysis of mean % DNA in comet tails (mean ± SEM, n = 3).
*P < 0.05, t test relative to control.
High NaCl (to 500 mosmol/kg) induces double-strand breaks (DSBs)
is accompanied by the caffeine-sensitive chk1 phosphorylation, G2/M cell
cycle arrest and γH2AX induction. NaCl bathing mIMCD3 cells was elevated
by adding NaCl to total osmolality 500 mosmol/kg for 22 h, then the addi-
tional NaCl was eliminated, lowering the osmolality to 300 mosmol/kg in the
presence or absence of 2 mM caffeine. At indicated time points after low-
ering NaCl phosphorylated Chk1 (“P-Chk1”), γH2AX and the number of cells
in mitosis were determined. (A and B) Chk1 becomes phosphorylated and G2/
M arrest is activated when NaCl is lowered. Caffeine decreases the chk1
phosphorylation and abrogates the cell cycle arrest. (A) Western blot analysis
of P-Chk1. (A Upper) Representative Western blot. (A Lower) Densitometry
(mean ± SEM, n = 3, *P < 0.05 relative to 0 time;#P < 0.05 relative to caffeine,
t test). (B) Analysis of % cells in mitosis by staining with anti–P-histone H3
(mean ± SEM, n = 4; *P < 0.02, t test relative to 0 time). Dashed line shows
percent in mitosis of cells in control medium before NaCl is elevated (4.8% ±
0.2, n = 16). (C and D) Western blot analysis of γH2AX changes in response to
high NaCl, bleomycin, and UV irradiation. (C) γH2AX increases after NaCl is
lowered, and this increase is prevented by caffeine. γH2AX is maximal 15 min
after NaCl is reduced and gradually decreases as DNA breaks decrease (Fig. 1).
(C Upper) Representative Western blot. (C Lower) Densitometry (mean ±
SEM, n = 3;#P < 0.05, t test relative to caffeine). (D) mIMCD3 cells were
exposed to bleomycin (5 μg/mL for 30 min) or 15 J/m2of UV light as de-
scribed in Materials and Methods. The increase of γH2AX 15 min after re-
turn from high NaCl is comparable to that after exposure to UV radiation
Repair of high NaCl-induced DNA breaks following lowering of NaCl
Dmitrieva et al. PNAS
| December 20, 2011
| vol. 108
| no. 51
ATM/ATR-Dependent DNA Double-Strand Break Repair Response Is
Activated During Repair of High NaCl-Induced DNA Breaks. Common
DNA damage repair responses include transient cell cycle arrest,
during which the DNA repair occurs. We tested to see whether
that happens during the disappearance of the DNA breaks after
reduction of high NaCl (Fig. 2). Indeed, phosphorylation of
checkpoint kinase 1 (Chk1), which contributes to all defined cell
cycle checkpoints (21), occurs rapidly, and the phosphorylation is
reduced by caffeine, which is an ATM/ATR inhibitor (22) (Fig.
2A). Similarly, G2/M cell cycle arrest activates rapidly when NaCl
decreases, and this arrest is abrogated by caffeine (Fig. 2B).
Repair of the breaks after NaCl is reduced is accompanied by
formation of foci containing phosphorylated H2AX (γH2AX) (1,
5). This histone modification occurs around DNA double-strand
breaks and contributes to their repair (13, 14). γH2AX is in-
duced to a maximal level within 15 min after lowering NaCl, then
gradually decreases (Fig. 2C) accompanying repair of the DSBs
(Fig. 1C). The γH2AX induction is also sensitive to inhibition of
ATM/ATR by caffeine (Fig. 2C). These results indicate that,
when elevated NaCl is lowered, a classical ATM/ATR-dependent
DNA damage response becomes activated, and they further
confirm that DNA DSBs increase upon exposure to high NaCl.
