Molecular Human Reproduction Vol.10, No.3 pp. 203±209, 2004
Advance Access publication January 29, 2004
DOI: 10.1093/molehr/gah029
Structure of human sperm DNA and background damage,
analysed by in situ enzymatic treatment and digital image
analysis
Lourdes Muriel1, Enrique Segrelles2, Vicente Goyanes1,2, Jaime GosaÂlvez3 and
Jose Luis FernaÂndez1,2,4
1
SeccioÂn de GeneÂtica y Unidad de InvestigacioÂn, Hospital `Teresa Herrera', Complejo Hospitalario Universitario Juan Canalejo,
15006 La CorunÄa, 2Unidad de la Mujer, La CorunÄa and 3Unidad de GeneÂtica, Facultad de BiologõÂa, Universidad AutoÂnoma de
Madrid, Spain
4
To whom correspondence should be addressed at: SeccioÂn de GeneÂtica, Hospital `Teresa Herrera', Complejo Hospitalario
Universitario Juan Canalejo, 15006 La CorunÄa, Spain. E-mail: [email protected]
DNA breakage detection±¯uorescence in situ hybridization (DBD±FISH) is a procedure to detect and quantify DNA breaks in
situ, on a cell-by-cell basis. A comparison between sperm nuclei versus peripheral blood leukocytes using this method demonstrated that the nucleoids from mature human sperm are 12.7 times more sensitive to alkaline denaturation than those from
human peripheral blood leukocytes. To investigate the origin of this alkali sensitivity, different approaches were employed.
First, free 3¢-OH ends of background DNA breaks were labelled by Klenow polymerase, or by DNA polymerase I following the
in situ nick translation assay. Second, the presence of abasic sites, the other recognized DNA lesions that lends to constitutive
alkali sensitivity, and DNA breaks with blocked 3¢ ends, were determined by in situ exonuclease III digestion prior to the polymerase labelling. The results demonstrated that the sperm nucleoid contains ~2.5-fold higher density of background DNA breaks
with 3¢-OH ends, and also ~2.8-fold higher density of basal abasic sites and DNA breaks with blocked 3¢ termini, than leukocytes.
These differences only partially explain the signi®cant alkali sensitivity of sperm DNA. However, in situ digestion with mung
bean nuclease before DNA break labelling showed that sperm DNA is 9-fold more enriched in segments of ssDNA than DNA
from leukocytes. The high frequency of partially denatured regions may result from a greater torsional stress of DNA loops in
sperm chromatin due to its higher degree of compaction. Moreover, these short unpaired ssDNA stretches should be included in
the category of alkali-labile sites detected by all techniques that measure DNA breaks through an alkaline unwinding step.
These results provide new insights into the nature of DNA packaging in sperm nuclei.
Key words: alkali-labile sites/alkaline DNA unwinding/chromatin structure/DBD±FISH/sperm DNA
Introduction
The integrity of sperm DNA is of crucial importance for balanced
transmission of genetic information to future generations. There is
good evidence that sperm DNA damage may lead to conception
failure, abortions, malformations, and genetic diseases. Furthermore,
the origin of 80% of de novo structural chromosome aberrations in
man are of paternal origin (Tomar et al., 1984). During the last two
decades, an interspecies IVF system between human sperm and golden
hamster oocytes allowed the study of sperm-derived chromosomes.
These studies demonstrated that human sperm contain a higher
incidence of baseline numerical and structural chromosome aberrations than somatic cells, as well as a higher incidence of chromosome
aberrations after in vitro and in vivo exposure to different mutagens
(Martin et al., 1989; Genesca et al., 1990; Kamiguchi and Tateno,
2002). More recently, the use of ¯uorescence in situ hybridization
(FISH) demonstrated a signi®cantly higher incidence of structural
aberrations than numerical abnormalities in human sperm, con®rming
the high frequency previously measured by the hamster oocyte method
(Sloter et al., 2000). Moreover, it is now recognized in humans that
couples in whom the male semen contains a high percentage of sperm
cells with heavily fragmented DNA have a very low potential for
natural or assisted reproduction. Thus, recent clinical studies indicate
that DNA fragmentation levels >30%, as measured by the sperm
chromatin structure assay (SCSA) test, are incompatible with the
initiation and maintenance of a term pregnancy (Evenson et al., 1999;
Larson et al., 2000).
At the DNA level, different techniques have been applied to sperm
cells to analyse DNA damage. Many of them are based on the
detection of DNA strand breaks through an alkaline unwinding
treatment. The alkali transforms DNA breaks into ssDNA that are
initiated from the ends of the break (Rydberg, 1975). These techniques
differ in the ways in which the ssDNA is analysed. The alkaline singlecell gel electrophoresis (SCGE) is a well-established morphological
procedure that allows cell-by-cell evaluation. In this method, the
length and amount of DNA that migrates to the tail, as determined
after DNA staining with a ¯uorochrome, is proportional to the amount
of breaks per cell. Alkaline unwinding detects alkali-labile sites in
addition to single- and double-DNA strand breaks, i.e. DNA lesions or
modi®cations that are transformed into DNA breaks by the alkali.
