- Email: [email protected]
Sequence of centromere separation
Sequence of Centromere Separation Minor Satellite DNA Does Not Influence Separation of Inactive Centromeres in Transformed Cells of Mouse Baldev K. Vig and Neidhard Paweletz
ABSTRACT: Neoplastic cells may carry inactive centromeres on some multicentric, yet stable, chromosomes. We report that some inactive centromeres in L929 mouse cells do not contain minor satellite DNA, the DNA fraction which has been suggested to constitute the centromere. We compared the sequence of separation of inactive centromeres carrying the minor satellite with those lacking this fraction. The sequence of separation appears to be independent of whether or not the inactive centromeres carry the minor satellite DNA. The timing of replication of the inactive centrameres is also independent of this DNA. Hence, minor satellite of mouse is not a factor in holding together the subunits of inactive centromeres. Extension of these results to active centromeres might suggest that the minor satellite DNA is not a factor responsible for adhesion of the two centromere sub-units up until late meta-anaphase.
INTRODUCTION Several neoplastic cell lines carry stable dicentric and multicentric chromosomes. The centromeres of normal monocentric chromosomes in a genome separate into two sub-units in genetically controlled, non-random sequence at metaanaphase junction. However, in stable dicentric and multicentric chromosomes the inactive centromeres separate into two units in prophase, generally long before the active or functional centromeres do [10]. The sequence of separation of normal, functional centromeres has been studied for a number of species. This sequential separation seems to depend on the quantity and quality of pericentric heterochromatin [6]; hence, upon the timing of replication of the centromere [see [7]]. In species with qualitatively uniform heterochromatin, e.g., Mus musculus domesticus, the timing of separation is directly related to the quantity of pericentric heterochromatin; the larger the block of pericentric heterochromatin, the latter the separation of the associated centromere [6]. The inactive centromeres, though, separate early irrespective of the quantity of centromeric heterochromatin. This is apparently so because pericentric heterochromatin associated with inactive centromeres replicates earlier than the corresponding regions associated with the active centromeres, even when both are of similar base composi-
From the Department of Biology (B. K. V.), University of Nevada, Reno, Nevada; and Forschungsschwerpunkt 4 (N. P.), Deutsches Krebsforschungszentrum, Heidelberg, BRD. Address reprint requests to B. K. Vig, Department of Biology, University of Nevada, Reno NV 89557-0015. Received January 21, 1993; accepted June 1Z 1993.
© 1993 Elsevier Science P u b l i s h i n g Co., Inc. 655 A v e n u e of the Americas, N e w York, NY 10010
tion. For instance, in m o u s e the repetitive D N A associated with active centromeres replicates in late S but the D N A of the same composition w h e n associated with inactive cantromeres replicates in early S [1]. The minor satelliteof mouse is present in allchromosomes of the M. m. domesticus genome except in the Y chromosome. This D N A fraction has been implicated to be the centromere D N A [3, 4, 11].Ifso, then itis logical to assume that the minor satelliteis responsible for holding the two centromere subunits together up until meta-anaphase. Recently, we studied the distribution of minor satellite D N A in stable dicentric and multicentric chromosomes in L929 mouse cell line.W e reported that the inactivecentromeres in a heptacentric chromosome carry minor satelliteD N A , as do all active centromeres in that cell line.However, inactive centromeres in several dicentric chromosomes do not contain any detectable minor satellite[9].Hence, some primary constrictions can be formed without the participation of this fraction of DNA. Whereas the sequence of centromere separation is established for the active centromeres, it is not k n o w n whether inactive centromeres also follow a particular pattern and hierarchy of separation. Italso stands to reason that the minor satellitein mouse carries similar functions w h e n present in the inactivecentromeres as itdoes for the activecentromeres. Hence, those inactive centromeres that carry the minor satelliteshould separate laterthan those that do not carry this DNA. The present study was, therefore, carried out to find out whether there is a hierarchical sequence of separation among the inactive centromeres and whether the presence of minor satelliteD N A has any influence on the timing of separation of these centromeres.