During Repair of High NaCl-Induced DNA Breaks, γH2AX Is Mainly
Located in Gene Deserts, Whereas Bleomycin and UV Induce γH2AX
at Random Locations Throughout Genome. Given that high NaCl
induces DSBs that persist as long as the level of NaCl stays high,
it has not been clear how cells survive and function in the con-
tinued presence of those breaks, why the cell cycle does not
remain arrested, how DNA transcription and replication can
proceed despite the breaks, and how mutations and genomic
instability are prevented. To begin answering those questions, we
have determined the genomic location of the breaks. Our strat-
egy is based on the fact that during repair of DSBs, γH2AX is
induced in distinct foci (Fig. 3). γH2AX-containing foci occur at
DSBs (13, 14), and the number of γ-H2AX foci approximates the
number of DSBs induced (ref. 14 and reviewed in ref. 23),
making foci of γH2AX consistent and quantitative markers of
DSBs. Therefore, to detect locations of DSBs, we performed
chromatin immunoprecipitation (ChIP) by using anti-γH2AX
antibody, followed by massive parallel sequencing of isolated
DNA fragments (γH2AX ChIP-Seq) (24). Given that the maxi-
mal intensity of γH2AX foci occurs 15 min after NaCl is lowered
and that the intensity of the foci already decreases by 30 min
(Fig. 3B), we determined the genomic locations of the γH2AX
foci that are present 15 min after lowering NaCl. In addition, we
performed ChIP-Seq to determine the genomic locations of the
γH2AX induced by bleomycin and UV radiation. Both bleomy-
cin and UV increase γH2AX (Fig. 2D). However, pattern of
γH2AX immunostaining is different (Fig. 3A) because of the
different nature of damage induced by bleomycin and UV. Thus,
bleomycin induces DSBs directly (25) and produces distinct
γH2AX foci (Fig. 3A), whereas, after UV radiation, DSBs arise
only indirectly as a result of the action of repair or degradation of
arrested replication forks (26). After UV radiation, γH2AX is
increased both in foci and in more diffused staining (Fig. 3A)
that might not be related to DSBs but to some other types of
DNA damage (27). Thus, by γH2AX ChIP-Seq we analyzed
genomic locations of γH2AX induced by bleomycin, UV, and
during repair of high NaCl-induced DNA breaks. We aligned
sequence reads (tags) to the mouse genome and analyzed the tag
density in the University of California, Santa Cruz (UCSC) ge-
nome browser (Fig. 4 and Fig. S5).
15 min after lowering NaCl. NaCl bathing mIMCD3 cells was elevated by adding NaCl to total osmolality of 500 mosmol/kg for 22 h, then the additional NaCl
was eliminated, lowering the osmolality to 300 mosmol/kg. Cells were exposed to bleomycin (5 μg/mL for 30 min) or 15 J/m2of UV light. (A) Immunocyto-
chemical staining for γH2AX demonstrating γH2AX foci. (B and C) Analysis of brightness of γH2AX foci by laser scanning cytometry (LSC). (B) Cytograms (LSC),
plotting Maximal Pixel Fluorescence (Max Pixel) of γH2AX in nuclei (a measure of γH2AX brightness of foci) vs. Integral DAPI (blue) fluorescence (a measure of
nuclear DNA content). Numbers on cytogramms show % of cells in which Max Pixel of γH2AX fluorescence is above the level shown by horizontal lane.
Brightness of γH2AX fluorescence and proportion of cells in which it increases are maximal 15 min after NaCl is lowered. Bleomycin and UV increase γH2AX as
expected. (C) Cell distributions based on γH2AX fluorescence intensity. γH2AX Fluorescence intensity increases in a fraction of cells after treatment with
bleomycin and UV and 15 min after lowering NaCl.
γH2AX induction after lowering of NaCl occurs in distinct foci, similar to those from other causes of DSBs, and the brightness of the foci is maximal
| www.pnas.org/cgi/doi/10.1073/pnas.1114677108 Dmitrieva et al.