These DNA modi®cations are classically referred to as DNA abasic
Molecular Human Reproduction vol. 10 no. 3 ã European Society of Human Reproduction and Embryology 2004; all rights reserved
203
L.Muriel et al.
sites, apurinic or apyrimidinic (AP), or deoxyribose lesions (von
Sonntag, 1987). Though a steady-state level of DNA breaks and AP
sites spontaneously occurs in living cells (Atamna et al., 2000),
deoxyribose lesions are probably more characteristic of radical or
exogenous damage. Surprisingly, using the SCGE assay, it was found
that DNA from mature sperm cells contains a large number of alkalilabile sites (Singh et al., 1989; Haines et al., 1998). Alkaline elution
techniques also supported these results (Singh et al., 1989; Van Loon
et al., 1991, 1993). The nature of these alkali-labile sites has not been
de®ned, although it has been speculated that they may be a
characteristic of condensed chromatin. This assumption is supported
by the fact that chicken erythrocytes and mouse kidney DNA, both
highly condensed, revealed a similar high density of constitutive
alkali-labile sites (Singh et al., 1989; Fairbain et al., 1994).
In the present report we have applied a battery of in situ DNA
labelling techniques to gain information on the nature of the intense
background sensitivity to alkali of DNA from human sperm cells, as
well as to establish baseline DNA damage, in comparison with
somatic cells (i.e. human blood leukocytes).
Materials and methods
Preparation of nucleoids in agarose microgels
The study was carried out with human blood leukocytes and mature sperm from
the same donor, in two healthy donors, obtained with informed consent. Sperm
cells in raw semen were diluted in Roswell Park Memorial Institute 1640
medium (Gibco, Invitrogen, UK) to obtain a sperm concentration of 103106/
ml. Leukocytes in buffy coat and diluted sperm cells were separately mixed
with low-melting point agarose (Pronadisa; Hispanlab, Spain) to produce a ®nal
concentration of 0.7% agarose at 37°C. Then 28 ml of each mixture was
pipetted onto the same glass slide precoated with 0.65% standard agarose dried
at 80°C, covered with a cover slip (12360 mm), and left to solidify at 4°C. The
longitudinal half of the slide contained the leukocytes, whereas the other half
contained the sperm. Thus, both cell types were simultaneously processed in
the same glass slide, under the same conditions. Cover slips were carefully
removed, and the slides immediately immersed horizontally in a tray with a
large volume of lysing solution 1 [0.4 mol/l Tris, 0.8 mol/l dithiothreitol (DTT),
1% sodium dodecyl sulphate (SDS), 50 mmol/l EDTA, pH 7.5] for 10 min at
room temperature, followed by incubation in lysing solution 2 (0.4 mol/l Tris, 2
mol/l NaCl, 1% SDS, pH 7.5) for 5 min at room temperature. Finally, the slides
were thoroughly washed in phosphate-buffered saline (PBS), four times, 5 min
each.
DBD±FISH procedure to detect DNA breaks and alkali-labile
sites
After washing in PBS (Sigma, USA), the nucleoids were washed twice in 0.9%
NaCl, 5 min each, and then incubated in an alkaline unwinding solution (0.03
mol/l NaOH, 1 mol/l NaCl), for 2.5 min at 7°C in the dark. This transforms
DNA breaks and alkali-labile sites into short ssDNA regions. After
neutralization with a 5 min incubation in 0.4 mol/l Tris±HCl (pH 7.5), the
slides were washed in TBE buffer (0.09 mol/l Tris±borate, 0.002 mol/l EDTA,
pH 7.5) for 2 min, dehydrated in sequential 70, 90 and 100% ethanol baths, 2
min each, and air-dried. A human whole genome probe (4.3 ng/ml in 50%
formamide/23SSC, 10% dextran sulphate, 100 mmol/l calcium phosphate, pH
7.0) (13SSC is 0.015 mol/l Na citrate, 0.15 mol/l NaCl, pH 7.0), biotin-labelled
by nick-translation, was denatured and hybridized overnight at room
temperature on dried slides. After hybridization, slides were washed twice in
50% formamide/23SSC, pH 7.0, for 5 min, and twice in 23SSC pH 7.0, for 3
min, at room temperature. Hybridized probe was detected with streptavidin±
Cy3 (1:200) (Sigma, USA) and cells were counterstained with 4,6-diamidino-2phenylindole (DAPI) in Vectashield (Vector, USA) (FernaÂndez et al., 1998,
2002; FernaÂndez and GosaÂlvez, 2002).
Labelling of DNA breaks with 3¢-OH ends
Two methods were used to label DNA breaks with 3¢-OH ends.
204
End labelling with Klenow polymerase.
After the PBS washes, the slides were incubated four times, 5 min each, in
excess reaction buffer for Klenow polymerase (10 mmol/l Tris±HCl, 5 mmol/l
MgCl2, 7.5 mmol/l DTT, pH 7.5). Reaction buffer (100 ml) containing 25 IU of
Klenow polymerase (New England BioLabs, USA) and biotin-16-dUTP in the
nucleotide mix were pipetted onto the slide, covered with a plastic cover slip,
and incubated in a moist chamber for 30 min at 37°C. After washing in TBE
buffer, the slides were dehydrated in sequential 70, 90 and 100% ethanol baths,
and air-dried. The incorporated biotin-16-dUTP was detected by a 30 min
incubation with streptavidin±Cy3, and slides were DAPI-counterstained.
Reaction buffer only was placed on one area of the slide, as a control. The
microgel between the areas with and without the polymerase was scratched to
avoid the possible diffusion of the enzyme to the control area.
In situ nick translation (ISNT) labelling with DNA polymerase I.
