The arrangement of the chromatin bodies in the interphase nuclei of 6 erythrocytes has been investigated by means of 3-dimensi0nal reconstruction from electron micrographs of serial sections. When the borders of chromatin bodies are marked on the surface of each model, discrete areas of chromatin in contact with the nuclear envelope are revealed. The number of these areas is approximately equal to the number of chromosomes in the diploid set. The data suggest that each chromatin body corresponds to a condensed interphase chromosome and that each chromosome is attached to one discrete site on the nuclear envelope. The data are insufficient to show whether or not the condensed chromosomes are arranged in any orderly pattern in these nuclei.

The possibility that the individual chromosomes are arranged in an ordered pattern in the interphase nucleus has excited many observers in the last decade (for reviews see Comings, 1968, 1972). It is possible that such order, if it exists, may be reflected in an ordered arrangement of the chromosomes at prophase and on the metaphase plate. Although many investigations have been carried out on mitotic cells the results are contradictory, some authors failing to detect any order in the spatial configuration of the chromosomes (e.g. Fox, Mello-Sampayo & Carter, 1975), some observing only a limited degree of order involving certain chromosomes (e.g. Zang & Back, 1968; Stack & Brown, 1969) and some observing specific association patterns involving the entire chromosome complement (e.g. Costello, 1970; Ashley & Wagenaar, 1974; Ashley, 1976). It is clear that the question of order or disorder in the interphase nucleus can only be answered unequivocally by a direct study of interphase nuclei themselves. In the majority of interphase nuclei, the chromatin is more or less dispersed and individual chromosomes can no longer be identified. Hence direct observations are limited to the location of regions of chromatin with characteristic behaviour or appearance at interphase, for example, late-replicating chromatin (Fussell, 1975) and chromatin stained by C-band techniques (Fussell, 1977; Ghosh & Roy, 1977). However, in mature nucleated erythrocytes virtually all the chromatin is condensed to form the so called nuclear, or chromatin, bodies and there is also evidence which suggests that each body is derived from an individual chromosome (Anderson & Norris, 1960; Davies, 1961). In this paper we report studies on the arrangement of the condensed chromatin in models of newt erythrocyte nuclei constructed from electron micrographs of serial sections.

Preparation and model construction

Details of the methods used are given in Murray (1976). Briefly, erythrocytes from the newt Triturus cristatus (2n — 24) were prepared by allowing them to settle under gravity on to a flat surface. This orients them and also avoids the distortion which can arise due to close packing following centrifugation. Fixation was in 3 % glutaraldehyde in amphibian Ringer followed by washing and fixation in 1 % OsO 2. After dehydration cells were embedded in Spurr resin (1969).

Ribbons of 70 or more serial sections were cut using a Dupont diamond knife on an A. F. Huxley-Cambridge ultramicrotome modified to provide a specimen advance of 250 nm. After expansion with a heat-pen sections were mounted in groups of about 10 on Formvar-coated narrow slot (0 · 2 × 1 · 5 mm) copper grids. The average section thickness of each series was estimated from observation of the white light interference colour of the sections as they floated on the knife-boat.

In preliminary experiments, sections of thickness about 0 · 5 to 1 · 0 μ m were examined in a 10 V electron microscope (kindly made available to us at the National Physical Laboratory, Teddington, G.B.) with the idea of reducing the number of sections needed for reconstruction, but this proved not to be practical. As Dr L. D. Peachey (private communication) has pointed out, difficulty in reconstruction arises whenever the individual structures themselves are of the same order of size as the thickness of the section, or less. The details become ambiguous as a result of the 2-dimensi0nal nature of the projection image. The problems encountered can sometimes be overcome by stereoscopy, but we found it simpler to use thinner sections and this has the advantage that a 10-kV electron microscope can be employed.

Sections were stained by immersion in 2 % aqueous magnesium uranyl acetate for 30 min at 60 °C, followed by immersion for 6 min in lead citrate (Reynolds, 1963). After coating with a thin layer of carbon, sections were examined in a Siemens Elmiskop I operating at 10 kV with a 200-μm condenser aperture and a 50-μm objective aperture. Nuclei suitably oriented and entirely contained within the ribbon of sections were photographed at a nominal magnification of × 5000. The exact magnification was determined with a carbon-grating replica (Agar Aids).