There are clusters of increased density of mapped tags
(γH2AX peaks) in cells after reduction of high NaCl (Fig. 4A
and Fig. S5). The peaks occur mainly in large intergenic regions
(gene deserts). Thus, H2AX becomes phosphorylated at specific
locations within gene deserts during repair of the high NaCl-
induced DSBs, showing that the high NaCl-induced DSBs are
not randomly distributed throughout genome but occur within
gene deserts. The peaks do not occur in DNA from cells main-
tained continuously at 300 or 500 mosmol/kg (Fig. 4A, “Control”
and “High NaCl, 22h”). Bleomycin and UV also increase
γH2AX (Figs. 2D and 3), but peaks do not occur in DNA from
cells in which it has been damaged by Bleomycin or UV irradiation
(Fig. 4A, “Bleomycin” and “UV”). Thus, the DNA damage caused
by Bleomycin and UV occurs randomly throughout the genome, in
marked contrast to the localized DSBs caused by high NaCl.
The localization of peaks to gene deserts is apparent by visual
examination of tag density in the genome browser. Representa-
tive regions are shown in Fig. 4 and Fig. S5. To apply this ob-
servation to the whole genome, we extracted information about
peak coordinates from the UCSC genome browser and analyzed
gene density over the regions of identified peaks in comparison
with artificial peaks of the same width and number that were
randomly generated throughout genome (Fig. 4B, Upper). We
identified 215 peaks of mean width of 2.4 Mb (95% confidence
interval between 2.2 Mb and 2.7 Mb). Gene density over the
peak regions is greatly reduced compared with random peaks or
average gene density in mouse genome (Fig. 4B, Lower).
PFGE Identifies Increased DNA Fragmentation After Exposure of
mIMCD3 Cells to High NaCl, and the Distribution of the Lengths of
the Fragments Is Consistent with Spacing of the γH2AX Peaks. We
used PFGE, which separates large DNA fragments according to
size (28), to further test whether high NaCl induces double-
strand breaks and whether the breaks are not randomly distrib-
uted (Fig. 5). Exposure of cells to high NaCl increases the
number of DNA fragments that enter the gel and the size of the
fragments is not random, consistent with a nonrandom location
of the DNA breaks (Fig. 5A). Given the locations of the γH2AX
peaks and assuming that there is one DNA break in each peak,
we calculated a size distribution of the predicted DNA fragments
(Fig. 5B and SI Materials and Methods). The calculated distri-
bution of fragment sizes closely resembles the actual distribution
(Fig. 5), which supports the conclusion that high NaCl-induced
double-strand breaks occur predominantly in gene deserts.
Why there are gene deserts has remained a mystery. They were
discovered when whole genome sequencing showed that genes
are not evenly distributed. A substantial fraction of mammalian
genomes contains gene deserts, defined as long regions (>500
kb) containing no protein-coding sequences. Gene deserts oc-
immunoprecipitated by anti-γH2AX. Chromatin was digested with MNase to mononucleosome size. Mononucleosomes containing γH2AX were immuno-
precipitated with anti-γH2AX antibodyandpurified DNA fragments were sequenced by usinga Solexa 2G GenomeAnalyzer.(A) UCSCgenome browser view of
representative examples of sequence tags mapped to the mouse genome. The lowest track shows location of genes (UCSC genome browser). ChIP-seq was
performed by using mIMCD3 cells treated, as follows. Control cells maintained at 300 mosmol/kg. Bleomycin cells were exposed to 5 μg/mL bleomycin for
30 min. “UV” cells at 300 mosmol/kg exposed to UV. “High NaCl” cells maintained at 500 mosmol/kg (NaCl added) for 22 h. “Return to Control” cells: NaCl was
elevated to total osmolality of 500 mosmol/kg for 22 h, then NaCl was lowered to a total osmolality of 300 mosmol/kg for 15 min. Note the distinct peaks of
increased sequence tag density in Return to Control samples indicating that the γH2AX foci occur at specific locations within genome. Note further that such
peaks donot occur in DNA from cells under any ofthe other conditions. The gray projections of the peaks in Return to Control demonstrate locations in areas of
genome containing few genes (gene deserts). The lack of peaks in conditions other than return to control indicates that the γH2AX foci are located randomly in
the genome under those conditions. Additional examples are shown in Fig. S5. (B) Relation between the observed γH2AX peaks and gene density throughout
the mouse genome. (B Upper) Calculated number of genes per megabase within each peak of DNA immunoprecipitated by anti-γH2AX antibody during repair
of high NaCl-induced DSBs. The number of peaks containing each gene densities (Genes per MB) is plotted (Actual Peaks, red), compared with the gene
densities in randomly generated peaks of the same size and number in each chromosome (Random Peaks, blue). Evidently, gene density is low in γH2AX peaks
throughout the genome. (B Lower) The gene density in the γH2AX peaks is also significantly below the average gene density throughout the genome.