The protocol and reaction buffer were similar to that of Klenow polymerase. In
this case, 20 IU of DNA polymerase I (New England BioLabs) in 100 ml of
reaction buffer (10 mmol/l Tris±HCl, 5 mmol/l MgCl2, 7.5 mmol/l DTT, pH
7.5) were incubated on the slide.
Labelling of AP sites and DNA breaks with blocked 3¢ ends
After PBS washes, the nucleoids were incubated in excess exonuclease III
reaction buffer (0.66 mmol/l MgCl2, 66 mmol/l Tris±HCl, pH 7.0), four times,
5 min each. A transverse half of the slide, including an area of the microgel
containing sperm cells and another area of the microgel containing leukocytes,
were incubated with 100 IU of exonuclease III (New England BioLabs) in 50 ml
of reaction buffer, for 30 min at 37°C. The other half was covered by
exonuclease III buffer alone. Previously, the microgel between both halves had
been scratched. After four thorough washings in PBS, the slides were processed
for Klenow end labelling or ISNT with DNA polymerase I. as described above.
The increase in labelling within the exonuclease III-digested area compared to
the undigested area, (exonuclease III + Klenow or DNA polymerase I) ±
(Klenow or DNA polymerase I alone), is dependent on AP sites cleaved by the
nuclease as well as on possible DNA breaks with blocked 3¢ ends.
In another set of experiments, after exonuclease III digestion, the slides were
washed in TBE buffer, dehydrated in ethanol baths and air-dried. The ssDNA
generated by the nuclease from DNA breaks and AP sites was detected by
hybridization with the whole genome probe, as described before. This is
equivalent to FISH with an enzymatic denaturation (GosaÂlvez et al., 2002).
ssDNA labelling
The protocol was similar to that of the AP site labelling. In this case, initial
incubations were in excess mung bean nuclease reaction buffer (50 mmol/l Na
acetate, 100 mmol/l NaCl, 1 mmol/l ZnCl2, pH 5), followed by 25 IU of mung
bean nuclease (New England BioLabs) in 50 ml of reaction buffer, for 45 min at
37°C. Finally, the slides were processed for Klenow end labelling. The increase
in labelling within the mung bean nuclease-digested area compared to the
undigested area, (mung bean + Klenow) ± (Klenow alone), is dependent on
ssDNA segments cleaved and removed by the nuclease.
Fluorescence microscopy and digital image analysis
Slides were viewed under a DMRB epi¯uorescence microscope (Leica,
Germany) equipped with a DMRD photoexposer, PL Fluotar 3100 objective
and appropriate ¯uorescence ®lters for DAPI and Cy3. Cy3 Images were
acquired using a high sensitivity CCD camera (Ultrapix 1600; Astrocam) that
detects >16 000 grey levels and allows subtraction of the current dark image
and correction for non-uniform sample illumination. Groups of 50 digital
images, one per cell, were taken for each experimental point, stored in the ®le
format of the camera (.apf) and then converted to .img ®les. Exposure times
were different depending on the experiment, in order to avoid saturation of the
signal. Thus, DBD±FISH images were taken for 200 ms, images from DNA
break labelling with Klenow and ISNT, as well as from exonuclease III-treated
nucleoids, were captured for 1 s, and those from the experiment with mung
bean nuclease were exposed for 500 ms. Each experiment was repeated at least
twice. Image analysis was performed using a semi-automatic routine designed
with Visilog 5.1 software (Noesis, France). This allows for thresholding,
background subtraction, and measures the whole ¯uorescence intensity
[surface area (in pixels)3mean ¯uorescence intensity (in arbitrary grey levels)]
Structure and background damage of human sperm DNA
of the signal. Statistical analysis was carried out using Student's t-test. It should
be noted that, for comparison with leukocytes, signals from sperm were
multiplied by 2 to correct for the difference in DNA ploidy.
Results and discussion
DBD±FISH is a sensitive technique to detect and quantify DNA
breaks in single cells either in the whole genome or within speci®c
DNA sequence areas. Cells trapped in an inert agarose microgel over a
slide are lysed and immersed in an alkaline unwinding solution that
produces ssDNA motifs starting from the ends of the DNA breaks.
After dehydration, the nucleoids are incubated with DNA probes. The
amount of hybridized probe in a target is related to the amount of
melted ssDNA generated by the alkali, which is in turn proportional to
the degree of local DNA breakage (FernaÂndez and GosaÂlvez, 2002).
Using a whole genome probe, we had previously observed that
nucleoids from mature human sperm yielded very strong ¯uorescent
signals (FernaÂndez et al., 2000). Figure 1a and b presents DBD±FISH
signals from human blood leukocytes and mature sperm cells from the
same individual, processed in the same slide under the same
experimental conditions. Using a restrictive alkaline incubation (2.5
min at 7°C), both cell types exhibit a heterogeneous labelling of the
nucleoids, showing areas with different signal strengths. A quantitative analysis of the ¯uorescence (Table I) reveals that under these
conditions, and correcting for the DNA content, the sperm cells have
12.7-fold more ¯uorescent signal than that from leukocytes. This
extreme alkali sensitivity of DNA from sperm cells had also been
reported using the alkaline SCGE or alkaline elution assays (Singh
et al., 1989; Van Loon et al., 1991, 1993). Unlike those sperm cells
with heavily fragmented DNA, this result was only obtained when the
alkaline unwinding step was performed on previously deproteinized
cells (FernaÂndez et al., 2000), suggesting the presence of a high
density of constitutive alkali-labile sites in the sperm DNA, that were
masked by the higher-order chromatin organization with protamines.
