DNA fragmentation during apoptosis is caused by frequent single-strand cuts.

One of the hallmarks of apoptosis is the digestion of genomic DNA by an endonuclease, generating a ladder of small fragments of double-stranded DNA. We have examined the nature of the DNA breaks produced in mouse thymocytes triggered to undergo apoptosis by steroids or by stimulation of the T cell receptor. Whereas the typical ladder pattern of oligonucleosomal fragments was observed after agarose gel electrophoresis, numerous single-strand cuts were detected after electrophoresis under denaturing conditions. Single-strand nicks were found to be very frequent in the internucleosomal regions, but also to occur in the core particle-associated DNA. An identical pattern of single-strand nicks was obtained when chromatin DNA was exposed to the single-strand cleaving deoxyribonuclease I. The nicked DNA fragments, extracted from apoptotic thymocytes, were sensitive to the action of S1-nuclease. We propose that DNA fragmentation induced during apoptosis is not due to a double-strand cutting enzyme as previously postulated, but rather is the result of single-strand breaks. This ensures the dissociation of the DNA molecule at sites where cuts are found within close proximity.     

Detection of DNA fragmentation and endonucleases in apoptosis

DNA degradation during apoptosis is endonuclease mediated and proceeds through an ordered series of stages commencing with the production of large DNA pieces of 300 kb which are then degraded to fragments of 50 kb. The 50-kb fragments are further degraded, in some but not all cells, to smaller pieces (10-40 kb) releasing the small oligonucleosome fragments that are detected as a characteristic DNA ladder on conventional agarose gels. Methodology is presented for the detection of both DNA ladders and the initial stages of DNA fragmentation using pulsed-field gel electrophoresis. We have developed electrophoresis conditions that resolve large fragments of DNA and also retain the smaller fragments on the same gel. Methods for the detection of endonuclease activities responsible for the cleavage of DNA during apoptosis are also presented.     

DNA loop organization and DNA fragmentation during radiation-induced apoptosis in human lymphocytes

Apoptotic DNA fragmentation induced by gamma-rays has been compared with the DNA loop sizes in G0-human lymphocytes using pulsed field gel electrophoresis (PFGE). Genomic DNA was cleaved into the DNA loops at the topoisomerase II mediated attachment points using short treatment of cells with etoposide. The apoptotic fragmentation, with a distinct cut-off around 50 kb for a maximum length of fragments, appeared 5 h after irradiation when the most part of radiation-induced DNA double strand breaks (DSBs) have been repaired. The data indicate that apoptotic fragmentation of DNA in the G0-human lymphocytes begins when repair of radiation-induced DSBs has been completed. Similar apoptotic DNA fragmentation was also observed following the treatment of cells with etoposide. All genomic DNA was fragmented into 50-kb fragments during the final stages of apoptosis. Most of the DNA in resting lymphocytes is organized into Mb-size loops but loops of sizes down to 50 kb were also observed. A sharp border between the size distributions of DNA loops and apoptotic fragments was found. The data suggest that 50 kb apoptotic fragmentation is not based on excision of the DNA loops. No apoptotic fragments with the sizes more than 5.7 Mb were seen during the whole course of apoptosis. This observation indicates that despite intensive apoptotic fragmentation into the 50-kb fragments the chromosomes maintain integrity during radiation-induced apoptosis in human lymphocytes. We propose a model for radiation-induced apoptotic fragmentation in human lymphocytes that involves four stages: induction of DNA breaks and relaxation of DNA loops; DNA repair followed by reorganization of the DNA loops into the 50-kb units of condensed chromatin; co-operative fragmentation of the reorganized DNA loops into the distinct 50-kb fragments and resealing of the chromosome ends at the sites of this fragmentation; cleavage of the 50-kb fragments at the internucleosomal spacers.     

Determination of microbial genome sizes by two-dimensional denaturing gradient gel electrophoresis.

In two-dimensional denaturing gradient gel electrophoresis, DNA is digested with a restriction endonuclease and the resulting DNA fragments are separated as a function of size by conventional agarose gel electrophoresis. Following this first dimension electrophoresis, the fragment distribution is placed at the top of a denaturing gradient slab gel and electrophoresis is carried out parallel to the gradient direction. This second dimension separation is a complex function of the base sequence of each fragment. Analysis of the DNA fragment distribution as a function of fragment size allows the DNA size to be calculated. This method has been applied to calculate three microbial genome sizes: Mycoplasma capricolum, 724 kb; Acholeplasma laidlawii, 1646 kb; and Hemophilus influenzae, 1833 kb.     