During repair of high NaCl-induced DNA breaks, γH2AX is mainly located in gene deserts. ChIP-Seq analysis of genomic locations of DNA fragments
Dmitrieva et al.PNAS
| December 20, 2011
| vol. 108
| no. 51
cupy ≈38% of human, 34% of mouse, 23% of rat, and 20% of
dog genome (29). It is extremely unlikely that gene deserts
reached their observed maximal size of 5.1 Mb with 545 deserts
larger than 640 kb by chance (30), which raises the question of
what selective pressure might be acting.
Animal cells are universally exposed to NaCl, and the level of
NaCl may be high in animals exposed to marine or desiccated
terrestrial environments. During evolution of mammals, osmoreg-
ulatory mechanisms developed that maintain osmolality of most
extracellular fluids close to 300 mosmol/kg. Nevertheless, even in
mammals, NaCl concentration is constantly very high in some tis-
sues, particularly the renal medulla. Given our finding that DNA
breaks induced by high NaCl are concentrated in gene deserts, we
suggest that, as the size of genomes has increased, newly formed
regions are susceptible (for unknown reasons) to high NaCl-in-
duced DNA breaks and evolve to contain fewer genes, thus limiting
mutations and preventing genomic instability. This suggestion is
supported by several observations.
i) The neutral mutation rate (30) and the rate of genome
rearrangements associated with appearances of new centro-
meres (31) are both higher in gene deserts than in regions
ii) Before the evolution of vertebrates, the sizes of genomes
grew in proportion to the number of genes. However, gene
deserts began appearing in fish and increased in size to
occupy ≈38% of the genome in humans (29). Over the same
period, osmoregulatory mechanisms developed that main-
tain systemic osmolality close to 300 mosmol/kg. Estimates
from molecular clocks of the rates of evolution show that
the rates decreased significantly in vertebrates before the
origin of Osteichthyes (32). That could have been due to
a combination of decreased rate of mutations in protein
coding regions owing to more precise osmoregulation and
the low abundance of functional genes in gene deserts
where they would be susceptible to NaCl-induced breaks.
iii) Recently, a model was proposed relating the rate of molec-
ular evolution and the maximal size of genomes (33). The
theory assumes that for an organism to be viable, essential
genes must be functional. Further, it predicts that popula-
tions become extinct because of lethal mutagenesis when
the mutation rate exceeds approximately six mutations per
replication in essential parts of the genome in mesophilic
organisms and one or two mutations in thermophilic ones.
The theory therefore predicts that mutation rate limits es-
sential genome size; in other words, the higher the muta-
tion rate, the smaller the sustainable size of the genome.
This theory implies that increasing the size of the genome
required that genes not evolve in regions, like the present
gene deserts, that are more susceptible to DNA breaks
Our finding that high NaCl-induced DSBs are located in gene
deserts is an example of nonrandom induction of DNA breaks in
higher organisms. Although we are uncertain why high NaCl breaks
DNA, the gene deserts apparently have properties that render
them more susceptible. Limitation of high NaCl-induced DNA
breaks to gene deserts helps explain why they apparently are less
harmful than are the random breaks induced by genotoxic agents
like UV radiation, ionizing radiation, and oxidants. Further, our
finding suggests a possible role of high NaCl in evolution of the
structure of the animal genome.