High-contrast prints of the electron micrographs were made on sheets of transparent film and superimposed in transmitted light. Prints were aligned in pairs using the criterion of best fit. Knife marks and adjacent cells were used as an aid in alignment. A similar method has been reported by Sjöstrand (1974).

The image from each negative was projected in a photographic enlarger on to a sheet of expanded polystyrene at a magnification determined by the relationship between the average thickness of the sections and that of the polystyrene sheet. Profiles of the chromatin areas were drawn on the polystyrene sheet and cut out using an electrically heated wire. The models were assembled by glueing together the cut-out profiles in the position determined by the aligned prints. Polystyrene bridges, left behind so as to maintain isolated areas in position, were removed as the models were assembled. The completed models were cut into quarters, with the hot wire, to reveal the interior.

Six models were constructed, each containing between 30 and 40 sections, at a magnification of 27000.

Identification of borders

The borders of the chromatin bodies were identified in the photomicrographs as outlined below. Subsequently they were marked on the surface of each model. The borders could be clearly defined when areas of chromatin were separated by a clear gap arising, for example, from interchromatin channels terminating at the nuclear envelope (Fig. 1, single-headed arrows). In thin, say 90-nm, sections such chromatin channels can often be seen to terminate in nuclear pores, as is well known. But in the thicker 250 nm sections employed for reconstruction well defined pores could not be distinguished. The clear gap seen in these sections presumably arises from several pores lying in close proximity along a line normal to the section. Frequently, in the micrographs of the 250-nm section, 2 areas of chromatin were seen to be clearly separated apart from a narrow strip of chromatin adjacent to the nuclear envelope (Fig. 1, double-headed arrows). Such strips are thought to result from the spreading out of the edges of chromatin bodies into a thin layer in contact with and covering areas of membrane surrounding the pores, forming projections between them which link one chromatin body to another. Their width will be dependent on the way the chromatin body thins down in the regions adjacent to the pores. Thus strips will be seen whenever sections are cut perpendicular to lines of pores which are not in close proximity, or cut obliquely so as to include the thin layer of chromatin adjacent to the pores. Areas of chromatin connected by such strips are also classed as having clearly defined borders, and this conclusion is supported by 2 observations. First, finger-like processes of dimensions similar to the strips can be seen to project from the surfaces of chromatin bodies within the interior of the nucleus (Fig. 1, small arrow-heads). Second, the strips sometimes have a density which is lower than the adjacent areas, consistent with the section passing through a pore, or part of one. For the sake of simplicity the positions of clearly defined borders of adjacent chromatin bodies were marked on the surface of each model with single solid lines, rather than two separate ones.

Fig. 1.

Electron micrograph of one of a series of 0· 25-μm sections through a newt erythrocyte nucleus: the position of this section within the completed model is shown at arrows in Figs. 2 and 4. Some chromatin areas at the surface of the nucleus are separated by a clear gap: examples are shown at single-headed arrows. Other chromatin areas are connected by a narrow strip of chromatin: examples are shown at double-headed arrows. The region with little chromatin adjacent to the nuclear envelope is the site of a border which lies in the plane of the section: see lower left, between large arrow-heads. Small arrow-heads indicate chromatin projections. × 9300.

Fig. 1.

Electron micrograph of one of a series of 0· 25-μm sections through a newt erythrocyte nucleus: the position of this section within the completed model is shown at arrows in Figs. 2 and 4. Some chromatin areas at the surface of the nucleus are separated by a clear gap: examples are shown at single-headed arrows. Other chromatin areas are connected by a narrow strip of chromatin: examples are shown at double-headed arrows. The region with little chromatin adjacent to the nuclear envelope is the site of a border which lies in the plane of the section: see lower left, between large arrow-heads. Small arrow-heads indicate chromatin projections. × 9300.