Influence of DNA conformation on radiation-induced single-strand breaks

It is usually assumed that sparsely ionizing radiation produces randomly distributed DNA breakages. This seems to be supported by the finding that in some DNA fragments single-strand scissions occur uniformly at all nucleotide sites, regardless of sequence. We performed experiments on two DNA fragments of about 300 by having different conformation to test whether radiation-induced single-strand breakage is dependent on DNA conformation. Breakage analysis was carried out by denaturing polyacrylamide gel electrophoresis, which allows determination of the broken site at single nucleotide resolution. We found uniform cutting patterns in B-form regions. On the contrary, X- or-irradiation of curved fragments of kinetoplast DNA showed that the distribution of single-strand breaks was not uniform along the fragment, as the cleavage pattern was modulated in phase with the runs of A-T pairs. This modulation likely reflected the reduced accessibility of the sites which on hydroxyl-radical attack give rise to strand breaks. The cleavage pattern was phased with the runs of A-T pairs. Moreover, the overall yield of strand breaks was considerably lower in curved DNA fragments than in those with extended straight regions. The conformation effect found here indicates that the cleavage pattern reflects the fine structural features of DNA.     

Ca/Mg-dependent endonuclease from porcine liver. Purification, properties, and sequence specificity

A rapid and reproducible method for the purification of the Ca/Mg-endonuclease from porcine and rat liver and for the stabilization of the enzyme activity is presented. The optimum conditions for enzyme activity were determined. The enzyme degrades double-stranded DNA endonucleolytically. In the course of digestion of form I closed circular SV 40 DNA, the form II nicked circular DNA is the prominent intermediate product. Digestion of hen erythrocyte nuclei with added endonuclease produces a ladder of mono- and oligonucleosomal fragments similar to that generated by micrococcal nuclease digestion. Determination of the 5'-terminal nucleotides indicates the absence of a significant base specificity. Analyzing the cleavage pattern of end-labeled pBR322 restriction fragments on sequencing gels shows that the enzyme exhibits a weak preference for dinucleotides with A in the 5'-position; dinucleotides with 5%-C are less readily cleaved. Digestion of end-labeled pBR322 DNA, followed by electrophoresis in agarose gels produces a "smear"-like fragmentation pattern with weak superimposed bands, while micrococcal nuclease generates a different and much more distinct pattern. These data show that the sequence specificity of the enzyme is less pronounced than that of micrococcal nuclease and that the sites preferentially cleaved are not the same.     

Chromatin Collapse during Caspase-dependent Apoptotic Cell Death Requires DNA Fragmentation Factor, 40-kDa Subunit-/Caspase-activated Deoxyribonuclease-mediated 3′-OH Single-strand DNA Breaks

Apoptotic nuclear morphology and oligonucleosomal double-strand DNA fragments (also known as DNA ladder) are considered the hallmarks of apoptotic cell death. From a classic point of view, these two processes occur concomitantly. Once activated, DNA fragmentation factor, 40-kDa subunit (DFF40)/caspase-activated DNase (CAD) endonuclease hydrolyzes the DNA into oligonucleosomal-size pieces, facilitating the chromatin package. However, the dogma that the apoptotic nuclear morphology depends on DNA fragmentation has been questioned. Here, we use different cellular models, including MEF CAD^−/− cells, to unravel the mechanism by which DFF40/CAD influences chromatin condensation and nuclear collapse during apoptosis. Upon apoptotic insult, SK-N-AS cells display caspase-dependent apoptotic nuclear alterations in the absence of internucleosomal DNA degradation. The overexpression of a wild-type form of DFF40/CAD endonuclease, but not of different catalytic-null mutants, restores the cellular ability to degrade the chromatin into oligonucleosomal-length fragments. We show that apoptotic nuclear collapse requires a 3′-OH endonucleolytic activity even though the internucleosomal DNA degradation is impaired. Moreover, alkaline unwinding electrophoresis and In Situ End-Labeling (ISEL)/In Situ Nick Translation (ISNT) assays reveal that the apoptotic DNA damage observed in the DNA ladder-deficient SK-N-AS cells is characterized by the presence of single-strand nicks/breaks. Apoptotic single-strand breaks can be impaired by DFF40/CAD knockdown, abrogating nuclear collapse and disassembly. In conclusion, the highest order of chromatin compaction observed in the later steps of caspase-dependent apoptosis relies on DFF40/CAD-mediated DNA damage by generating 3′-OH ends in single-strand rather than double-strand DNA nicks/breaks.     