breaks occur predominantly in gene deserts during exposure to high
NaCl. Possibilities that we are considering include decreased DNA
mosmol/kg for the indicated times or 5μg/mL bleomycin was added for 30 min. Cells kept in high NaCl for 72 h were split once to maintain logarithmic growth
(representative of three independent experiments). High NaCl increases DNA fragmentation, and there are size peaks of the DNA fragments, consistent with
nonrandom location of the DNA breaks. (B) The distribution of DNA fragment lengths is consistent with that predicted from the distance between γH2AX
peaks. We assumed one break per γH2AX peak, present at a maximal probability at the middle of the peak. Upper shows the probability distribution for
fragments of lengths from 0 to 60 Mb. Because the limit of detection for PFGE is ≈6 Mb, Lower is restricted to predicted sizes <10 Mb. Note that the peaks of
predicted fragment sizes (labeled 1, 2, and 3) are consistent with the peaks seen experimentally in A.
(A) High NaCl increases DNA fragmentation in mIMCD3 cells (pulsed-field gel electrophoresis, PFGE). NaCl was elevated to a total osmolality of 500
| www.pnas.org/cgi/doi/10.1073/pnas.1114677108Dmitrieva et al.
repair in gene deserts similar to that in heterochromatin (34), pres- Download full-text
ence of specific target sequences for nucleases activated by high
that makes the DNA there more susceptible to damaging agents.
Materials and Methods
Methods were published for Western blot (1), exposure of cells to UV radi-
ation (5), treatment with H2O2(20), analysis of cells in mitosis (6), immu-
nostaining, and analysis of brightness of γH2AX foci by laser-scanning
cytometry (LSC) (5). More details are included in SI Materials and Methods, as
are details of doses and timing of drug application, analysis of gene density
at genomic locations enriched with γH2AX-immunoprecipitated sequence
tags, and analysis of expected distribution of DNA fragment sizes, based on
genomic locations of γH2AX ChIP-Seq peaks.
Cell Culture. mIMCD3 cells (35) were grown in medium containing 45% DME
Low Glucose (Invitrogen), 45% F12 Coon’s Modification (No. F6636; Sigma),
and 10% FBS (HyClone). Osmolality of control medium was 300–320 mosmol/
kg. High NaCl medium was prepared by adding NaCl to the total osmolality
of 500 mosmol/kg. All of the experiments were performed on logarithmi-
cally growing cells at ≈80% confluence. To elevate NaCl, control medium
was replaced by the high NaCl media.
Analysis of Double-Strand and Single-Strand DNA Breaks by Comet Assay. Two
different assays were used: neutral comet assay modified for detection of
double-strand breaks and alkaline comet assay, which detects DNA SSBs,
double-strand breaks, and alkali-labile sites. Those assays were performed as
described (19) with minor modifications. See SI Materials and Methods and
Fig. S4 for details.
ChIP and Illumina Library Construction for Sequencing. ChIP was performed by
using Enzymatic Chromatin IP kit (No. 9003; Cell Signaling Technology).
Conversion of the ChIP-enriched DNA into libraries suitable for sequencing
using the Illumina Genome Analyzer was performed by using the published
protocol (36). See SI Materials and Methods and Fig. S4 for the detailed ChIP-
Solexa Pipeline Analysis. Sequence tags were obtained and mapped to the
mouse genome by using the Solexa Analysis Pipeline as described (37). The
unique reads were retained and converted to browser extensible data (BED)
files for viewing the data in the UCSC genome browser. The read number
and genomic coordinates were summarized in 300-bp windows.
