Distribution of Chromomeres As a Basis of Chromosomal Coiling

J. Cell Sd. 80, 193-205 (1986) 193 Printed in Great Britain © The Company of Biologist! Limited 1986

DISTRIBUTION OF CHROMOMERES AS A BASIS OF CHROMOSOMAL COILING

VEIKKO SORSA Department of Genetics, University of Helsinki, Finland

SUMMARY Periodicity in the distribution of prominent bands was analysed from the light and electron microscopic maps of salivary gland chromosomes of Drosophila melanogaster. The data obtained indicate that a similar distribution of prominent chromomeres in an individual interphase chromatid results in a unilateral accumulation of chromatin at the chromonema stage, if the helical axis of chromonema consists of ~S-9 interchromomere + chromomere units per turn. Orientation of the largest chromomeres mainly on one lateral half and the smallest chromomeres mairrly on the opposite lateral half of the chromonema apparently bends it to form the chromosomal 'macro* coil. Thus the increase in DNA content in the chromomeric loops located at specific intervals along the chromatids may have an important role in the evolution of coiling hierarchy in the eukaryotic chromosomes.

INTRODUCTION At about the same time in the last century as the helical coiling of chromosomes was first depicted by Baranetzky (1880) the polytene chromosomes were also described as striated cords in the cell nuclei of certain insects (Balbiani, 1881). Since then an abundance of reports has been published concerning both the coiling of chromosomes (see e.g. Kaufmann, 1936, 1948; Ris, 1961; Ris & Korenberg, 1979) and the banding pattern in the polytenized interphase chromosomes (see Beermann, 1962). Although chromosomal coiling has been clearly demonstrable by means of light microscopy (LM) (e.g. see Ohnuki, 1968), electron microscopy (EM) of whole mounts has often failed to uncover it, and favoured the folded fibre or radial loop models of mitotic chromosomes instead of coiling hierarchy (DuPraw, 1970; Labhart, Koller & Wunderli, 1982; Utsumi, 1982). Occasionally, coiling has also been detected in whole-mounted chromosomes by means of both transmission EM (e.g. see Colomb & Bahr, 1974; Haapala & Nokkala, 1982) and scanning EM (e.g. see Harrison, Britch, Allen & Harris, 1981; Mullinger & Johnson, 1983). EM of cross-sectioned mitotic chromosomes has strongly supported the radial loop organization of coiled chromosomes (Marsden & Laemmli, 1979; Adolph, 1980a,6) but the existence and nature of the chromosome 'scaffold' (see Earnshaw & Laemmli, 1983) are still disputable (e.g. see Okada & Comings, 1980; Hadlaczky, Praznovsky & Bisztray, 1982; Nasedkina & Slesinger, 1982). Essential contributions in favour of the side loop model of chromatid structure have been the demonstration of the

Key words: chromomeres, chromosomal coiling, Drosophila. 194 V. Sorsa lampbrush type of organization in polytene chromosomes (Sorsa, Pusa, Virrankoski & Sorsa, 1970) as well as in the meiotic prophase of insect chromosomes (Keyl, 1975) and the DNA side loops in spread mitotic chromosomes (Paulson & Laemmli, 1977).