Otherwise, DTT could generate sulphur-centred radicals on the
protamines, and not on histones, due to the disruption of disulphide
bridges. This could potentially originate damage in the close DNA
molecule, speci®cally in sperm nucleoids. Nevertheless, additional
experiments using either proteinase K (1 mg/ml, overnight, at 37°C)
instead of DTT, or using the lysing solutions complemented with
freshly added 10% dimethylsulphoxide as radical scavenger, showed
the same differential result between sperm cells and leukocytes, after
DBD±FISH. It discards an artefact created by the cell preparation
technique. The nature of these constitutive alkali-labile sites has not
been established. Besides masked DNA breaks, they may correspond
to AP sites, both turning into ssDNA by alkaline treatment (TeÂoule,
1987; von Sonntag, 1987). In fact, DNA breaks and AP sites are
known to be at a steady-state level in living cells, continuously being
generated and repaired.
Different in situ labelling protocols were performed to study the
nature of the strong alkali-labile sensitivity of human sperm DNA.
These protocols assayed for DNA breaks or AP sites. Background
DNA breaks with 3¢-OH ends were labelled using both Klenow and
DNA polymerase I (Dierendonck, 2002; Wood, 2002). The Klenow
fragment of Escherichia coli DNA polymerase I is routinely employed
for in situ labelling of DNA breaks. This enzyme catalyses the
addition of nucleotides to the 3¢-OH end of a DNA break using the
complementary DNA strand as template. Though the 5¢±3¢ exonuclease activity from DNA polymerase I is absent, the Klenow
fragment conserves its 3¢±5¢ exonuclease proofreading activity, which
could produce a nick translation effect. Nevertheless, in the presence
of dNTP the polymerase activity counteracts further degradation, so
the nucleotide at the 3¢-OH end will be continuously replaced. The
®nal result is the gap or end-®lling, and subsequent labelling by the
incorporation of biotin-16-dUTP. Using this approach, sperm
nucleoids exhibit 2.6-fold more signal than leukocytes (Figure 1c, d;
Table I). That is, the density of basal single-strand breaks, nicks or
gaps with 3¢-OH ends, appears to be 2.6-fold higher in sperm than in
leukocytes. In the case of DNA polymerase I, in addition to the 5¢±3¢
polymerase activity and the 3¢-5¢ exonuclease activity, the enzyme has
a concurrent 5¢±3¢ exonuclease action. Therefore, the primer strand is
hydrolysed at the 5¢ side of the nick while the synthetic activity
catalyses the addition of nucleotides to the 3¢ side. The simultaneous
hydrolysis and synthesis results in the translation of the nick along the
DNA duplex in the 5¢±3¢direction, thus resulting in the incorporation
of a great amount of biotin-16-dUTP molecules per single DNA strand
break. This is re¯ected in a higher background labelling of DNA
breaks in both leukocytes and sperm with the DNA polymerase I than
when using the Klenow fragment (Figure 1e, f). Nevertheless, the
difference in signal between both cell types is very close to that found
for Klenow polymerase, that is, sperm cells contain 2.3 higher density
of background DNA breaks than leukocytes (Table I). It is accepted
that basal DNA breaks in the living cell correspond to single-strand
breaks or nicks, double-strand breaks being extremely rare, unless
apoptotic degradation or damage is induced by exogenous agents. The
labelling systems we employed require that the DNA break should
contain a free 3¢-OH end, as in the terminal uridine nucleotide end
labelling (TUNEL) assay, using the terminal deoxynucleotidyl
transferase (TdT). Though the latter enzyme is primer and template
independent, this is only of interest in the case of the necessity for
labelling blunt-ended or 5¢-recessed fragments, that cannot be detected
by the polymerases (Walker et al., 2002). Nevertheless, this only
should be the case when analysing double-strand breaks, which are
practically absent in living cells.
Free 3¢-OH ends are typical of the transient DNA breaks generated
by the excision repair machinery to correct for spontaneous base
damage and AP sites in the living cell. Though 3¢-OH ends may be
more predominant in background damage, 3¢-ends with phosphate or
phosphoglycolate groups could be present in other single-strand
breaks. These seem to be a consequence of oxidative damage, and
behave as blocks to the polymerase activity (TeÂoule, 1987; von
Sonntag, 1987). Other DNA lesions present under physiological
conditions that may lead to alkali sensitivity are AP sites. These
modi®cations are generated by spontaneous base loss through
cleavage of the glycosyl bond between deoxyribose and purines or
pyrimidines. The alkali incises the DNA 3¢ to the AP site, transforming it into a single-strand break. It has been established that the
spontaneous depyrimidination rate seems to be 20 times slower than
that of depurination (Lindahl and KarlstroÈm, 1973). Though with high
variability depending on the reports, the steady-state level of AP sites
in living cells has been recently estimated as <1 AP site per 106
nucleotides, and may increase with ageing (Atamna et al., 2000). To
investigate other DNA alterations, such as presence of AP sites under
physiological conditions, that may lead to alkali sensitivity, we used a
different experimental approach based on E. coli exonuclease III
activity. Exonuclease III is a 3¢±5¢ exonuclease speci®c for doublestranded DNA, which removes nucleotides starting from 3¢±ends,
producing ssDNA patches. These 3¢ ends may contain either a free
hydroxyl group or modi®cations that block the polymerases, since
exonuclease III removes the 3¢ blocking ends, to form free 3¢-OH
groups, suitable substrates for DNA polymerases. This enzyme also
has AP endonuclease activity, so that a 3¢-OH end is speci®cally
generated after in situ treatment of DNA. 3¢-OH ends generated by
exonuclease III can be labelled by the Klenow fragment or by ISNT
with DNA polymerase I (Liu et al., 2002). This is a similar approach to
that carried out by Legault et al. (1997), but they employed T4 DNA
205
L.Muriel et al.