Analysis of DNA Fragmentation by PFGE. Agarose embedded DNA was pre-
pared by using the CHEF Mammalian Genomic DNA Plug Kit (No. 170–3591;
Bio-Rad). Briefly, cells were rinsed with PBS, scraped off the dish, resus-
pended in 1% CleanCut Agarose from the kit at a final concentration of 12
million cells/mL. The agarose/cell suspension was solidified at 4 °C for 10 min
in a casting mold, followed by incubation of the agarose plugs in Proteinase
K solution at 50 °C for 3 d to digest proteins. PFGE was performed as de-
scribed (28) by using the CHEF-DR II system (Bio-Rad) and the following
parameters: 1% Megabase Agarose (No. 161–3108; Bio-Rad), 0.5× TBE run-
ning buffer, 120° reorientation angle, 6 V·cm−1, and 14 °C. Gels were run for
16 h with switch time of 16 s, followed by 30-h run with switch time of 80 s
DNA Size Markers were as follows: Schizosaccharomyes pombe, Saccharo-
myces cerevisiae, and Hansenula wingei chromosomes (Bio-Rad). Gels were
stained with SYBR Gold (Invitrogen).
ACKNOWLEDGMENTS. We thank Drs. Chris Combs and Daniela Malide at
the National Heart, Lung, and Blood Institute (NHLBI) Light Microscopy Core
Facility for help with microscopy and images processing and Dr. Iouri
Chepelev at the NHLBI Laboratory of Molecular Immunology for assistance
with sequencing data analysis. This research was supported by the Intra-
mural Research Programs of the National Institutes of Health, NHLBI.
1. Dmitrieva NI, Cai Q, Burg MB (2004) Cells adapted to high NaCl have many DNA
breaks and impaired DNA repair both in cell culture and in vivo. Proc Natl Acad Sci
2. Kültz D, Chakravarty D (2001) Hyperosmolality in the form of elevated NaCl but not
urea causes DNA damage in murine kidney cells. Proc Natl Acad Sci USA 98:1999–2004.
3. Dmitrieva NI, Celeste A, Nussenzweig A, Burg MB (2005) Ku86 preserves chromatin
integrity in cells adapted to high NaCl. Proc Natl Acad Sci USA 102:10730–10735.
4. Dmitrieva NI, Ferraris JD, Norenburg JL, Burg MB (2006) The saltiness of the sea breaks
DNA in marine invertebrates: Possible implications for animal evolution. Cell Cycle 5:
5. Dmitrieva NI, Bulavin DV, Burg MB (2003) High NaCl causes Mre11 to leave the nu-
cleus, disrupting DNA damage signaling and repair. Am J Physiol Renal Physiol 285:
6. Dmitrieva NI, Bulavin DV, Fornace AJ, Jr., Burg MB (2002) Rapid activation of G2/M
checkpoint after hypertonic stress in renal inner medullary epithelial (IME) cells is
protective and requires p38 kinase. Proc Natl Acad Sci USA 99:184–189.
7. Michea L, et al. (2000) Cell cycle delay and apoptosis are induced by high salt and urea
in renal medullary cells. Am J Physiol Renal Physiol 278:F209–F218.
8. Dmitrieva NI, Burg MB (2004) Living with DNA breaks is an everyday reality for cells
adapted to high NaCl. Cell Cycle 3:561–563.
9. Dmitrieva NI, Burg MB (2008) Analysis of DNA breaks, DNA damage response, and
apoptosis produced by high NaCl. Am J Physiol Renal Physiol 295:F1678–F1688.
10. Bankir L (1996) Urea and the kidney. The Kidney, ed Brenner BM (WB Saunders
Company, Philadelphia), Vol 5, pp 571–606.
11. Lamitina ST, Morrison R, Moeckel GW, Strange K (2004) Adaptation of the nematode
Caenorhabditis elegans to extreme osmotic stress. Am J Physiol Cell Physiol 286:
12. Sherwood L, Klandorf H, Yancey P (2005) Animal Physiology. From Genes to Organ-
isms (Thomson Learning, Belmont, CA).
13. Rogakou EP, Pilch DR, Orr AH, Ivanova VS, Bonner WM (1998) DNA double-stranded
breaks induce histone H2AX phosphorylation on serine 139. J Biol Chem 273:
14. Sedelnikova OA, Rogakou EP, Panyutin IG, Bonner WM (2002) Quantitative detection
of (125)IdU-induced DNA double-strand breaks with gamma-H2AX antibody. Radiat
15. Ostling O, Johanson KJ (1984) Microelectrophoretic study of radiation-induced DNA
damages in individual mammalian cells. Biochem Biophys Res Commun 123:291–298.