31 Cancer Genet Cytogenet70:31-38(1993) 0165-4608/93/$06.00
32
J
140 120_~ 100c'(D
8o E
~o
60-
0
rr
40-
20 ! O
3
°
0
2.4
5.5 8.2 10.1 12.8 Chromosome length (microns)
15.2
Figure I A Giemsa-stained cellexhibitingthe heptacentricand a dicentricchromosome with allcentromeres held as unseparated units. The inactive centromeres are marked with small arrowheads, active centromeres with large arrowheads.
Sequence of Centromere Separation M A T E R I A I S AND METHODS L929 cells of mouse origin were u s e d in these studies. The cells were grown in McCoy's 5A m e d i u m w i t h 10% fetal calf serum. The cell line grows equally well w h e t h e r CO 2 is available or not. Before fixing in 3:1 methanol:acetic acid, the cultures were exposed to Colcemid (1 hour) and to hypotonic (0.075 M KCI, 10 minutes). The air-dried slides were stored at room temperature until used. For detecting heterochromatin the chromosomes were either C-banded or stained w i t h DAPI. C-banding was carried out by exposing the cells to 0.2 N HC1 (50 minutes), saturated Ba (OH)2 (6 minute) and reconstituting in 2 × SSC (2 hours at 65°C). The presence of m i n o r satellite DNA was confirmed by in situ hybridization using the entire 120-base (bp) repeats or w i t h a 42-met oligodeoxynucleotide, AACGT GTATA TCATA GAGTT ACAAT GAGAA ACATG G A A A A riG, homologous to a part of the mouse m i n o r satellite [11]. The details of results obtained with the 120-bp repeats are published elsew h e r e [9]. The 42-mer oligomer was synthesized on an app l i e d Biosystems 380a DNA synthesizer as previously reported [5]. Ten p m o l of oligomer were labeled w i t h 1 n m o l biotin-11-dUTP (Sigma)/1.5 n m o l dATP using 50 U terminal deoxynucleotidyltransfemse (Boehringer) in final volume of 30 B1 of reaction buffer s u p p l i e d by the manufacturer. After 3 h incubation at 37°C, the oligomers were ethanol precipitated in the presence of 20 ~g E. coli t-RNA (Boehringer). Finally, the d r i e d pellets was dissolved in tris-EDTA buffer. Prior to in situ hybridization, the cells were digested with RNase (100 ~g/ml PBS). These were, then, denatured in 70% formamide/2 x SSC at 70°C for 2 minutes and dehydrated in chilled methanol series. Five ~1 of the hybridization mixture (2 × SSC, 30% formamide, 0.5 ng/~l labeled oligomer DNA, 0.5 ~g/~l E. coli t-RNA) was a p p l i e d to the slides under a cover slip and sealed with rubber cement. After allowing overnight hybridization at 37°C, the slides were washed in 0.1 M NaHCO3/0.05% Tween 20 at 37°C. Biotin was detected using FITC avidin [Sigma) w i t h one r o u n d of signal amplification. Chromosomes were counterstained with a 1 ~l/ml of p r o p i d i u m i o d i d e in antifade (DABCO) solution. To detect DNA synthesis in the centromere region, async h r o n o u s l y growing cells were exposed to 10 .6 M 5-brom o d e o x y u r i d i n e (BrdU) during the last 5 hours of culturing. For the detection of the signal, the cells were treated w i t h anti-BrdU antibody for 30 minutes at 37°C. Followed
33 by washing in Dulbecco's phosphate buffer, the cells were reincubated for a d d i t i o n a l 30 minutes at 37°C in goat antimouse secondary antibody tagged with FITC. Chromosomes were counterstained w i t h 1:1000 dilution of e t h i d i u m bromide. Those centromeres that replicated before the a d d i t i o n of BrdU w o u l d not show the label; those replicating after the a d d i t i o n of the t h y m i d i n e analog w o u l d be labeled. This approach, however, would not d i s t i n g u i s h between all those centromeres that replicated before the a d d i t i o n of BrdU or between those that replicated after the a d d i t i o n of BrdU. Nonetheless, this treatment was found suitable for this study, i.e., for distinguishing between a sufficient n u m b e r of centromeres that replicated at different times. The fluorescence-tagged cells were studied using a Zeiss epifluorescent apparatus fitted with filter no. 16 (450-480 A range}. Fluorescent intensity of the minor satellite hybridized to the octacentric chromosome was studied using a confocal microscope, MRC 600 Confocal Scanning Laser Microscope (BioRad) attached to a Nikon Diaphot inverted microscope. Sequence of centromere separation was studied only for inactive centromeres, both on the heptacentrics as well as the dicentrics. All inactive centromeres in a given cell were analyzed for their having separated or not as seen u n d e r the light microscope. A line drawing representing the various centromeres along the length of the heptacentric was used to mark the positions of various centromeres and to designate whether a particular centromere in that chromosome h a d separated ( + ) or not ( - ). For BRdU-labeled cells, similar line drawings were u s e d to denote the presence/absence of the label. Photographs were made on black-and-white {Kodak TechPan, ASA 400, 8-10 second exposure) or color film (Kodak Ecktachrome, ASA 400, 45 second exposure). Giemsa-stained cells were photographed on a TechPan, 400 ASA, film using automatic exposure.