The role of topoisomerase II in the excision of DNA loop domains during apoptosis

Disintegration of nuclear DNA into high molecular weight (HMW) and oligonucleosomal DNA fragments represents two major periodicities of DNA fragmentation during apoptosis. These are thought to originate from the excision of DNA loop domains and from the cleavage of nuclear DNA at the internucleosomal positions, respectively. In this report, we demonstrate that different apoptotic insults induced apoptosis in NB-2a neuroblastoma cells that was invariably accompanied by the formation of HMW DNA fragments of about 50-100 kb but proceeded either with or without oligonucleosomal DNA cleavage, depending on the type of apoptotic inducer. We demonstrate that differences in the pattern of DNA fragmentation were reproducible in a cell-free apoptotic system and develop conditions that allow in vitro separation of the HMW and oligonucleosomal DNA fragmentation activities. In contrast to apoptosis associated with oligonucleosomal DNA fragmentation, the HMW DNA cleavage in apoptotic cells was accompanied by down-regulation of caspase-activated DNase (CAD) and was not affected by z-VAD-fmk, suggesting that the caspase/CAD pathway is not involved in the excision of DNA loop domains. We further demonstrate that nonapoptotic NB-2a cells contain a constitutively present nuclease activity located in the nuclear matrix fraction that possessed the properties of topoisomerase (topo) II and was capable of reproducing the pattern of HMW DNA cleavage that occurred in apoptotic cells. We demonstrate that the early stages of apoptosis induced by different stimuli were accompanied by activation of topo II-mediated HMW DNA cleavage that was reversible after removal of apoptotic inducers, and we present evidence of the involvement of topo II in the formation of HMW DNA fragments at the advanced stages of apoptosis. The results suggest that topo II is involved in caspase-independent excision of DNA loop domains during apoptosis, and this represents an alternative pathway of apoptotic DNA disintegration from CAD-driven caspase-dependent oligonucleosomal DNA cleavage.     

Site specificity of bleomycin-mediated single-strand scissions and alkali-labile damage in duplex DNA.

Form II PM2 DNA, which contained bleomycin-mediated single-strand breaks, was purified and treated with the extracellular endonuclease from Alteromonas BAL 31. This enzyme cleaves the phosphodiester backbone opposite a single-strand break to yield a double-strand break. The locations of these double-strand breaks were determined relative to the cleavage sites produced by the restriction enzyme HindIII. The experimental procedure was as follows. Form I PM2 DNA was treated with bleomycin to produce alkali-labile bonds. These were hydrolyzed by alkali treatment and the DNA, now containing single-strand breaks, was purified and treated with the BAL 31 enzyme and the HindIII enzyme to determine the positions of the original alkali-labile bonds. It was found that the single-strand breaks and alkali-labile bonds were introduced at preferred sites on the PM2 genome, since electrophoretic analyses of the DNA after the HindIII digestion revealed DNA bands of discrete sizes. The molecular weights of the DNA fragments produced by these treatments indicate that single-strand breaks and alkali-labile bonds occur at the same sites as those previously determined for direct double-strand scissions introduced by bleomycin at neutral pH. Some of the specific sites of double-strand scissions mediated by bleomycin at neutral pH (Lloyd et al., 1978b) are also shown here to be relatively more reactive than other sites when the DNA contains superhelical turns.     

Characterization of glyoxalase I purified from pig erythrocytes by affinity chromatography

The linear order of nine fragments generated by the action of endonuclease AvaI on the DNA of bacteriophage lambda was determined from the altered fragmentation patterns of bacteriophages containing known deletions and of hybrids of bacteriophages lambda and phi80. Digestion of 5'-terminally 32P-labelled bacteriophage-lambda DNA was used to identify the terminal fragments. Measurement of relative fragment lengths permitted rough mapping of the endonuclease-AvaI cleavage sites relative to the ends of the bacteriophage-lambda chromosome. The fragment order was confirmed and the map refined by analysis of the fragmentation of derivative phages containing single cleavage sites for endonuclease EcoRI.     