Axial coiling model of chromosomes The whole light-microscopically recognizable hierarchy of coiling with minor, major and supercoils was depicted by Cleveland (1949) from the chromosomes of flagellates. Many of his drawings give an impression that the coiling hierarchy exists in the axial part of chromosomes. Accordingly, recent advances in the studies on chromosome organization indicate that the coiling of chromatids starts with the formation of an 'axial fibre' to which the chromomeric loops are laterally attached (Nokkala & Nokkala, 19856). The contraction of interchromomeric DNA pulls the adjacent chromomeres more closely together. The tight packaging of chromomeric material, particularly in the largest chromomere loops, bends the laterally located 'axial fibre' to form helical coils, which may then be stabilized by 'scaffolding' proteins (Adolph, Cheng & Laemmli, 1977; Laemmli et al. 1978; Earnshaw & Laemmli, 1983, 1984). By coiling of the axial cord all the chromomeric material is orientated radially outwards from the helical axis. This stage, which obviously corresponds to the 'minor' coil in terminology of Cleveland (1949) is called the 'chromonema' stage by Nokkala & Nokkala (19856). According to the-axial coiling model, the further condensation of chromatin compels the chromonema to form larger helical coiling. This order of coiling is called the chromosomal (macro) coil and it evidently corresponds to the 'major' coil in the terminology of Cleveland (1949). However, it is unclear what makes a coiled chromonema form a higher order of helical structure. One reason could be the unilateral accumulation of chromomeric material on one side of the chromonema, which necessitates further coiling. Apparently, distribution of large chromomeres mainly on one side of the chromonema makes its structure bilateral and tends to bend it. Unilateral accumulation of large loops in the chromonema stage implies that they should be located at certain intervals along the axial fibre. It means that a corresponding periodicity in location of large chromomeres should be detectable already in the interphase chromatids, and this could be best studied on the polytenized interphase chromosomes. In the present study the periodicity of heavy bands has been analysed from the salivary gland chromosomes of Drosophila melanogaster in which the banding pattern has been mapped most exactly.

MATERIALS AND METHODS The sites of the ~S20 most prominent bands, which are usually easily detectable in the photomicrographs of the salivary gland chromosomes 1-3 of D. melanogaster, were determined according to the revised reference maps of Bridges (see Lindsley & Grell, 1968). In the 2L chromosome and in the distal half of the X chromosome the sites of prominent marker bands were also localized with the EM maps (Sorsa, 1982, 1984; Sorsa, Saura & Heino, 1983). The average Chromomeres and chromosomal coiling 195 intervals between the marker bands (or doublets or band complexes) were determined as interband + band units. Presuming that the intervals of prominent bands in polytenized interphase chromosomes also represent the intervals of prominent chromomeres in the individual interphase chromatids, the results were tested against the concept of axial coiling hierarchy of chromosomes. The axial coiling model of chromosomes has been strongly supported by the recent results of Nokkala & Nokkala (1985a,6).

RESULTS AND DISCUSSION When compiling the reference maps of the salivary gland chromosomes of D. melanogaster, C. B. Bridges already noticed that there are more prominent bands at certain rather regular intervals along the polytene chromosomes. Bridges (1935) utilized this special distribution of distinct bands by using them as border-lines of divisions and subdivisions in his guide maps of the salivary gland chromosomes of D. melanogaster. The distribution of heavy bands is also clear in the photographic maps of Lefevre (1976). A comparison of the photo map of Lefevre with the revised camera lucida maps of Bridges (Lindsley & Grell, 1968) and with the electron microscopic maps (Sorsa, 1982, 1984) shows that many of the heavy bands in the photo maps are actually formed by groups of closely adjacent bands. In terms of the polyteny hypothesis this implies that also in every individual interphase chromatid the large chromomeres and groups of them are located at similar intervals as shown by the banding pattern in the polytenized interphase chromosomes. To study the distribution of prominent bands shown by the photomicrographs of salivary gland chromosomes of D. melanogaster, the sites of those bands were marked on the revised reference maps of Bridges and on the EM maps. The average distances of marker bands were determined as interband (+band) units (i.b. units).

Distribution of prominent bands in the polytene chromosomes of D. melanogaster The total axial length of the salivary gland chromosomes X, 2 and 3 of D. melanogaster is ~2173 ^m according to the revised reference maps of Bridges. This length is approximately 4*5 % of the total DNA length of the D. melanogaster genome, which is estimated to be ~47-6mm (Laird, 1971). The number of prominent bands, which can usually be recognized in the light micrographs of chromosomes X, 2 and 3 is approximately 520. The total number of i.b. units in the chromosomes X, 2 and 3 is —-5010 according to the revised reference maps (see Lindsley & Grell, 1968) giving on average —9-6 i.b. units per prominent marker band. If the short interbands in between the halves of the doublet bands of Bridges are excluded, the respective numbers are —3695 and ~7 i.b. units. The number of i.b. units in regions between the marker bands varies from 2 to ~20. Correspondingly, there are in a single chromatid of polytene chromosome 2-20 interchromomere + chromomere units (i.e. units) between the same prominent markers. According to the axial coiling model a marker chromomere and its adjacent region consisting of a certain number of i.e. units correspond to a complete turn of 196 V. Sorsa

