Unfixed and fixed human chromosomes show different staining patterns after restriction endonuclease digestion

Unfixed and fixed human chromosomes show different staining patterns after restriction endonuclease digestion. Restriction endonucleases (REs) have been widely used to produce banding patterns on chromosomes, but it remains uncertain to what extent the patterns are due to the sequence specificity of the enzymes, and to what extent chromatin structure influences the pattern of digestion. To throw light on this question, we have digested with restriction endonucleases unfixed chromosomes prepared in two different ways (iso-lated, and whole metaphase cells spread with a cytocentrifuge) and compared the results with those obtained on conventionally fixed chromosomes. Unfixed isolated chromosomes are easily destroyed by REs; after fixation with cold methanol, which produced minimal alteration to the chromatin structure, the chromo- somes are resistant to the action of REs, and conventional methanol-acetic acid fixation is required to permit the induction of banding patterns by REs. Unfixed cytocentrifuge preparations, in which the chromosomes are still surrounded by cytoplasm, are much more resistant to the action of REs, and again banding patterns were only induced after methanol-acetic acid fixation. We conclude that the action of restriction endonucleases on chromosomes is strongly influenced by chromatin organisation, and that methanol-acetic acid fixation is required to permit the induction of conventional banding patterns on chro- mosomes.

The possibility of using restriction endonucleases (REs) to induce banding patterns in metaphase chromosomes has resulted in the widespread use of these enzymes in cytogenetics in recent years (MEZZA-NOTE et al. 1983a, b, 1985MILLER et al. 1983;KAELBLING et al. 1984;BIANCHI et al. 1985;BABU and VERMA 1986;MEZZANOTTE 1986;FERRUCCI et al. 1987;MARCHI and MEZZANOTTE 1988;BURK-HOLDER 1989). Certain REs produce bands whose location coincides with that of certain highly repetitive DNA fractions, while others induce linear differentiation generally resembling G-banding patterns (MEZZANOTTE et al. 1983a(MEZZANOTTE et al. , b, 1985MILLER et al. 1983).
The mechanism and significance of RE-induced bands are still under investigation. Some authors claim that the DNA in all regions of chromosomes is available for digestion with REs (MILLER et al. 1983) and that "chromatin structure is not the main cause underlying chromosome banding" produced by such enzymes (BIANCHI et al. 1985). On the other hand, it is clear that although REs can act only by cleaving specific DNA sequences, their effects on chromosomes could potentially be limited, firstly by the accessibility of the chromosomal DNA to the enzymes, and secondly by the ability of the cleaved DNA to diffuse out of the chromosomes. Clearly chromatin structure could be an important factor in determining the patterns of banding produced by digesting chromosomes with restriction enzymes. On the one hand, regions such as heterochromatin are believed to have more compact organisation than euchromatin, perhaps dependent upon the presence of different proteins (STRAUSS and VARSHAV-SKY 1984). On the other hand the standard procedures for preparation of metaphase chromosomes, particularly methanol-acetic acid fixation, are known to extract many proteins from chromosomes (HANCOCK and SUMNER 1982;BURKHOLDER and DUCZEK 1980;HUTCHISON and WEINTRAUB 1985).
In the present study we have investigated the action of restriction endonucleases on unfixed meta-phase chromosomes prepared in two fundamentally different ways, and compared the results with those obtained on chromosomes fixed with minimal change in methanol, as well as with those from chromosomes fixed in methanol-acetic acid. We have found that the action of REs on chromosomes is greatly influenced by the method of preparation, and that methanol-acetic fixation is generally necessary for REs to induce banding on metaphase chromosomes. We conclude, in agreement with other observations (SUMNER et al., in preparation) that chromatin organisation is indeed an important determinant of RE-induced banding patterns on chromosomes.

Cytocentr$uge preparations
Unfixed metaphase chromosome spreads were made as described by PERRY and THOMSON (1986). Briefly, human lymphocytes were cultured in the presence of phytohaemagglutinin, arrested at metaphase with colcemid, treated with hypotonic solution, and spun down on to slides using a Shandon Cytospin centrifuge. The preparations were allowed to dry thoroughly before further treatment.

Isolated chromosome preparation;\
Chromosomes were isolated from human lymphoblastoid cells arrested in metaphase with colcemid using the polyamine method of SILLAR and YOUNG (1981) or the KCM buffer method of GOODERHAM and JEPPESEN (1983). The isolated chromosomes were centrifuged on to 13 mm diameter coverslips in a multiwell plate using a Sorvall ST 60,000 refrigerated centrifuge at lo00 rpm. The chromosome preparations on the coverslips were kept in the appropriate isolation buffer at 4°C until required for further treatment.

Fixation
Both Cytospin preparations, and preparations of isolated chromosomes prepared by both methods, were examined unfixed, or fixed in cold 80 O/o methanol at -20°C for 20 min, or in methanol-acetic acid (3: 1) for 20 min at room temperature.

Restriction enzyme digestion
Slides were incubated with the enzymes, obtained from Boehringer, dissolved in the appropriate buf-fers according to the manufacturer's recommendation, as follows: Control slides were incubated in the appropriate buffers in the absence of the enzyme. Unless otherwise stated, all incubations were for 3h at 37°C; in a few experiments the incubation time was reduced to lh. After incubation the slides or coverslips were stained with Gurr's Giemsa R66 (BDH Ltd) diluted 1:20 in distilled water, for 10 min.