Figure 1. Comparison of different ¯uorescent labellings between nucleoids from intact human blood leukocytes and from mature sperm. (a±b) DNA breakage
detection±¯uorescence in situ hybridization (DBD±FISH) signal after hybridization of a whole genome probe in leukocytes (a) and in sperm cells (b). (c±d)
End labelling of DNA breaks with the Klenow polymerase in leukocytes (c) and in sperm (d). In the latter, a sperm cell with fragmented DNA is shown in the
upper left, as a control of adequate enzymatic activity. (e±f) DNA breakage labelling by in situ nick translation (ISNT) with DNA polymerase I in leukocytes
(e) and sperm cells (f). (g±h) DNA breakage and apurinic or apyrimidinic (AP) sites labelled together by sequential exonuclease III±Klenow polymerase, in
leukocytes (g) and sperm (h). In this case, a sperm with heavily fragmented DNA is also shown as a control of adequate enzymatic labelling of DNA breaks.
(i±j) DNA breaks and AP sites labelled together by a sequential exonuclease III±ISNT, in leukocytes (i) and sperm (j). (k±l) FISH with a whole genome probe
to detect the ssDNA generated by exonuclease III on leukocytes (k) and on sperm (l). (m±p) End labelling by Klenow polymerase of background DNA breaks
in leukocytes (m) and sperm cells (o), and after cleavage of ssDNA sites by mung bean nuclease in both leukocytes (n) and sperm (p). Camera exposure times:
200 ms (a, b), 1 s (c±l) and 500 ms (m±p).
206
Structure and background damage of human sperm DNA
Table I. Values of whole ¯uorescence intensity (mean 6 SEM) obtained after different labelling protocols
in human blood leukocytes and sperm
DBD±FISH
Klenow
DNA Pol I
ExoIII + Klenow
ExoIII + DNA Pol I
(ExoIII + Klenow) ± Klenow
(ExoIII + DNA Pol I) ± DNA Pol I
ExoIII-FISH
Mung bean 0 IU + Klenow
Mung bean 0.5 IU + Klenow
(MB 0.5 IU + Kl) ± (MB 0 IU + Kl)
Leukocytes
Sperm
Sperma/
leukocytes
1243.4 6 612.2
1235.1 6 440.5
4776.4 6 1893.8
2638.4 6 649.1
11 964.1 6 2418.2
1403.3 6 208.6
7187.7 6 524.4
3294.9 6 966.8
652.4 6 169.9
921.9 6 226.2
269.5 6 56.3
7879.4 6 1959.1
1641.5 6 719.8
5576.0 6 1753.9
3809.5 6 1279.9
15 100.1 6 4176.7
2168.0 6 560.1
9524.2 6 2422.8
7418.5 6 2074.7
679.2 6 189.2
1895.5 6 917.8
1216.3 6 728.6
12.7
2.6
2.3
2.9
2.5
3.1
2.6
4.5
2.1
4.1
9.0
Values shown were divided by a thousand.
The exposure times of the CCD camera were 200 ms for DBD±FISH images, 1 s for Klenow, DNA
polymerase I and exonuclease III images, and 500 ms for mung bean images.
DNA Pol I = DNA polymerase I; Exo III = exonuclease III; MB = mung bean nuclease; Kl = Klenow
polymerase.
aTo correct for the difference in DNA content, sperm values were multiplied by 2.
polymerase to label free 3¢-OH ends, whereas the sum of AP sites and
single-strand breaks with both free 3¢-OH and 3¢ blocking groups were
labelled using E. coli endonuclease IV previously to the polymerase,
instead of exonuclease III. Nucleoids preincubated with exonuclease
III show the expected increase in the signal due to the extra labelling
of AP sites and DNA breaks with blocked 3¢ termini as compared to
nucleoids treated with either Klenow or DNA polymerase I to label
DNA breaks (compare Figure 1c±f with g±j). Thus, the contribution of
AP sites and these kinds of DNA breaks is estimated by subtracting the
value of the polymerase signal from the exonuclease III±polymerase
signal. In the case of the Klenow labelling, DNA from sperm cells
exhibits 3.1-fold higher density of AP sites and DNA breaks with 3¢
blocking groups than that from leukocytes (Figure 1g, h; Table I);
whereas after ISNT this difference is 2.6-fold (Figure 1i, j; Table I).
Otherwise, the ssDNA generated by exonuclease III from both the
DNA breaks and the cleaved AP sites can be detected by incubation
with a whole genome probe (Figure 1k, l). This is the equivalent to a
FISH using an enzymatic denaturation (GosaÂlvez et al., 2002). The
quantitative analysis showed that sperm cells have 4.5-fold more
signal from background DNA breaks, with 3¢-OH and blocked 3¢
termini, and AP sites together, than leukocytes. It is higher than the
difference detected when these whole lesions were labelled with
exonuclease III±Klenow or exonuclease III±DNA polymerase I
(Table I). This could be explained because sperm nucleoids show
internal areas of strong hybridization of the whole genome probe after
exonuclease III±FISH, far more marked than in leukocytes (Figure 1k,
l). These areas may correspond to chromocentres of clustered
centromeric±pericentromeric repetitive satellite DNA sequences
(Zalensky et al., 1995), which hybridize more strongly when using a
whole genome probe (FernaÂndez et al., 2000).