Table 1. Distance of prominent light-microscopic marker bands according to the revised reference maps of Bridges given as the average number of i.b. units Average distance of Chromosome markers (i.b. units) S.D. (x) X(l) 6-9 2-08 2L 7-1 1-94 2R(2) 7-4 1-98 3L 6-8 1-98 3R(3) 7-5 1-92

Means for (1W3) 714 1-98

The regions of salivary gland chromosomes between the marker bands are proposed to correspond to chromonemal coils in mitotic chromosomes of D. melanogaster. The intervals of the prominent marker bands consisting of 12 or more i.b. units are proposed to form two turns in the chromonema.

chromonemal coil in mitotic chromosomes. Assuming that the intervals of marker chromomeres containing 12 or more i.b. units correspond to two turns in the chromonema between the nearest dominant loops, the average number of i.e. units representing a single complete turn at the chromonema stage would be 7-14 in chromosomes X, 2 and 3 (see Table 1). In principle, the axial attachment sites of two adjacent marker loops (or tight groups of them) are located on the same lateral half of the chromonema if their distance apart is less than 1-5 times but more than 0-5 times the average coil length in the chromonemal axis (Fig. 1). Analysis of mutual intervals of prominent marker bands in the polytene chromosomes indicates that about 25 % of intervals of prominent bands are more than 1*5 times the average interval. On the other hand, less than 10% of intervals are shorter than 0-5 times the average distance. This means that in the chromonemal coil of an individual chromatid, formed with a regular length of ~7 i.e. units per turn, roughly about 70% of prominent loops should be able to orientate on the same lateral half of the axis. If a variation of ±2 i.e.

Fig. 1. A schematic representation of the distribution of large chromomeres forming the prominent marker bands (identified by larger numbers) in a region of polytene chromosome (A), their location in an individual chromatid (B), as well as in the coiled axis of the chromonema stage formed according to coiling model I (c) and according to model II (D). Model I (c) proposes that most of the axial coils are generally formed by an equal number of (7) i.e. units, and smaller or larger coils are only needed for shifting the coiling frame in the sites of disturbance. A regularly coiled region of the chromonemal axis with 7 i.e. units per turn is depicted in front and side views. Opposite lateral halves (a and b) are separated by a dotted line. The proposed nucleosomal organization in the axial fibre is shown only in cross-section, and in the first coil in side view (see Fig. 2). Model II (D) proposes that the number of i.e. units varies turn by turn along the chromonemal axis to maintain the most optimal orientation of large chromomeric loops. The long intervals are drawn to form two axial turns between adjacent marker chromomeres. The orientation of large chromomere loops in a chromonema formed according to model I is shown schematically in E. The basic idea is that unilateral accumulation of chromatin bends the chromonema to form a larger, chromosomal coil as shown in Fig. 2. Chromomeres and chromosomal coiling 197

39 43 4546 5051 53 198 V. Sorsa

Table 2. The average number of i.b. units counted between the LM marker bands or groups in the EM maps of divisions 1-10 of chromosome X (Sorsa, 1982; Sorsa et al. 1983) and 21-40 of chromosome 2L (Sorsa, 1984) Average no. of i.b. Estimated total no. Chromosome units between of chromonemal region marker bands s.D. (x) coils X(l-10) 8-05 2-55 54 7-04* 1-74* 64*

2L 7-57 3-07 94 7-01* 2-13* 104*

• The distances of marker bands containing 12 or more i.b. units arc interpreted to represent two turns in the axial coil of chromonema. units (roughly corresponding to the standard deviation obtained in this analysis) is allowed in the average length per axial turn, practically all the dominant loops can be orientated to the same lateral side of the helical axis in the chromonema stage. The difference in the average amount of chromatin included in the chromomere loops on opposite lateral halves of the chromonemal axis is difficult to estimate. A rough comparison of the axial length of bands in the polytene chromosomes, which is shown to be correlated with the amount of DNA in chromomeres forming the bands (Laird, Ashburner & Wilkinson, 1980), indicates that the amount of chromatin in the chromomeres of prominent bands is about 10-20 times greater than that in the chromomeres of the narrowest bands. As expected, a similar difference exists in the need for space for the packaging of large and small chromomeric loops on their side of the chromonema. Thus the strongly unilateral location of chromatin compels the chromonemal axis to bend in the direction of the smallest chromomeres and to form a larger helical coil. In this chromosomal macro coil all the largest loops tend to orientate outwards.