Cytospin preparations
Alu I largely destroyed unfixed chromosomes, although with a shorter, lhr, digestion, some residual C-bands could be seen. Cytospin preparations fixed with cold methanol were apparently unaffected by the enzyme (Fig. la), while methanol-acetic fixed preparations were often C-banded ( Fig. 1 b). Control slides incubated in buffer showed uniformly stained chromosomes with good morphology.
Unfixed chromosomes digested with Hae I11 showed some C-bands, and occasionally some Gbands (Fig. 2). Poor quality G-banding was also seen on Cytospin preparations fixed with cold methanol or methanol-acetic acid. Unfixed control slides again showed uniformly stained chromosomes with good morphology.
After Hind I11 digestion, chromosomes of Cytospin preparations, whether unfixed or fixed, were essentially unbanded, or at best showed only indistinct banding, and thus did not differ significantly from control slides (Fig. 3).
Hinf I digestion proved to be very destructive to unfixed chromosomes (Fig. 4a), although control slides incubated only in buffer showed much better morphology (Fig. 4b). Chromosomes fixed in cold methanol or in methanol-acetic acid and digested with Hinf I often showed paler C-bands, however, (Fig. 4c), and patchy digestion (Fig. 4d).

Isolated chromosomes
No substantial differences were seen between the action of restriction enzymes on chromosomes isolated by the KCM method and those isolated by the polyamine method. In all cases unfixed chromo-   No banding pattern. somes were completely destroyed or dispersed as a Hue 111, Hind 111, or Hinf I recognisable, uniformly result of 3h digestion, although unfixed controls, stained chromosomes were retained, while lhr treated only with buffer, largely retained their digestion with A h I still destroyed chromosomes. morphology and stained uniformly (Fig. 5a), except Isolated chromosomes fixed with cold methanol for buffer H (control for HinfI), which itself tended were also largely destroyed or swollen by all the to destroy the chromosomes. After l h digestion with enzymes tested. but chromosomes fixed in metha- nol-acetic acid proved to be rather more resistant. After A h I digestion, isolated chromosomes fixed in methanol-acetic were heavily extracted, but retained some C-bands (Fig. 5b). Hue 111 apparently had no effect on such chromosomes, while Hind I11 extracted much of the chromosomal material, leaving the centromeres or chromosome cores. Hinf I again destroyed methanol-acetic fixed chromosomes.

Discussion
Our results show clearly that the action of restriction endonucleases on metaphase chromosomes is strongly influenced by the method of preparation. In general, unfixed chromosomes are attacked most readily by REs, and are often largely destroyed. Fixation of either isolated chromosomes or cytospin-prepared metaphase cells increased their resist-

Fig. 5a and b. Isolated chromosomes. a Unfixed chromosomes incubated in Buffer A and stained with Giemsa. Morphology well preserved, and no banding. b Chromosomes fixed in methanol-acetic acid, digested with A h I and stained with Giemsa. Chromosomal material largely extracted. but residual C-bands remain.
ance to digestion. There is also a clear difference between the effect of REs on Cytospin preparations, which are more resistant, and isolated chromosomes, which are relatively labile even after fixation. Most importantly, it is only after methanolacetic acid fixation that the banding patterns, known to be induced by REs in conventionally prepared metaphase chromosomes, can be seen, although results are generally less clear than on conventional preparations.
These +esults can be explained readily in the context of the available information about the different types of chromosome preparations. In Cytospin preparations, the chromosomes are still surrounded by a quantity of cytoplasm, which is clearly still present after fixation (see Figures) and which no doubt helps to stabilise the chromosome structure as well as inhibiting access of REs. In contrast, the enzymes have direct access to isolated chromosomes, and their greater lability is shown by their frequent loss or distortion during the procedures described in this paper. Indeed, it was the lability of completely unfixed chromosomes that led us to use cold methanol fixation, a procedure reported to result in least loss of chromatin organisation and of chromosomal components (HUTCHISON and WEINTRAUB 1985). Use of chromosomes fixed in cold methanol should therefore provide information on the effect of REs on stabilized chromosomes that, nevertheless, have an organisation close to that of native chromosomes. Our observations indicate that, in general, such chromosomes are not substantially atfected by the restriction endonucleases tested. It requires fixation by methanol-acetic acid, a procedure known to alter the structure of chromatin, and to extract significant amounts of protein (HANCOCK and SUMNER 1982; BURKHOLDER and DUCZEK 1980; HUTCHISON and WEINTRAUB 1985) to permit sufficient access of the REs to produce clear banding patterns. Although our experiments do not address directly the question of whether, in conventional methanol-acetic acid fixed chromosomes, the banding pattern is influenced by chromatin organisation, rather than simply reflecting the distribution of recognition sites for the enzyme, they do indicate that chromatin organisation does affect restriction endonuclease digestion of chromosomes. It is important to note that the retention of a more nearly native organisation in methanol-fixed chromosomes could affect the results of RE digestion in two ways; firstly by preventing access of the enzymes to the chromosomal DNA, and secondly, by preventing loss of the DNA fragments resulting from nuclease digestion.
Our results differ from those of BURKHOLDER and SCHMIDT (1986) and of BURKHOLDER (1989), who obtained, on isolated unfixed mouse chromosomes, banding patterns induced by REs that resembled those obtained on conventionally prepared, methanol-acetic acid fixed chromosomes. These differences can probably be explained to a large extent by the use of different procedures for preparation of unfixed chromosomes, and by different protocols for RE digestion. Nevertheless, BURKHOLDER (1989) does present evidence that in some situations access of a restriction enzyme to chromosomal DNA can be inhibited.