In some micrographs areas of condensed chromatin occasionally had an appearance which suggested the presence of 2 chromatin bodies whose borders could not be clearly defined. These areas were characterized by a combination of the following: discontinuities in the chromatin profile; chromatin with an inhomogeneous or mottled appearance; the presence of a definite border in a corresponding position in the preceding and succeeding sections. Such areas were presumed to mark the junction between 2 separate chromatin bodies with ill-defined, or what we will term obscured borders. Obscured borders were marked on the surface of the models with a broken line. The ratio of obscured to clearly defined borders, i.e. length of broken to solid lines, on the surface of the models was approximately 1 to 6.

We envisage that the nuclear pores, singly, 01 in groups, are generally confined to the boundary areas between chromatin bodies and form curvilinear arrays on the nuclear surface outlining the sites of attachment. This concept receives some support from freeze-fracture studies (La Fountain & La Fountain, 1973) on the generative and vegetative nuclei in pollen from Tradescantia paludosa. In the former relatively inactive nucleus there were curvilinear arrays of pores with comparatively large areas free from pores.

The numbers in parentheses (col. 2) refer to alternative possibilities in models where identification of 2 areas as separate was uncertain.

All 6 models of nuclei were flattened ellipsoids in shape. The volume of each nucleus was calculated to be approximately the same for all models, to within 15%. Minor differences between the models were observed, in particular the extent to which the nuclear surfaces were indented. Two of the nuclei had membrane inclusions, one in the form of a membrane continuous with the nuclear envelope, and lying in the haemoglobin-containing regions of the nucleus, the other in the form of a membrane passing through the condensed chromatin itself.

The border lines marked on each polystyrene sheet, when joined together on the surface of the model, enclose a number of discrete areas each of which represents an area of chromatin in contact with the nuclear envelope and separate from the adjacent chromatin (Figs. 2-5). We call these surface areas the nuclear envelope-attachment sites. Table 1 shows the total number of envelope-attachment sites found in each model. An average of 25 · 5 ± 1 · 4 sites per nucleus was found. This number is very close to 24, the diploid chromosome number in newt.

Table 1.

The number of envelope-attachment sites and the numbers of separate bodies observed in each model (see text).

The number of envelope-attachment sites and the numbers of separate bodies observed in each model (see text).
The number of envelope-attachment sites and the numbers of separate bodies observed in each model (see text).
Fig. 2-5.

Photographs of a model of the nucleus in a newt erythrocyte rotated about its long axis through 90° intervals. Arrows mark the position of the section shown in Fig. 1. Clearly defined borders are shown by solid black lines, obscured borders by dotted lines (see text). The envelope-attachment sites of chromatin bodies are numbered randomly. C marks those bodies which are clearly separate in the interior of the nucleus. A star marks the position of a line drawn from the nucleolus to intersect, normally, the nearest surface. × 6500.

Fig. 2-5.

Photographs of a model of the nucleus in a newt erythrocyte rotated about its long axis through 90° intervals. Arrows mark the position of the section shown in Fig. 1. Clearly defined borders are shown by solid black lines, obscured borders by dotted lines (see text). The envelope-attachment sites of chromatin bodies are numbered randomly. C marks those bodies which are clearly separate in the interior of the nucleus. A star marks the position of a line drawn from the nucleolus to intersect, normally, the nearest surface. × 6500.

When the interiors of the models were examined each was seen to contain only 1–3 condensed chromatin bodies completely separate from the remaining chromatin. This indicates that most of the bodies must have come into close contact a short distance from the envelope, so as to make them indistinguishable one from another. In addition a number of bodies could be identified as probably-separate, i.e. most of the body, apart from a small region, was separated from the remainder of the chromatin by a clear gap. Table 1 shows the total number of separate and probably-separate bodies found per nucleus. Every body identified as separate or probably-separate was seen to have a single envelope-attachment site.