Distance of heavy marker bands in the EM maps A more detailed study of the mutual distances of light-microscopic marker bands was carried out by using the electron-microscopic maps of divisions 1—10 of chromosome X and 21-40 of 2L (Sorsa, 1982, 1984; Sorsa et al. 1983). The results of this study are given in the Table 2. According to the EM maps the average interval is ~8 i.b. units between the 54 marker bands of chromosome X divisions 1-10, and ~7-6 i.b. units between the 94 marker bands found in chromosome 2L. The standard deviations were calculated to be ~2-5 and 3 i.b. units, respectively. However, there are ~10 intervals in the mapped area of both X and 2L, which are apparently long enough to form two chromonemal coils between the nearest markers. If all the intervals of 12 or more i.e. units in chromatids are proposed to be able to form two turns in the chromonemal axis, the total number of chromonemal coils is Chromomeres and chromosomal coiling 199

~64 in the distal half of X and ~104 in 2L chromosome. The average length of all these coils is ~7 and the standard deviation is ~2 i.e. units in both chromosomes.

Models of chromosomal coiling The earlier radial loop models proposing a random distribution of radial loops in the chromonema stage do not satisfactorily explain the appearance of macro coiling in eukaryotic chromosomes (e.g. see Comings, 1977; Marsden & Laemmli, 1979; Adolph & Kreisman, 1983; Pienta & Coffey, 1984). The present model suggests that the periodic distribution of large chromomeric loops could be the basis for further coiling of the chromonema. As is shown schematically in Fig. 1, two slightly different models for chromonemal coiling can be proposed. Both of the models are capable of orientating most of the prominent chromomeric loops laterally on one side, and most of the smallest chromomeric loops on the opposite side of the chromonemal axis (see Fig. 2). Model I (Fig. lc) is based on the principle that the adjacent coils in the chromonemal axis tend to be composed of equal numbers of i.e. units. Usually this is enough to orientate large loops on the same side of the axis. However, in certain regions of chromatids some different sizes of coils are apparently needed to retain the normal coiling frame and to maintain the unilateral orientation of large chromomeres (see Fig. 1E). Model II (Fig. ID) proposes that the optimal orientation of large chromomeric loops on one side of the axis is obtained by continuous variation in the length of coils, i.e. by the variation in the number of i.e. units per turn and in the tightness of coiling in the axial cord. EM studies on the structure of band chromatin in polytene chromosomes have shown that the nucleosome-chromatosome type of organization exists in chromatin fibres (e.g. see Sorsa, 1976). In mitotic chromosomes the large chromomeric loops are evidently shortened by the formation of larger, so-called 'solenoidal' coils as has been observed both in thin sections and in whole mounts (Marsden & Laemmli, 1979; Adolph, 1980a,6; Harrison et al. 1981; Utsumi, 1981). The mode of condensation of interloop DNA is not known. Assuming that a nucleosomal organization also exists in the axial DNA, the interloop regions of chromatids probably tend to form a 'solenoid'-sized coil with three to seven chromatosomes per turn as is generally found in chromatin (e.g. see Finch & Klug, 1976; Marsden & Laemmli, 1979; Ris & Korenberg, 1979; Adolph, 1980a,fe; Zatsepina, Polyakov & Chentsov, 1983; Derenzini, Hernandez-Verdun & Bouteille, 1983). The proposed mode of coiling in the 'axial fibre' (Nokkala & Nokkala, 19856) is shown in Figs lc and 2. According to the axial coiling model the length of turns in a chromonemal coil is mainly determined by the number of i.e. units and the tightness of 'solenoidal coil' in the axial fibre. An interesting question is the possible role of the regions of chromonemal coils that are orientated inwards in the chromosomal macro coils and are mainly composed of small chromomeric loops? Are there some special binding sites for contractile 'scaffolding' proteins (Adolph et al. 1977; Laemmli et al. 1978; 200 V. Sorsa