RESULTS
Location of Centromeres The L929 cell line has one heptacentric c h r o m o s o m e and one to three dicentrics. The heptacentric c h r o m o s o m e has three inactive centromeres in one arm and two in the other; the m i d d l e centromere is c o m p o s e d of two units and is the
Figure 2 Part of a mouse brain tumor cell showing the heptacentric chromosome and a dicentric chromosome. The DAPI-stained cell (a) shows heterochromatin blocks associated with inactive (small arrows) as well as active (large arrows) centromeres. Note the presence of two heterochromatin blocks in the heptacentric (open arrows). These are not associated with any centromere (cf Fig. 1). Upon in situ hybridization with the 42-mer partially homologous to the mouse minor satellite (b), all centromeres on the heptacentric and active centromere in dicentric exhibit the signal. The inactive centromere on the dicentric (small arrow in b) and non-centromeric heterochromatin blocks in the heptacentric do not show minor satellite DNA. F i g u r e 3 Confocal microscopy of the heptacentric chromosome reveals the intensity of the fluorescence due to FITC label used to detect the minor satellite DNA. Note the highest amount in the centromere in the middle, the active centromere. The next higher intensity is present in the regions of centromeres 1 and 5, which carry about equivalent amount. The lowest intensity is found in numbers 2 and 4. The peak points of the intensities for inactive centromeres are numbered; the active centromere is identified with an arrow. Other points of fluorescent intensity represent background fluorescence at various sites along the length of the heptacentric, x-axis = chromosome length marking the position of the various centromeres; y-axis = fluorescence intensity in relative arbitrary units.
34
B . K . Vig and N. Paweletz
functional centromere (Fig. 1). As previously suggested [9], this heptacentric chromosome is derived from an octacentric that appeared to be an isochromosome. The dicentrics have the active centromere located at one terminus and the inactive centromere about the middle of the chromosome (Fig. 1). All centromeres, active or not, are flanked by pericentric heterochromatin. Starting with the end of the long arm of the heptacentric chromosome, we shall designate the inactive centromeres as numbers 1, 2, 3 (in the long arm) and 4, 5 (in the short arm). This places centromeres 3 and 4 proximal to the active centromere, which is not numbered in this communication.
Sequence of Centromere Separation: Presence versus Absence of Satellite DNA The initial study dealt with comparing the frequency of separated centromeres carrying the satellite DNA with those that lacked it. The centromeres carrying satellite DNA should separate later than those that do not have this component, by virtue of late replication of this DNA. If satellite DNA is responsible for holding the active centromeres together it is expected to do so in the inactive centromeres also. We scored 134 cells from four different fixations over a 3-month period. There appeared to be no consensus that the centromeres bearing satellite DNA separated with frequency lower than those not carrying the satellite DNA. Even though these studies were not carried out with synchronous populations, our sample indicated that about as many inactive centromeres in the heptacentric had separated (58.85%) as had those in the dicentrics (60%). The data (Table 1) show the frequency of separated and unseparated inactive centromeres in the dicentric when none, 1, 2, 3, 4, or all 5 inactive centromeres in the heptacentric had separated. It is clear that some inactive centromeres in the dicentrics had separated when no centromere in the heptacentric had yet separated. Conversely, in some cells all inactive centromeres in the heptacentric had separated but some inactive centromeres in the dicentric had not yet separated (Fig. 4). In summary, the data show that the presence of minor satellite of mouse has no preferential control over the timing of separation of inactive centromeres.