The suggestion that each chromatin body is derived from an individual chromosome is based on light-microscope observations on erythrocytes from Amphiuma (Anderson & Norris, 1960) and Rana pipiens (Davies, 1961). In favourable preparations of erythrocytes in which the nuclei have flattened on the surface of the slide, discrete chromatin bodies can be seen and their number is roughly similar to the diploid number of chromosomes. However, only a few separate bodies can be distinguished in the models constructed from electron micrographs of serial sections. One possible explanation is that the chromatin bodies become slightly separated in the flattened preparations viewed by light microscopy. Davies & Tooze (1966, figs. 18, 22) also noted that the 2 chromatids, known from light microscopy of whole cells to comprise each metaphase chromosome, often appeared fused in electron micrographs of thin sections. In one favourable model (No. 1 in Table 1) 8 bodies were identified as separate and probably-separate. Together they occupied, at a rough estimate, one third of the total chromatin volume and this, also, is consistent with each body being derived from one chromosome.

Our finding that the number of chromatin body attachment-sites per nucleus is very close to 24, the diploid chromosome number, supports the hypothesis relating chromatin bodies to interphase chromosomes. This number of attachment-sites can be explained in 2 ways. The first possibility is that each condensed interphase chromosome is attached to one site on the envelope. The second possibility, supported by experimental observations on certain cells approaching division, is that each chromosome is attached to 2 separate sites on the nuclear envelope. This would lead, of course, to a number of attachment-sites which is double the chromosome number. But numerical similarity would be restored if, for example, the chromosomes were linked end-to-end to form a chain, with each pair of telomeres attached to an area on the nuclear envelope corresponding to one of our attachment sites. The interactions of chromosomes with the nuclear envelope were reviewed by Franke & Scheer (1974). Attachment of the ends of chromosomes to each other and to the nuclear envelope, was described in cells of Ornithogalum virens (2n = 6) at prophase by Ashley & Wagenaar (1974) and Ashley (1976). In several different species at the pachytene stage of meiosis, the bivalents are also attached by their ends to separate sites on the nuclear envelope (Moens, 1973). In our models of interphase nuclei, however, all the chromatin bodies which could be identified as separate, or probably-separate, were found to be attached to only one site on the envelope and this favours the first of the above possibilities. The absence of distinctive features, centromeres and well-defined telomeres, in the condensed chromosomes of erythrocytes makes it impossible to decide whether there is a specific region on the chromosome involved in attachment to the nuclear envelope.

It is possible that a specific 3-dimensional configuration of the chromosomes within the interphase nucleus could be brought about by end-to-end linkage of chromosomes in a definite order and maintained by attachment of chromosomes to the nuclear envelope. Evidence for linkage of some of the chromosomes in interphase nuclei of Allium cepa was obtained by counting C-band spots (Fussell, 1977). Evidence for linkage in a definite order was obtained in Ornithogalum virens where the individual prophase chromosomes could be identified by a modified Giemsa technique (Ashley, 1976). Our observations on newt erythrocytes show that all the chromosomes are attached to the nuclear envelope but they provide little information about a possible non-random arrangement. Even if all the chromatin bodies had been clearly-separate, it is likely that the differences in their geometry would have been insufficient to permit the identification of individuals, which can be achieved with chromosomes at metaphase. When the blocks of chromatin were compared in the different models there was no detectable similarity. The presence of the separate chromatin bodies is inconsistent with linkage of all the chromosomes by their telomeres. Furthermore these separate bodies did not occupy the same position in all the models. The nucleolus, which acts as a marker for the position of the nucleolar-organizing chromosomes, was also found at a different site in each model. When the size and shape of the envelope-attachment sites were compared no clear correlation could be established. However this does not exclude an orderly arrangement on the surface since the geometry of each site could be variable, that is, not completely characteristic for a particular chromosome.

We thank Professor M. H. F. Wilkins, F.R.S., for his continued interest, Mr D. Back for help in constructing the models and Mr Z. Gabor for help in preparing the plates. A. B. Murray was an M.R.C. Scholar.