Altogether, taking into account the differences in genome size,
sperm DNA contains ~2.5-fold more background single-strand breaks
with free 3¢-OH ends and a close magnitude of extra AP sites and
possible DNA breaks with 3¢ blocking groups, than leukocyte DNA.
Besides being physiological, there is the possibility of generation of
some of these lesions in vitro, during the technical processing.
Nevertheless, working on an inert agarose matrix minimizes this
problem, stabilizing the relaxed DNA loops, avoiding both chemical
oxidation during phenol extraction and physical shearing (Legault
et al., 1997), being the base of the SCGE and DBD±FISH assays to
detect induced DNA breaks. Moreover, there should be a similar
sensitivity of nucleoids from sperm or leukocytes, to spontaneously
produced DNA lesions during the incubations, so that the observed
differences cannot be attributed to them. Both DNA breaks and AP
sites exist in a steady-state in living leukocytes, being continuously
produced and repaired. Nevertheless, the mature sperm lacks DNA
repair capacity (Robbins, 1996), and appears highly vulnerable to
radiation and some chemicals or their metabolites (Kamiguchi and
Tateno, 2002). Therefore, DNA damage could accumulate as sperm
age and/or are in contact with potential exogenous damaging agents.
Thus, the DNA repair activity of the penetrated oocyte should be of
enormous importance to correct for the background damage of human
sperm DNA (Genesca et al., 1992).
According to the DBD±FISH results where an extreme alkalisensitivity of naked sperm DNA was evident in sperm cells compared
to leukocytes, the relatively higher density of single-strand breaks and
AP sites is insuf®cient to fully explain the situation. It is evident that
these spontaneous DNA lesions should not be the only modi®cations
that confer alkali sensitivity in undamaged cells. It has been stated that
DNA in chromatin from mitotic cells is more sensitive to denaturation
than is the DNA in chromatin of interphase cells (Darzynkiewicz et al.,
1987; Darzynkiewicz, 1994). The ssDNA was detected by ¯ow
cytometry after acridine orange staining. The same group also
demonstrated that this sensitivity was related to a higher abundance
in mitotic chromosomes of DNA segments sensitive to single strandspeci®c nucleases, such us S1 and mung bean nucleases (Juan et al.,
1996). Based on these observations, it was hypothesized that the
stronger torsional tension in DNA loops of mitotic chromosomes may
lead to the formation of a higher frequency of single-stranded hairpinloop structures which are sensitive to single strand-speci®c nucleases.
These localized unpaired DNA segments may be too small to be
detectable by Acridine Orange staining or by FISH. Nevertheless, they
would behave as initiation sites from which the denaturing agents
would initiate the generation of extensive ssDNA that becomes
detectable by these procedures.
The chromatin of mature mammalian sperm is speci®cally arranged
in an extremely compact and stable organization. Based on the
doughnut loop model, the protamine-bound DNA is coiled into loops
that collapse into a toroid of tightly packed chromatin. Each toroid
would represent a single DNA loop domain of ~40±50 kb, attached at
its base to the nuclear matrix (Allen et al., 1997; Ward, 1997). As a
result, when chromatin is organized in a protamine complex, it
207
L.Muriel et al.
becomes six times more condensed than in metaphase chromosomes
(Ward and Coffey, 1991). This organization is stabilized by inter- and
intramolecular disulphide bonds in protamines, resulting in a very
rigid, crystalline-like structure (Balhorn, 1982). Moreover, the DNA
loops from sperm chromatin are smaller than those from somatic cells,
the latter with an average size of 90 kb, or even longer, so their relative
density should be higher in sperm cells (Nadel et al., 1995; Klaus et al.,
2001). In the case of somatic cells, S1-hypersensitive sites have been
mapped close to the matrix attachment regions (MAR), located at the
base of the DNA loops (Targa et al., 1994). MAR contain baseunpairing regions (BUR) that may ligate speci®c nuclear proteins
which could be involved in the establishment of the tissue-speci®c
nuclear architecture and regulation (Shanon, 2003; Cai et al., 2003).
These facts imply that the presence of localized unwound sites by
severe topological tension should be far more frequent in the
chromatin from mature sperm cells. We have compared the in situ
end labelling of the DNA breaks generated by mung bean nuclease in
nucleoids from sperm and leukocytes (Figure 1m±p). Mung bean
nuclease was chosen because, unlike S1 nuclease, this enzyme will not
cleave the DNA strand opposite a nick, so it is more speci®c of
ssDNA. After subtraction of the signal corresponding to Klenow
labelling alone, i.e. the component due to background DNA breaks, it
was found that the nucleoids of sperm cells contain 9-fold higher
density of short ssDNA stretches than the nucleoids from leukocytes
(Figure 1 m±p; Table I). This supports the great abundance of partially
denatured ssDNA segments as the main cause of the strong alkali
sensitivity of sperm DNA. In the case of histone-organized chromatin,
these ssDNA segments had been detected by nuclease digestion,
without previous protein extraction (Juan et al., 1996). In our case, it is
dif®cult to establish their presence in intact sperm nuclei, since their
compaction level restricts the accessibility of nucleases in comparison
with histone-organized chromatin, and it should be borne in mind that
the alkali sensitivity of sperm DNA was only detected after protein
removal. Nevertheless, it is interesting to note that the DNA segment
sensitive to S1 nuclease in organized chromatin is also sensitive as
naked DNA, when it is stressed in recombinant DNA plasmids (Larsen
and Weintraub, 1982). In any case, besides AP sites, the presence of
partially denatured DNA segments should be considered as genuine
alkali-labile sites, either spontaneous or induced by DNA damaging
agents, and be taken into consideration in all the procedures that
employ an alkaline denaturation to detect and quantify DNA breakage.