16. Dmitrieva NI, Burg MB (2007) Osmotic stress and DNA damage. Methods Enzymol
17. Collins AR, Dobson VL, Dusinská M, Kennedy G, Stĕtina R (1997) The comet assay:
What can it really tell us? Mutat Res 375:183–193.
18. Collins AR (2004) The comet assay for DNA damage and repair: Principles, applica-
tions, and limitations. Mol Biotechnol 26:249–261.
19. Olive PL, Banáth JP (2006) The comet assay: A method to measure DNA damage in
individual cells. Nat Protoc 1:23–29.
20. Olive PL, Johnston PJ (1997) DNA damage from oxidants: Influence of lesion com-
plexity and chromatin organization. Oncol Res 9:287–294.
21. Dai Y, Grant S (2010) New insights into checkpoint kinase 1 in the DNA damage re-
sponse signaling network. Clin Cancer Res 16:376–383.
22. Zhou BB, et al. (2000) Caffeine abolishes the mammalian G(2)/M DNA damage
checkpoint by inhibiting ataxia-telangiectasia-mutated kinase activity. J Biol Chem
23. Kinner A, Wu W, Staudt C, Iliakis G (2008) Gamma-H2AX in recognition and signaling
of DNA double-strand breaks in the context of chromatin. Nucleic Acids Res 36:
24. Wold B, Myers RM (2008) Sequence census methods for functional genomics. Nat
25. Hecht SM (2000) Bleomycin: New perspectives on the mechanism of action. J Nat Prod
26. Limoli CL, Giedzinski E, Bonner WM, Cleaver JE (2002) UV-induced replication arrest in
the xeroderma pigmentosum variant leads to DNA double-strand breaks, gamma -H2AX
formation, and Mre11 relocalization. Proc Natl Acad Sci USA 99:233–238.
27. de Feraudy S, Revet I, Bezrookove V, Feeney L, Cleaver JE (2010) A minority of foci or
pan-nuclear apoptotic staining of gammaH2AX in the S phase after UV damage
contain DNA double-strand breaks. Proc Natl Acad Sci USA 107:6870–6875.
28. Herschleb J, Ananiev G, Schwartz DC (2007) Pulsed-field gel electrophoresis. Nat
29. Salzburger W, Steinke D, Braasch I, Meyer A (2009) Genome desertification in eu-
therians: Can gene deserts explain the uneven distribution of genes in placental
mammalian genomes? J Mol Evol 69:207–216.
30. Ovcharenko I, et al. (2005) Evolution and functional classification of vertebrate gene
deserts. Genome Res 15:137–145.
31. Lomiento M, Jiang Z, D’Addabbo P, Eichler EE, Rocchi M (2008) Evolutionary-new
centromeres preferentially emerge within gene deserts. Genome Biol 9:R173.
32. Peterson KJ, et al. (2004) Estimating metazoan divergence times with a molecular
clock. Proc Natl Acad Sci USA 101:6536–6541.
33. Zeldovich KB, Chen P, Shakhnovich EI (2007) Protein stability imposes limits on or-
ganism complexity and speed of molecular evolution. Proc Natl Acad Sci USA 104:
34. Goodarzi AA, Jeggo P, Lobrich M (2010) The influence of heterochromatin on DNA
double strand break repair: Getting the strong, silent type to relax. DNA Repair
35. Rauchman MI, Nigam SK, Delpire E, Gullans SR (1993) An osmotically tolerant inner
medullary collecting duct cell line from an SV40 transgenic mouse. Am J Physiol 265:
36. Schmidt D, et al. (2009) ChIP-seq: Using high-throughput sequencing to discover
protein-DNA interactions. Methods 48:240–248.
37. Barski A, et al. (2007) High-resolution profiling of histone methylations in the human
genome. Cell 129:823–837.
Dmitrieva et al.PNAS
| December 20, 2011
| vol. 108
| no. 51