Earnshaw & Laemmli, 1983, 1984), which effectively shorten this side of the chromonema and accomplish its bending into the next order of coiling (see Haapala &Nokkala, 1982)? As shown in Fig. 3 the multistranded axial cord, which often appears by spreading the mitotic chromosomes (e.g. see DuPraw, 1970; Yunis & Bahr, 1979; Mullinger & Johnson, 1980), may be formed by entangling and stretching of the chromomeric loops orientated along the main axis of chromosomes as was previously proposed by Earnshaw & Laemmli (1984). Occasionally, the edges of this multifibrillar 'axis' may form two parallel bundles resembling 'half chromatids' (see Comings & Okada, 1970).

Fig. 2. A schematic drawing of different phases (a-f) of mitotic coiling as based on the axial coiling models presented above, a. Lampbrush organization in an individual chromatid before the condensation of chromatid axis. b. Nucleosomal organization appears in the axial DNA. c. Formation of axial fibre by further coiling of the nucleosome-chromatosome fibre between the chromomere loops, d. Tight packaging of chromatin loops laterally to the axis bends it to form a helical coil with all the chromomeric loops radiating outwards, e. Periodical distribution of large chromomeres along the chromatid axis is able to orientate the largest loops on one side of the chromonema, which then compels the chromonema to bend and form a chromosomal coil with the largest loops outwards. (Large loops attached to the front of the upper chromosomal coil are not depicted, because of showing the axial coiling of the chromonema.) f. The macro coils in a mitotic chromosome with outward protruding large loops. Chromomeres and chromosomal coiling 201

Coiling of chromosomes 2L and XofD. melanogaster By using the number and average length of chromonemal coils, as estimated above on the basis of EM maps, and the number of chromosomal coils in the mitotic chromosomes X and 2L some outlines of the coiling process in the chromosomes of Drosophila can be proposed. Approximately 94 prominent bands or band groups can

Fig. 3. A schematic explanation of the appearance of half-chromatid-like splitting of the multistranded axis formed by the entangling of axially orientated chromomeric loops of adjacent chromosomal coils (a) by the spreading of mitotic chromosomes. Locally the longitudinally orientated loops may be tightly bundled by the axial fibre (b). The gradual uncoiling of loop chromatin (c) down to nucleosome level (d) and to DNA (e) is depicted in some chromomeric loops (see the text). 202 V. Sorsa generally be counted in the light micrographs of chromosome 2L. According to the EM map of chromosome 2L (Sorsa, 1984) the average mutual distance given as the number of i.b. units between those marker bands is ~7-6. If the intervals of 12 or more i.e. units in chromatids are split between two coils, the total number of chromonemal coils is 104 and their average length is ~7 i.e. units (see Table 2). In the chromonema stage of the 2L chromosome the dominant chromomeres in all turns of the axial coil tend to orientate so that they are able to protrude outwards in the chromosomal coil. The number of chromosomal macro coils, which can be counted directly from the preparations of mitotic chromosomes, is, according to Nokkala (personal communication), ~S-6 in 2L and X. By dividing the number of chromonemal coils (104 in 2L) by the number of chromosomal coils (5-5) the number of chromonemal coils per macro coil is obtained. In the 2L chromosome it is ~19. Accordingly, it can be calculated that the mapped distal half of X chromatid, which forms ~3 macro coils in mitotic metaphase, contains ~21 chromonemal coils per macro coil.