Location of Minor Satellite In order to determine the distribution of minor satellite, in situ hybridization was carried out with the 42-mer oligonucleotide. In the octacentric, the distribution of the signal coincides with all centromeres, active as well as inactive. At occasions the central, active centromere expressed two coalescing or coinciding signals, indicating the presence of two units harboring the minor satellite. In dicentrics, however, the signal was confined only to the region of the active centromere with no indication of its being present at the site of the inactive centromere (Fig. 2). These results are similar to what has been observed using the entire 120-bp minor satellite segment. Relative Quantity of Minor Satellite In order to estimate the relative quantity of the minor satellite at various centromeres, we used the FITC-generated fluorescent intensity of the centromeric regions after in situ hybridization with the entire 120-bp repeat. The confocal microscopy used for this purpose assumes that the compaction of the various inactive centromeres is more or less equal. Under the light microscope, it seems to be so. Figure 3 shows that the highest amount of satellite DNA is present in the region of the active centromere. Centromeres 1, 3, and 5 carry lesser but, essentially, about similar quantities. The other two centromeres, numbers 2 and 4, carry the least amount of minor satellite but may be considered near equal in their intensity. These similarities and differences made it possible to study whether the timing of separation is correlated with the quantity of minor satellite. Table 1
Separation of Centromeres with Varying Quantifies of Satellite DNA If inactive centromeres with or without satellite DNA do not express preference in relative timing of separation, then do the centromeres carrying satellite DNA follow any hierarchy of separation? We addressed this question by studying the sequence of separation of the inactive centromeres in the heptacentric chromosome. Because centromeres I and 5 are thought to be the replicates of the same original centromere due to isochromosome formation, we compared the separation of these two centromeres. Among 72 cells analyzed, 50 showed the separation of centromere 5 having occurred. Centromere I had separated only in 36 cells. Centromere 2 ap-
Frequency of inactive centromeres expressing separation with regards to the presence of minor satellite DNA (heptacentrics) or its absence (dicentrics) Inactive centrorneres in the Heptacentric
Dicentric
Cells analyzed
No. separated per heptacentric
Total separated/unseparated
Total separated/unseparated
20 13 19 27 14 41 Total 134
0 1 2 3 4 5
0/100 13/52 38/57 81/54 56/14 205/0
37/14 21/7 28/13 42/24 19/8 33/89
393/277 (58.7%)
180/120 (60%)
Sequence of Centromere Separation
35
p e a r e d to separate in between these two frequencies having separated in 46 cells. Centromeres I and 5 carry about equivalent amounts of m i n o r satellite as d e t e r m i n e d by fluorescent intensity of the FITC tag. One w o u l d expect these two centromeres to separate w i t h equal frequency; however, this was not so (~2 = 2.2, p < 0.20). This might indicate that equivalent amounts of the DNA u n d e r consideration do not guarantee equivalent frequencies of separation of the bearer centromeres. Apparently, the a m o u n t of satellite DNA has no bearing on the timing of separation of these two centromeres. The other pair of centromeres w i t h almost similar quantities of satellite DNA, viz, centromeres 2 and 4, however, separated in about equal frequency; 46 for 2 and 43 for 4 (~2 = 1.2, p ~ 0.7). Considering the relative frequencies of separation of the four centromeres (1, 2, 4, and 5), we can c o n c l u d e that larger quantities of the satellite DNA do not delay the separation of centromeres, e.g., n u m b e r 5, w h i c h carries a larger amount, separates more frequently. Besides, w h e n we analyzed the data for the two earlier separating centromeres, numbers 1 and 5 were represented more frequently than numbers 2 and 4 (Table 2). Centromere 3, with an intermediate a m o u n t of satellite DNA, was only once among the first two centromeres to separate but was present most frequently among those that separated last. Centromere 1, nonetheless, was present most frequently among the two that separated last. W h e n we s t u d i e d w h i c h of the two centromeres with equivalent quantities of satellite DNA separated earlier than the other, we found no preference for one over the other; either centromere could separate ahead of its counterpart (Fig. 5a,b). In our sample, 20 of 72 cells showed centromere 5 to have separated ahead of centromere 1. This figure is slightly h i g h e r than 16 of 72, the frequency w i t h w h i c h centromere 1 separated earlier t h a n centromere 5. Centromeres 2 and 4 again showed equivalent frequencies in separating ahead of each other. Overall, these data indicate that the timing of separation of inactive centromeres is not strictly d e p e n d e n t u p o n the quantity of satellite DNA.