Hydrogen peroxide induces rapid digestion of oligodendrocyte chromatin into high molecular weight fragments

High molecular weight (HMW) fragmentation of nuclear chromatin was studied in cultured rat oligodendrocytes (OL) exposed to hydrogen peroxide (H2O2). Intact genomic DNA was isolated by agarose embedding, and analyzed by field inversion gel electrophoresis, with and without S1 endonuclease digestion to detect and discriminate between single and double stranded fragmentations, respectively. The exposure of OL to H2O2 resulted in a very rapid degradation of chromosomal DNA into HMW fragments that reflect native chromatin structure. Hence, within 10 min after the addition of 1 mM H2O2, a discrete pool representing approximately 45% of the nuclear chromatin underwent single strand digestion into >400 kb fragments likely at AT-rich matrix attachment regions. Subsequent accumulation of single stand breaks at these regions led to bifilar scission. Ultimately, chromatin within this susceptible pool was cleaved at remaining matrix attachment regions into 50-200 kb fragments. Chromatin digestion could be elicited with H2O2 concentrations as low as 50 microM. After the removal of H2O2, most >400 kb fragments were religated within 2 h; however, digestion into 50-200 kb fragments was irreversible. The DNA digestion was not accompanied by the degradation of nuclear proteins, i.e., lamins A/C and poly (ADP-ribose) polymerase indicating that chromatin fragmentation is unlikely to be mediated by proteolysis. In conclusion, H2O2 at pathologically relevant concentrations induces a very rapid and extensive digestion of OL chromatin into HMW fragments. Because the chromatin fragmentation is only partly reversible, it may be a decisive factor in committing oxidatively stressed OL to degeneration and/or death.     

Downregulation of Bcr-Abl in K562 cells restores susceptibility to apoptosis: characterization of the apoptotic death

We examined the susceptibility of a variety of human leukemic cell lines to the induction of apoptosis. K562, a chronic myelogenous leukemic cell line which expresses the bcr-abl fusion gene, was found to be extremely resistant to apoptosis, irrespective of the inducing agent. This resistance can be attributed to the deregulated Abl kinase activity of bcr-abl, as downregulation of its expression using antisense oligodeoxynucleotides targeted to the beginning of the abl sequence in this chimeric gene rendered these cells susceptible to cytotoxic drug-induced apoptosis. Examination of the morphological and biochemical features of apoptosis in K562 cells revealed the typical membrane blebbing and chromatin condensation associated with this form of cell death. In situ TdT-mediated end labeling of the DNA revealed the presence of strand breaks in the treated cells and field inversion gel electrophoresis revealed the presence of large 10-50 kb fragments. However there was an absence of oligonucleosomal DNA fragmentation, whether or not Bcr-Abl was expressed. Thus, while inhibition of expression of Bcr-Abl renders K562 cells susceptible to apoptosis, the absence of oligonucleosomal DNA fragmentation in these cells is independent of the function of this molecule.     

Role of DNAS1L3 in Ca2+- and Mg2+-dependent cleavage of DNA into oligonucleosomal and high molecular mass fragments

Ca2+- and Mg2+-dependent endonucleases have been implicated in DNA fragmentation during apoptosis. We have demonstrated that particular nucleases of this type are inhibited by poly(ADP-ribosyl)ation and suggested that subsequent cleavage of PARP by caspase-3 might release these nucleases from poly(ADP-ribosyl)ation-induced inhibition. Hence, we purified and partially sequenced such a nuclease isolated from bovine seminal plasma and identified human, rat and mouse homologs of this enzyme. The extent of sequence homology among these nucleases indicates that these four proteins are orthologous members of the family of DNase I-related enzymes. We demonstrate that the activation of the human homolog previously specified as DNAS1L3 can induce Ca2+- and Mg2+-dependent DNA fragmentation in vitro and in vivo. RT-PCR analysis failed to detect DNAS1L3 mRNA in HeLa cells and nuclei isolated from these cells did not exhibit internucleosomal DNA fragmentation when incubated in the presence of Ca2+and Mg2+. However, nuclei isolated from HeLa cells that had been stably transfected with DNAS1L3 cDNA underwent such DNA fragmentation in the presence of both ions. The Ca2+ionophore ionomycin also induced internucleosomal DNA degradation in transfected but not in control HeLa cells. Transverse alternating field electrophoresis revealed that in nuclei from transfected HeLa cells, but not in those from control cells, DNA was cleaved into fragments of >1000 kb in the presence of Mg2+; addition of Ca2+in the presence of Mg2+resulted in processing of the >1000 kb fragments into 50 kb and oligonucleosomal fragments. These results demonstrate that DNAS1L3 is necessary for Ca2+- and Mg2+-dependent cleavage of DNA into both oligonucleosomal and high molecular mass fragments in specific cell types.     