Finally, since ssDNA is known to be much more sensitive than doublestranded DNA to several types of DNA damage (Shapiro, 1981;
Legault et al., 1997), the higher density of unpaired DNA segments
could explain the relatively higher frequency of background DNA
damage and sensitivity to several mutagens, in mature sperm. It may
be useful to evaluate the behaviour of germ cells from different stages
of maturation, using our experimental approach. Since remarkable
changes in chromatin structure and organization are apparent, this
could result in different responses to the unwinding treatment.
Acknowledgements
We are very grateful to Barbara Hamkalo (Department of Molecular Biology
and Biochemistry, University of California, Irvine) for the critical reading of
the manuscript. This work was supported by the Fondo de Investigaciones
Sanitarias (FIS 01-3113), Plan Gallego de InvestigacioÂn y Desarrollo
TecnoloÂgico de la Xunta de Galicia (PGIDIT-01BIO02E) and PGIDIT
02PXIC91603PN, and the Consejo de Seguridad Nuclear from Spain.
References
Allen MJ, Bradbury EM and Balhorn R (1997) AFM analysis of DNA
protamine complexes bound to mica. Nucleic Acids Res 25,2221±2226.
208
Atamna H, Cheung I and Ames BN (2000) A method for detecting abasic sites
in living cells: age dependent changes in base excision repair. Proc Natl
Acad Sci USA 97,686±691.
Balhorn R (1982) A model for the structure of chromatin in mammalian
sperm. J Cell Biol 93,298±305.
Cai S, Han H and Kohwi-Shigematsu K (2003) Tissue-speci®c nuclear
architecture and gene expresion regulated by SATB1. Nature Genet 34,42±
51.
Darzynkiewicz Z (1994) Acid-induced denaturation of DNA in situ as a probe
of chromatin structure. Methods Cell Biol 41,527±541.
Darzynkiewicz Z, Traganos F, Carter SP and Higgins PJ (1987) In situ factors
affecting stability of the DNA helix in interphase nuclei and metaphase
chromosomes. Exp Cell Res 172,168±179.
Dierendonck JH (2002) DNA damage detection using DNA polymerase I or its
Klenow fragment: applicability, speci®city, limitations. Methods Mol Biol
203,81±108.
Evenson DP, Jost LK, Marshall D, Zinaman MJ, Clegg E, Purvis K, de
Angelis P and Claussen OP (1999) Utility of the sperm chromatin structure
assay as a diagnostic and prognostic tool in the human fertility clinic. Hum
Reprod 14,1039±1049.
Fairbain DW, Reyes WA and O'Neill KL (1994) Alkali-labile sites are
prevalent in kidney tissue DNA. Cancer Lett 81,67±76.
FernaÂndez JL and GosaÂlvez J (2002) Application of FISH to detect DNA
damage: DNA Breakage Detection±FISH (DBD±FISH). Methods Mol Biol
203,203±216.
FernaÂndez JL, Goyanes VJ, Ramiro-DõÂaz J and GosaÂlvez J (1998) Application
of FISH for in situ detection and quanti®cation of DNA breakage.
Cytogenet Cell Genet 82,251±256.
FernaÂndez JL, VaÂzquez-GundõÂn F, Delgado A, Goyanes VJ, Ramiro-DõÂaz J, de
la Torre J and GosaÂlvez J (2000) DNA breakage detection±FISH (DBD±
FISH) in human spermatozoa: technical variants evidence different
structural features. Mutat Res 453,77±82.
FernaÂndez JL, Goyanes V and GosaÂlvez J (2002) DNA Breakage DetectionFISH (DBD±FISH). In Rautenstrauss B and Liehr T (eds) FISH
TechnologyÐSpringer Laboratory Manual. Springer-Verlag, Heidelberg,
pp. 282±290.
Genesca A, Barrios L, Miro R, CaballõÂn MR, Benet J, Fuster C, Bon®ll X and
Egozcue J (1990) Lymphocyte and sperm chromosome studies in cancertreated men. Hum Genet 84,353±355.
Genesca A, CaballõÂn MR, Miro R, Benet J, Germa JR and Egozcue J (1992)
Repair of human sperm chromosome aberrations in the hamster egg. Hum
Genet 82,181±186.
GosaÂlvez J, de la Torre J, LoÂpez-FernaÂndez C and FernaÂndez JL (2002) The
combined use of nuclease in situ digestion and FISH (NU-FISH). In
Rautenstrauss B and Liehr T (eds) FISH TechnologyÐSpringer Laboratory
Manual. Springer-Verlag, Heidelberg, pp 272±281.
Haines G, Marples B, Daniel P and Morris I (1998) DNA damage in human
and mouse spermatozoa after in vitro-irradiation assessed by the comet
assay. In del Mazo J (ed) Reproductive Toxicology, Plenum Press, New
York, pp 79±93.
Juan G, Pan W and Darzynkiewicz Z (1996) DNA segments sensitive to
single-strand-speci®c nucleases are present in chromatin of mitotic cells.
Exp Cell Res 227,197±202.