Replication of Centromeres with Equivalent Amounts of Minor Satellite The timing of r e p l i c a t i o n of inactive centromeres that separate at different times s h o u l d also be different. We therefore
Table 2
looked for the replication pattern of the various inactive centromeres. Following the fact that n o r m a l l y heterochromatin (made up of satellite DNA) is the last segment of the chrom o s o m e to replicate, the centromeres that carry the m i n o r satellite s h o u l d replicate later than those that do not have this fraction. Besides, the centromeres associated w i t h larger quantities of satellite DNA s h o u l d replicate later t h a n their counterparts associated w i t h lesser amounts of repetitive DNA. However, there appeared to be no relationship between the presence of the m i n o r satellite and delayed replication of that centromere in the BrdU-labeled population. Several dicentrics replicated in the region of the inactive centromere before the equivalent segments on the heptacentric did. The reverse was also true: inactive centromeres in the heptacentric showed no sign of BrdU incorporation in several cells in w h i c h the dicentrics h a d incorporated the t h y m i d i n e analogue at the site of inactive centromeres. The situation with satellite DNA-bearing centromeres was also similar w h e t h e r these carried similar or different quantities of the satellite DNA. Not only d i d the timing of replication of, for example, centromeres I and 2 exhibit indifference for earlier or later replication but also the centromeres with equivalent satellite DNA, e.g., numbers I and 5 differed from each other in this respect. For instance, in some cells centromere 1 a p p e a r e d to have replicated ahead of centromere 5 (Fig. 6a). In other instances the reverse was true; centromere 5 showed no BrdU incorporation w h e n centromere 1 d i d so (Fig. 6b). Thus the replicating of inactive centremeres, unlike those of the active centromeres, does not dep e n d u p o n the presence or absence of the satellite DNA nor does the quantity of satellite DNA dictate its timing of replication. In brief, these studies indicate that 1) inactive centromere may or may not carry m i n o r satellite DNA in mouse L929 cells; yet all appear to constitute bona fide primary constrictions; 2) the inactive centromeres separate ahead of the active centromeres but fail to show any sequential hierarchy whether or not these carry m i n o r satellite DNA; 3) there is no relationship between the quantity of m i n o r satellite DNA and the p o s i t i o n of separation of an inactive centromere in the sequence; 4) inactive centromeres harboring similar quantities of satellite DNA may separate at different times; and 5) the inactive centromeres w i t h similar quantities of
C o m p a r i s o n b e t w e e n various satellite DNA-bearing inactive centromeres on 72 heptacentric c h r o m o s o m e s regarding their frequency of separation
Centromere number
1
2
3
Times separated unseparated Times separated among the first two the last two First to separate compared to that with equivalent amount of minor satellite DNA centromere 1 vs. 5 centromere 2 vs. 4
36 36
46 26
11 22
4 5
° .. indicates the position "ofthe active centromere.