Anderson
,
N. G.
&
Norris
,
C. B.
(
1960
).
III. The effects of amines on the structure of isolated nuclei
.
Expl Cell Res
.
19
,
605
618
.
Ashley
,
T.
(
1976
).
Chromosomal order in nuclei
.
J. Cell Biol
.
70
,
395a
.
Ashley
,
T.
&
Wagenaar
,
E. B.
(
1974
).
Telomeric associations of gametic and somatic chromosomes in diploid and autotetraploid Ornithogalum virens
.
Can. J. Genet. Cytol
.
16
,
61
76
.
Comings
,
D. E.
(
1968
).
The rationale for an ordered arrangement of chromatin in the interphase nucleus
.
Am. J. hum. Genet
.
20
,
440
460
.
Comings
,
D. E.
(
1972
).
The structure and function of chromatin
.
Adv. Hum. Genet
.
3
,
237
–431.
Costello
,
D. P.
(
1970
).
Identical linear order of chromosomes in both gametes of the acoel turbellarian Polychoerus carmelensis: A preliminary note
.
Proc. natn. Acad. Set. U. S. A
.
67
,
1951
-
1958
.
Davies
,
H. G.
(
1961
).
Structure in nucleated erythrocytes
.
J. biophys. biochem. Cytoi
.
9
,
671
687
.
Davies
,
H. G.
&
Tooze
,
J.
(
1966
).
Electron- and light-microscope observations on the spleen of the newt Triturus cristatus. the surface topography of the mitotic chromosomes
.
J. Cell Sci
.
1
,
331
350
.
Fox
,
D. P.
,
Mello-Sampayo
,
T.
&
Carter
,
K. C.
(
1975
).
Chromosome distribution in neuroblast metaphase cells of Locusta migratoria L
.
Chromosoma
53
,
321
333
.
Franke
,
W. W.
&
Scheer
,
U.
(
1974
).
Structures and functions of the nuclear envelope
.
In The Cell Nucleus
, vol.
1 (ed
.
H.
Busch
), pp.
244
256
.
New York, London
:
Academic Press
.
Fussell
,
C. P.
(
1975
).
The position of interphase chromosomes and late replicating DNA in centromere and telomere regions of Allium cepa L
.
Chromosoma
50
,
201
210
.
Fussell
,
C. P.
(
1977
).
Telomere associations in interphase nuclei of Allium cepa demonstrated by C-banding
.
Expl Cell Res
.
110
,
111
117
.
Ghosh
,
S.
&
Roy
,
S. Ch
. (
1977
).
Orientation of interphase chromosomes as detected by Giemsa C-Bands
.
Chromosoma
61
,
49
55
.
La Fountain
,
J. R.
&
La Fountain
,
K. L
. (
1973
).
Comparison of density of nuclear pores in vegetative and generative nuclei
.
Expl Cell Res
.
78
,
472
476
.
Moens
,
P. B.
(
1973
).
Mechanisms of chromosome synapsis at meiotic prophase
.
Int. Rev. Cytol
.
35
,
117
-
134
.
Murray
,
A. B.
(
1976
).
Electron Microscope Studies of Condensed Interphase Chromosomes, Their Arrangement and Their Substructure, in Various Blood Cells
.
Ph.D. Thesis, King’s College
,
University of London
.
Reynolds
,
E. S.
(
1963
).
The use of lead citrate at high pH as an electron-opaque stain in electron microscopy
.
J. Cell Biol
.
17
,
208
212
.
Sjöstrand
,
F.
(
1974
).
A search for the circuitry of directional selectivity and neural adaptation through three-dimensional analysis of the outer plexiform layer of the rabbit retina
.
J. Ultra-struct. Res
.
49
,
60
156
.
Spurr
,
A. R.
(
1969
).
A low viscosity epoxy resin embedding medium for electron microscopy
.
J. Ultrastruct. Res
.
26
,
31
43
.
Stack
,
S. M.
&
Brown
,
W. V.
(
1969
).
Somatic pairing, reduction
and recombination
:
an evolutionary hypothesis of meiosis
.
Nature, Land
.
222
,
1275
1276
.
Zang
,
K. D.
&
Back
,
E.
(
1968
).
Quantitative studies on the arrangement of human metaphase chromosomes
.
Cytogenetics
7
,
455
470
.