Kamiguchi Y, and Tateno H (2002) Radiation- and chemical-induced
structural chromosome aberrations in human spermatozoa. Mutat Res
504,183±191.
Klaus AV, McCarrey JR, Farkas A and Ward WS (2001) Changes in DNA
loop domain structure during spermatogenesis and embryogenesis in the
Syrian golden hamster. Biol Reprod 64,1297±1306.
Larsen A and Weintraub H (1982) An altered DNA conformation detected by
S1 nuclease occurs at speci®c regions in active chick globin chromatin. Cell
29,609±622.
Larson KL, De Jonge C, Barnes A, Jost L and Evenson DP (2000)
Relationship between assisted reproductive techniques (ART) outcome
and status of chromatin integrity as measured by the sperm chromatin
structure assay (SCSA). Hum Reprod 15,1717±1722.
Legault J, Tremblay A, Ramotar D and Mirault M (1997) Clusters of S1
nuclease-hypersensitive sites induced in vivo by DNA damage. Mol Cell
Biol 17,5437±5452.
Lindahl T and KarlstroÈm O (1973) Heat-induced depyrimidination of
deoxyribonucleic acid in neutral solution. Biochemistry 12,5151±5154.
Liu PK, Cui J, Moore N and Huang D (2002) The in situ detection of apurinic/
apyrimidinic sites and DNA breaks bearing extension blocking termini.
Methods Mol Biol 203,235±244.
Structure and background damage of human sperm DNA
Martin RH, Rademaker A, Hildebrand K, Barnes M, Arthur K, Ringrose T,
Brown IS and Douglas G (1989) A comparison of chromosomal aberrations
induced by in vivo radiotherapy in human sperm and lymphocytes. Mutat
Res 226,21±30.
Nadel B, de Lara J, Finkernagel SW and Ward WS (1995) Cell-speci®c
organization of the 5S ribosomal RNA gene cluster DNA loop domains in
spermatozoa and somatic cells. Biol Reprod 53,1222±1228.
Robbins WA (1996) Cytogenetic damage measured in human sperm following
cancer chemotherapy. Mutat Res 355,235±252.
Rydberg B (1975) The rate of strand separation in alkali of DNA of irradiated
mammalian cells. Radiat Res 61,274±287.
Shanon MF (2003) A nuclear address with in¯uence. Nature Genet 34,4±6.
Shapiro R (1981) Damage to DNA Caused by Hydrolysis. Plenum Publishing
Corp, New York.
Singh NP, Danner DB, Tice RR, McCoy MT, Collins GD and Schneider EL
(1989) Abundant alkali-sensitive sites DNA of human and mouse sperm.
Exp Cell Res 184,461±470.
Sloter ED, Lowe X, Moore II, DH, Nath J and Wyrobek AJ (2000) Multicolor
FISH analysis of chromosomal breaks, duplications, deletions, and
numerical abnormalities in the sperm of healthy men. Am J Hum Genet
67,862±872.
Targa FR, Razin SV, de Moura Gallo CV and Scherrer K (1994) Excision
close to matrix attachment regions of the entire chicken alpha-globin gene
domain by nuclease S1 and characterization of the framing structures. Proc
Natl Acad Sci USA 91,4422±4426.
TeÂoule R (1987) Radiation-induced DNA damage and repair. Int J Radiat Biol
51, 573±589.
Tomar D, Magenis E, Chamberlin J, Allen L, Olson S, Donlon T, Barton S,
Sheehy R and Waggoner D (1984) Preferential paternal origin of de novo
structural chromosome rearrangements. Am J Hum Genet 36,115S.
VanLoon AAWN, Den Boer PJ, Van der Schans GP, Mackenbach P,
Grootegoed JA, Baan RA and Lohman PHM (1991) Immunochemical
detection of DNA damage induction and repair at different cellular stages of
spermatogenesis of the hamster after in vitro or in vivo exposure to ionizing
radiation. Exp Cell Res 193,303±309.
VanLoon AAWN, Sonneveld E, Hoogerbrugge J, Van der Schans GP,
Grootegoed JA, Lohman PHM and Baan RA (1993) Induction and repair of
DNA single-strand breaks and DNA base damage at different cellular stages
of spermatogenesis of the hamster upon in vitro exposure to ionizing
radiation. Mutat Res 294,139±148.
vonSonntag C (1987) The Chemical Basis of Radiation Biology. Taylor &
Francis, London, pp 210±213.
Walker PR, Carson C, Leblanc J and Sikorska M (2002) Labelling DNA
damage with terminal transferase: applicability, speci®city, limitations.
Methods Mol Biol 203,3±19.
Ward WS (1997) Chromosome organization in mammalian sperm nuclei. In
Barratt C, de Jonge C, Mortimer D and Parinaud J (eds) Genetics of Human
Male Infertility. Editions EDK, Paris, pp 205±221.
Ward WS and Coffey DS (1991) DNA packaging and organization in
mammalian spermatozoa: comparison with somatic cells. Biol Reprod
44,569±574.
Wood KA (2002) Labelling DNA breaks in situ by Klenow enzyme. Methods
Mol Biol 203,109±119.
Zalensky AO, Allen MJ, Kobayashy A, Zalenskaya IA, Balhorn R and
Bradbury EM (1995) Well-de®ned genome architecture in human sperm
nucleus. Chromosoma 103,577±590.
Submitted on September 4, 2003; resubmitted on November 18, 2003;
accepted on November 20, 2003
209