a
4
5
35 37
43 29
50 22
1 17
9 11
13 12
16
20 13
13
36
B . K . Vig a n d N. P a w e l e t z
F i g u r e 4 A mouse brain tumor cell showing the differences in the timing of separation of the inactive centromeres in the heptacentric chromosome (large arrow) carrying satellite DNA and the dicentrics (small arrows), w h i c h do not exhibit this DNA. Note that all inactive centromeres in the heptacentric have separated but some in the dicentrics are still held together. F i g u r e 5 Differential separation of centromeres 1 and 5 in the heptacentric. In (a) centromere 5 (large arrowhead) is held as one unsplit unit, whereas centromeres 1, 2, and 3 have separated. Centromere 4 is also unseparated (small
Sequence of Centromere Separation
m i n o r satellite DNA may replicate at different times in the cell cycle. DISCUSSION We have previously demonstrated that the centromeres in a given genome separate in a n o n r a n d o m , genetically controlled manner. This sequence appears to be controlled by, or d e p e n d e n t upon, the amount a n d base composition of the pericentric heterochromatin [7]. The inactive centromeres, however, separate before the cell reaches meta-anaphase stage, some time in prophase. But it is not k n o w n w h e t h e r the inactive centromeres also separate in a given sequence. This study reveals that the inactive centromeres do not follow the constraints defining the hierarchy that governs the sequence of separation of the active centromeres; the former neither separate in a specific sequence nor does the process appear to be controlled by the quantity of minor satellite DNA. The true cause for p r e m a t u r e separation of centromeres is still not clear; however, these do not associate w i t h s p i n d l e microtubules nor do they participate in migratory processes. This p h e n o m e n o n does raise the question whether active centromeres w o u l d also fail to function p r o p e r l y should, for some reason, they separate prematurely. The case of the hum a n X c h r o m o s o m e [2], however, points to this fact. One possible control for the sequential centromere separation may be the timing of replication of the centromere. It has been suggested that meta-anaphase separation of normal centromeres might reflect late replication of centromeres [1] followed by some sort of "maturation" before these separate into two sub-units. The inactive centromere appears to defy this concept of late replication. It has been shown that, in mouse, heterochromatin, and, hence, the satellite DNA associated w i t h active centromeres, replicates late. But this is not true for the inactive centromeres. Even t h o u g h the inactive centromeres are associated w i t h pericentric heterochromatin of similar base c o m p o s i t i o n as those found in the equivalent r e g i o n of the active centromeres, the pericentric heterochromatin in the former replicates in earlier part of S [8]. Even then it does not e x p l a i n why the early replicating satellite DNA w o u l d show no relationship between the timing of replication a n d the quantity of satellite DNA, as observed in the present study. W h e t h e r or not the inactive centromeres carry m i n o r satellite DNA appears to make no difference either in the timing of centromere separation or in the timing of replication of the centric region. Considering the constraints i m p o s e d on the active centromeres with regard to the quantity of
37 pericentric heterochromatin, we expected some sort of relationship between the quantity of centric/pericentric satellite DNA and the two parameters m e n t i o n e d above. In mouse, all chromosomes except the Y carry minor satellite DNA. The Y c h r o m o s o m e is also the first one in the genome to separate at its centromere [6]. If the centromere owes its property of being h e l d as one unit to the centromeric DNA, a property suggested for the m i n o r satellite DNA in mouse, then we expect the centromeres w i t h m i n o r satellite to separate later than those that do not carry this fraction. This, however, is not the case with inactive centromeres; these separate r a n d o m l y irrespective of the presence or absence of the m i n o r satellite fraction. Besides, the inactive centromeres even with similar quantity of satellite DNA separate at different times, e.g., centromere 1 versus 5. This would also be true of the total quantity of periceutric heterochromatin. Regardless of the a m o u n t of heterochromatin, w h i c h is constant for any given inactive centromere, these centromeres separate in a r a n d o m manner. Hence, the inactive centromeres do not obey the laws that govern the sequence of separation of active centromeres.
This study was carried out while one of us (B.K.V.)was on sabbatical leave from the University of Nevada, Reno. B.K.V. is grateful to Alexander vonHumboldt Stiftung for the fellowship. We thank Dr. Harry Scherthan for synthesis of the 42-mer and for carrying out in situ hybridization; Mr. Burt Richards and Ms. Karen McCoy for help with generating Figure 3; Mrs. M. Enulescu for help in tissue culture; and Mr. Gerald Withers with photography. Dr. D. Schroeter read the manuscript carefully and made constructive suggestions, for which we are thankful.