Specificity and mode of cleavage of the pH 4.0 endonuclease from adenovirus type 2 - infected KB cells.

Adenovirus type 2 or lambda DNA was digested with the pH 4.0 endonuclease, purified from adenovirus 2-infected KB cells. The enzyme produces a limit digest of approximate size in the range of 140-210 base pairs long. The termini of the DNA fragments generated by the endonuclease digestion had 3'-P and 5'-OH groups. The 3' and 5' end groups of the products were analyzed. Our data indicate that 3' end group was a purine (68-76%), dA occuring about twice the frequency of dG. The 5' end group was either dG or dC with equal frequency. Data obtained by treatment of the 5' labeled endonuclease product of lambda DNA with single-strand specific S1 nuclease from Asperigillus oryzae or exonuclease VII from Escherichia coli indicated that the majority of the products had a short 5' protruding ends. The mode of cleavage of this endonuclease seems to be through initial formation of several single-strand breaks with some base specificity. If these breaks are at close proximity on opposite strands, double-stranded fragments with protruding ends are generated.     

Sulindac activates nuclear translocation of AIF, DFF40 and endonuclease G but not induces oligonucleosomal DNA fragmentation in HT-29 cells

Sulindac is one of the most widely studied nonsteroidal anti-inflammatory drugs in the prevention of colon cancer. Thus, from the viewpoint of colon cancer chemotherapy it is important to reveal the mechanism of sulindac-induced cell death. This study was undertaken to dissect the molecular mechanism underlying sulindac-induced apoptosis in human colon cancer cell line HT-29 (mutant p53), focusing on nuclear translocation of AIF, DFF and endonuclease G. On induction of apoptosis by sulindac, it was associated with decreased mitochondrial membrane potential, nuclear expression of active caspase-3, cleavage of poly(ADP-ribose) polymerase, translocation of mitochondrial proteins to the nucleus, and morphological evidence of nuclear condensation. However, sulindac led to only disintegration of nuclear DNA into high molecular weight DNA fragments of about 100-300 kbp as determined by a pulse-field gel electrophoresis, suggesting a predominantly AIF-mediated cell death process. In summary, our findings indicate that sulindac induces large-scale DNA fragmentation without oligonucleosomal DNA fragmentation. This result suggests that nuclear translocation of DFF and endonuclease G are not sufficient for the induction of oligonucleosomal DNA fragmentation in HT-29 cells.     

Chromatin isolated from viable human PBLs contains DNA fragmented to >/=50 kb

Massive chromatin fragmentation (around 50 kb, to several hundred kb) is observed in nuclear lysates of human peripheral blood lymphocytes (PBLs) upon their treatment with nuclease-free protein-denaturants. There is a consistent variation in the fragment size distributions that parallels the proliferative activity of the cells. Predominantly approximately 50 kb fragmentation is exhibited in samples from cells immunsuppressed via CD4 crosslinking, as opposed to the heterogeneous, higher molecular weight DNA of anti-TcR/CD3-, phytohemagglutininor concanavalin A-stimulated cells. Tritiated thymidine incorporated into DNA in the latter cultures can be detected in the approximately 50 kb band. Direct lysis of agarose-embedded, live cells in alkali+detergent also yields fragmented DNA, with a single-strand size of >/=50 kb. These data suggest that (i) the cells yielding fragmented DNA were alive at the time of DNA extraction, (ii) either regularly arranged, preformed nicks or hypersensitive sites may be present at every roughly 50-100 kb in the chromatin of PBLs, (iii) these sites or the fragmentation mechanism acting upon them, appear to be regulated in concert with the transit of cells between the resting and proliferative compartments.     