Evolutionary basis of chromosomal coiling Keyl (1965) suggested that the DNA of individual chromomeric loops may occasionally be doubled by tandem duplications due to unequal crossing over between homologues. Independent control of the amount of DNA in individual chromomeres apparently has several important evolutionary advantages. It has been proposed that the DNA of the central parts of chromomeric loops most frequently participates in the crossing over between homologues (see Edstrom, 1975; Sorsa, 1975). Thus the crossing-over distance of genes may be partly controlled by the increase and decrease in the length of individual chromomeric loops. It now seems that the increase in the size of certain chromomeres located at suitable intervals from each other along the whole chromosome may induce the macro coiling of chromonema. In the chromosomes of D. melanogaster the average intervals of ~7—8 i.e. units between the prominent chromomere loops seem to offer the structural basis to accomplish the coiling observed in the mitotic chromosomes by orientating most of the largest chromomeres outwards at the chromosomal macro coil. Depending on the differences in periodicity a variation in the length of coiling units may exist in different chromosomes and between different species. Accordingly, it may be predicted that in the interphase chromatids of those species that have in their chromosomes the capability of forming an even higher order of coiling, there may also exist a higher order of periodic grouping of chromomeric material. This longer periodicity is able to induce a unilateral accumulation of chromatin also in so-called 'secondary chromonema' (Nokkala & Nokkala, 1985a) and to bend it to form a higher order of coiling. This largest order obviously corresponds to the 'super' coil in the terminology of Cleveland (1949). Why the 'secondary chromonema' appears only in the first meiotic division of certain species is an open question. Similarly, the possible role of repetitive DNA sequences in the recognition and lining up of the chromomeric loops during the different phases of coiling processes is not known. Chromomeres and chromosomal coiling 203

The results of the present analysis on the polytene and mitotic chromosomes of D. melanogaster indicate that the periodical location of large chromomeres in the interphase chromatids is regular enough to act as the structural basis for chromosomal macro coiling in mitosis. Thus the results apparently support the hypothesis that the distribution and the size of chromomeric loops are essential in inducing macro coils in eukaryotic chromosomes. The development of an optimal distribution of large chromomeres with specific intervals is apparently disturbed by chromosomal rearrangements as well as the independent and sometimes contradictory evolutionary selection pressure controlling the crossing-over frequency between the neighbouring genes.

The study was supported financially by the National Research Council of Sciences of Finland. I thank Dr Seppo Nokkala for valuable discussions and unpublished data concerning the organization of mitotic chromosomes of Drvsophila.