REFERENCES 1. Broccoli D, Paweletz N, Vig BK (1989): Sequence of centromere separation: characterization of multicentric chromosomes in a rat cell line. Chromosoma 98:13-22. 2. Fitzgerald (1989): Aberrant chromatid separation and aneuploidy. In: Mechanisms of Chromosome Distribution and Aneuploidy, MA Resnick, Vig BK, eds. Alan R Liss, New York, pp. 104-110. 3. Joseph A, Mitchell AR, Miller OJ (1989): Organization of mouse satellite DNA at centromeres. Exp Cell Res 183:494-500. 4. Kipling D, Ackford H, Taylor BA, Cooke HJ (1991): Mouse minor satellite DNA genetically maps to the centromere and is physically linked to the acrocentric telomere. Genomes 11:235-241. 5. Scherthan H, Kohler M, von Malsch K, Schweizer D (1992): Nonisotopic chromosomal in situ hybridization with double labelled
arrowhead). In this cell two dicentrics (small arrowheads) have separated at their inactive centromeres whereas the third dicentric is still not resolved into sub-units at its inactive centromere (intermediate-sized arrowhead). In (b) centromere 1 (thick arrow) is held as one unit while centromeres 3 and 5, with about equivalent satellite DNA, have resolved into two units. Note that the two dicentrics in this cell show no evidence of separation at their inactive centromeres (small arrowhead). F i g u r e 6 Parts of two cells labeled with BrdU during the last 5 hours of culture. In (a) the label is present in the region of centromere 5 (large arrow) but not in that of centromere 1 (small arrow), indicating that centromere 5 had replicated before the addition of BrdU. The reverse is seen in (b), where centromere 1 (large arrow) replicated later than centromere 5 (small arrow).
38 DNA: signal amplification at the probe level. Cytogenet Cell Genet 60:4-7. 6. Vig BK (1982): Sequence of centromere separation: role of centromeric heterochromatin. Genetics 102:795-806. 7. Vig BK (1987): Sequence of centromere separation: a possible role for repetitive DNA. Mutagenesis 2:155-159. 8. Vig BK, Broccoli D (1988): Sequence of centromere separation: differential replication of pericentric heterochromatin in multicentric chromosomes. Chromosoma 96:311-317.
B. K, Vig a n d N. P a w e l e t z
Vig BK, Richards B (1992): Formation of primary constrictions and heteroehromatin in mouse does not require minor satellite DNA. Exp Cell Res 201:292-298. 10. Vig BK, Sternes K, Paweletz N (1991): Centromere structure and function in neoplasia. Cancer Genet Cytogenet 43:151-178. 11. Wong AKC, Rattner JB (1988): Sequence organization and cytological localization of the minor satellite of mouse. Nucleic Acid Res 16:11645-11661.
ABSTRACT: Neoplastic cells may carry inactive centromeres on some multicentric, yet stable, chromosomes. We report that some inactive centromeres in L929 mouse cells do not contain minor satellite DNA, the DNA fraction which has been suggested to constitute the centromere. We compared the sequence of separation of inactive centromeres carrying the minor satellite with those lacking this fraction. The sequence of separation appears to be independent of whether or not the inactive centromeres carry the minor satellite DNA. The timing of replication of the inactive centrameres is also independent of this DNA. Hence, minor satellite of mouse is not a factor in holding together the subunits of inactive centromeres. Extension of these results to active centromeres might suggest that the minor satellite DNA is not a factor responsible for adhesion of the two centromere sub-units up until late meta-anaphase.
INTRODUCTION Several neoplastic cell lines carry stable dicentric and multicentric chromosomes. The centromeres of normal monocentric chromosomes in a genome separate into two sub-units in genetically controlled, non-random sequence at metaanaphase junction. However, in stable dicentric and multicentric chromosomes the inactive centromeres separate into two units in prophase, generally long before the active or functional centromeres do [10]. The sequence of separation of normal, functional centromeres has been studied for a number of species. This sequential separation seems to depend on the quantity and quality of pericentric heterochromatin [6]; hence, upon the timing of replication of the centromere [see [7]]. In species with qualitatively uniform heterochromatin, e.g., Mus musculus domesticus, the timing of separation is directly related to the quantity of pericentric heterochromatin; the larger the block of pericentric heterochromatin, the latter the separation of the associated centromere [6]. The inactive centromeres, though, separate early irrespective of the quantity of centromeric heterochromatin. This is apparently so because pericentric heterochromatin associated with inactive centromeres replicates earlier than the corresponding regions associated with the active centromeres, even when both are of similar base composi-
From the Department of Biology (B. K. V.), University of Nevada, Reno, Nevada; and Forschungsschwerpunkt 4 (N. P.), Deutsches Krebsforschungszentrum, Heidelberg, BRD. Address reprint requests to B. K. Vig, Department of Biology, University of Nevada, Reno NV 89557-0015. Received January 21, 1993; accepted June 1Z 1993.