Cloning of the DNA BglII fragments of bacteriophage T4 in the vector plasmid pSCC31

An attempt has been made to clone six BglII fragments of T4 DNA in the range of 3.3-8.1 kb in the vector plasmid pSCC31 containing a single BglII site within the gene for endonuclease EcoRI and pL promoter of phage lambda. DNA fragments were extracted from the corresponding bands of agarose gel. The following BglII fragments were cloned: the 3.3 kb fragment No. 9 containing a portion of gene 20, the gene 21 and a portion of gene 22; the 4.2 kb fragment No. 8.1 with genes 17, 18, 19 and a portion of gene 20; the 5.2 kb fragment No. 7.1 with genes 25-29 and a portion of gene 48. In the case of the fragment No. 7.1, the recombinant plasmids pRL705 and pRL707 with different orientation of phage DNA fragment were obtained. An attempt to clone the fragments No. 8.2 (4.2 kb), No. 7.2 (5.45 kb) and No. 6 (8.1 kb) was unsuccessful and this probably indicates the presence of the genes, whose products are deleterious to the growth of bacterial cell.     

Lack of internucleosomal DNA fragmentation is related to Cl(-) efflux impairment in hematopoietic cell apoptosis.

The heterodimeric DNA fragmentation factor (DFF) is responsible for DNA degradation into nucleosomal units during apoptosis. This process needs the caspase-dependent release of ICAD/DFF-45, the inhibitory subunit of DFF. Here we report that triggering apoptosis via a hyperosmotic shock in hematopoietic cells causes the appearance of mitochondrial and cytosolic alterations, activation of caspases, chromatin condensation, nuclear disruption, and DNA fragmentation. However, oligonucleosomal but not high molecular weight (50-150 kb) DNA cleavage is abolished if Cl(-) efflux is prevented by using NaCl to raise extracellular osmolarity or by Cl(-) channel blockers, even when apoptosis is initiated by other agents (staurosporine, anti-Fas antibody). In these conditions, all the apoptosis hallmarks investigated remain detectable, including the cleavage of ICAD/DFF-45. In vitro assays with lysates of cells in which Cl(-) efflux is blocked confirm the lack of internucleosomal DNA degradation. These findings establish that neither caspase activation nor ICAD/DFF-45 processing per se is sufficient to induce oligonucleosomal DNA fragmentation and that high molecular weight DNA degradation and chromatin condensation appear independently of it. Finally, they suggest that Cl(-) efflux is a necessary cofactor that intervenes specifically in the activation of the DFF endonuclease.     

Electrophoretic DNA analysis for the detection of apoptosis

There are many techniques available for the detection of apoptotic cells; some are based on morphological changes, others on biochemical events. However, electrophoretic detection of the systematic cleavage of DNA into oligonucleosomal multimers of 180–200 bp remains the “hallmark” of apoptosis. Conventional constant field agarose gel electrophoresis of DNA from apoptotic cells can be used to resolve the multimers into the characteristic DNA ladders indicative of apoptotic cell death. More recently, it has become clear that the generation of the lower molecular weight oligonucleosomal DNA is preceded by the generation of higher molecular weight fragments. In some cell types, DNA cleavage proceeds no further than the formation of 300 and/or 50 kbp cleavage products. DNA fragmentation of this size can only be resolved using a form of pulsed-field gel electrophoresis. Basic “starter” protocols for conventional and pulsed field electrophoresis for the detection of apoptotic cell DNA are described in this article.     

Control of partial digestion combining the enzymes dam methylase and MboI.

A method is described which allows the preparation of reproducible partial digests without previous establishment of the incubation conditions. It is based on a combined application of dam methylase and the restriction endonuclease MboI, both recognizing the sequence 5'-GATC-3' but MboI unable to cut the methylated site. Due to their competition for the same substrate the DNA is partially digested, with the size of the resulting fragments strongly dependent on the ratio of enzymes. The Km of the dam methylase was determined to be 115 ng DNA/microliters indicating a variance in fragment sizes generated at low DNA-concentrations. This effect is minimized above 150 ng/microliters. Any influence of digestion time is avoided, because the reaction runs until complete modification of all sites. The dependence on enzyme concentration and presence of agarose was checked. Knowledge of these parameters allows an accurate prediction of fragment sizes generated at different conditions. The technique was successfully used to construct libraries from different sources, in particular chromosome-specific libraries from small amounts of flow-sorted material.