REFERENCES ADOLPH, K. W. (1980

HADLACZKY, G., PRAZNOVSKY, T. & BISZTRAY, G. (1982). Structure of isolated protein-depleted chromosomes in plants. Chromosoma 86, 643-659. HARRISON, C. J., BRITCH, M., ALLEN, T. D. & HARRIS, R. (1981). Scanning electron microscopy of the G-banded human karyotype. Expl Cell Res. 134, 141-153. KAUFMANN, B. P. (1936). Chromosome structure in relation to the chromosome cycle. l.Bot. Rev. 2, 529-553. KAUFMANN, B. P. (1948). Chromosome structure in relation to the chromosome cycle. II. Bot. Rev. 14, 57-126. KEYL, H.-G. (1965). Duplikationen von Untereinheiten der Chromosomalen DNS wahrend der Evolution von Ckinmomus thummi. Chromosoma 17, 139-180. KEYL, H.-G. (1975). Lampbrush chromosomes in spermatocytes of Chironomus. Chromosoma 51, 75-91. LABHART, P., KOLLER, T. & WUNDERU, H. (1982). Involvement of higher order chromatin structures in metaphase chromosome organization. Cell 30, 115-121. LAEMMU, U. K., CHENG, S. M., ADOLPH, K. W., PAULSON, J. R., BROWN, J. A. & BAUMBACH, W. R. (1978). Metaphase chromosome structure: The role of nonhistone proteins. Cold Spring Harbor Symp. quant. Biol. 42, 351-360. LAIRD, C. D. (1971). Chromatid structure: Relationship between DNA content and nucleotide sequence diversity. Chromosoma 32, 378—406. LAIRD, C. D., ASHBURNER, M. & WILKINSON, L. (1980). Relationship between relative dry mass and average band width in regions of polytene chromosomes of Drosophila. Chromosoma 76, 175-189. LEFEVRE, G., JR. (1976). A photographic representation and interpretation of the polytene chromosomes of Drosophila melanogaster salivary glands. In Genetics and Biology of Drosophila (ed. M. Ashburner & E. Novitski), vol.la, pp. 31-66. London: Academic Press. LrNDSLEY, D. L. & GRELL, E. H. (1968). Genetic Variations of Drosophila melanogaster, Publ. 627. Washington, D.C.: Carnegie Instn of Washington. MARSDEN, M. P. F. & LAEMMU, U.K. (1979). Metaphase chromosome structure. Evidence for a radial loop model. Cell 17, 849-858. MULLJNGER, A. M. & JOHNSON, R. T. (1980). Packing DNA into chromosomes. J. Cell Set. 46, 61-86. MULLINGER, A. M. & JOHNSON, R. T. (1983). Units of chromosome replication and packing. Jf. Cell Set. 64, 173-193. NASEDKTNA, T. V. & SLESINGER, S. I. (1982). The structure of partly decondensed metaphase chromosomes. Chromosoma 86, 239-249. NOKKALA, S. & NOKKALA, C. (1985a). Spiral structures of meiotic chromosomes in plants. Hereditas 103 (in press). NOKKALA, S. & NOKKALA, C. (19856). Coiled internal structure of chromonema within chromosomes suggesting hierarchical coil model for chromosome structure. Hereditas 104 (in press). OHNUKI, Y. (1968). Structure of chromosomes. I. Morphological studies of the spiral structure of human somatic chromosomes. Chromosoma 25, 402—428. OKADA, T. A. & COMINGS, D. E. (1980). A search for protein cores in chromosomes: Is the scaffold an artifact? Am.J. hum. Genet. 32, 814-832. PAULSON, J. R. & LAEMMU, U. K. (1977). The structure of histone-depleted metaphase chromosomes. Cell 12, 817-828. PIENTA, K. J. & COFFEY, D. S. (1984). A structural analysis of the role of the nuclear matrix and DNA loops in the organization of the nucleus and chromosome. J'. Cell Sci. Suppl. 1, 123—135. Ris, H. (1961). The annual invitation lecture: Ultrastructure and molecular organization of genetic systems. Can.J. Genet. Cytol. 3, 95-120. Ris, H. & KORENBERG, J. (1979). Chromosome structure and levels of chromosome organization. In Cell Biology 2, The Structure and Replication of Genetic Material (ed. D. M. Prescott & L. Goldstein), pp. 268-361. New York: Academic Press. SORSA, V. (1975). A hypothesis for the origin and evolution of chromomere DNA. Hereditas 81, 77-84. SORSA, V. (1976). Beaded organization of chromatin in the salivary gland chromosome bands of Drosophila melanogaster. Hereditas 84, 213-220. Chromomeres and chromosomal coiling 205

SORSA, V. (1982). Electron microscopic map for the salivary gland X chromosome of Drosophila melanogaster divisions 1-5. lnAdvances in Genetics, Development and Evolution of Drosophila (ed. S. Lakovaara), pp. 23-32. New York: Plenum. SORSA, V. (1984). Electron microscopic mapping and ultrastructure of Drosophila polytene chromosomes. In Insect Ultrastructure 2 (ed. R. C. King & H. Akai), pp. 75-107. New York: Plenum. SORSA, V., PUSA, K., VIRRANKOSKI, V. & SORSA, M. (1970). Electron microscopy of an induced "lampbrush stage" of polytene chromosomes in Drosophila. Expl Cell Res. 60, 466—469. SORSA, V., SAUKA, A. 0. & HEINO, T. I. (1983). Electron microscopic analysis of the banding pattern in the salivary gland chromosomes of Drosophila melanogaster. Divisions 6 through 10 of X. Hereditas 98, 181-200. UTSUMI, K. R. (1982). Scanning electron microscopy of Giemsa-stained chromosomes and surface-spread chromosomes. Chromosoma 86, 683-702. YUNIS, J. J. & BAHR, G. F. (1979). Chromatin fiber organization of human interphase and prophase chromosomes. Expl Cell Res. 122, 63-72. ZATSEPINA, O. V., POLYAKOV, V. Yu. & CHENTSOV, YU. S. (1983). Chromonema and chromomere. Chromosoma 88, 91-97.

(Received 5 June 1985-Accepted 25 July 1985)

Note added in proof The axial coiling model of mitotic chromosomes introduced above obviously is in good agreement with the recent results of J. B. Rattner & C. C. Lin (1985). Cell 42, 291-296.