© 1993 Elsevier Science P u b l i s h i n g Co., Inc. 655 A v e n u e of the Americas, N e w York, NY 10010
tion. For instance, in m o u s e the repetitive D N A associated with active centromeres replicates in late S but the D N A of the same composition w h e n associated with inactive cantromeres replicates in early S [1]. The minor satelliteof mouse is present in allchromosomes of the M. m. domesticus genome except in the Y chromosome. This D N A fraction has been implicated to be the centromere D N A [3, 4, 11].Ifso, then itis logical to assume that the minor satelliteis responsible for holding the two centromere subunits together up until meta-anaphase. Recently, we studied the distribution of minor satellite D N A in stable dicentric and multicentric chromosomes in L929 mouse cell line.W e reported that the inactivecentromeres in a heptacentric chromosome carry minor satelliteD N A , as do all active centromeres in that cell line.However, inactive centromeres in several dicentric chromosomes do not contain any detectable minor satellite[9].Hence, some primary constrictions can be formed without the participation of this fraction of DNA. Whereas the sequence of centromere separation is established for the active centromeres, it is not k n o w n whether inactive centromeres also follow a particular pattern and hierarchy of separation. Italso stands to reason that the minor satellitein mouse carries similar functions w h e n present in the inactivecentromeres as itdoes for the activecentromeres. Hence, those inactive centromeres that carry the minor satelliteshould separate laterthan those that do not carry this DNA. The present study was, therefore, carried out to find out whether there is a hierarchical sequence of separation among the inactive centromeres and whether the presence of minor satelliteD N A has any influence on the timing of separation of these centromeres.
31 Cancer Genet Cytogenet70:31-38(1993) 0165-4608/93/$06.00
32
J
140 120_~ 100c'(D
8o E
~o
60-
0
rr
40-
20 ! O
3
°
0
2.4
5.5 8.2 10.1 12.8 Chromosome length (microns)
15.2
Figure I A Giemsa-stained cellexhibitingthe heptacentricand a dicentricchromosome with allcentromeres held as unseparated units. The inactive centromeres are marked with small arrowheads, active centromeres with large arrowheads.
Sequence of Centromere Separation M A T E R I A I S AND METHODS L929 cells of mouse origin were u s e d in these studies. The cells were grown in McCoy's 5A m e d i u m w i t h 10% fetal calf serum. The cell line grows equally well w h e t h e r CO 2 is available or not. Before fixing in 3:1 methanol:acetic acid, the cultures were exposed to Colcemid (1 hour) and to hypotonic (0.075 M KCI, 10 minutes). The air-dried slides were stored at room temperature until used. For detecting heterochromatin the chromosomes were either C-banded or stained w i t h DAPI. C-banding was carried out by exposing the cells to 0.2 N HC1 (50 minutes), saturated Ba (OH)2 (6 minute) and reconstituting in 2 × SSC (2 hours at 65°C). The presence of m i n o r satellite DNA was confirmed by in situ hybridization using the entire 120-base (bp) repeats or w i t h a 42-met oligodeoxynucleotide, AACGT GTATA TCATA GAGTT ACAAT GAGAA ACATG G A A A A riG, homologous to a part of the mouse m i n o r satellite [11]. The details of results obtained with the 120-bp repeats are published elsew h e r e [9]. The 42-mer oligomer was synthesized on an app l i e d Biosystems 380a DNA synthesizer as previously reported [5]. Ten p m o l of oligomer were labeled w i t h 1 n m o l biotin-11-dUTP (Sigma)/1.5 n m o l dATP using 50 U terminal deoxynucleotidyltransfemse (Boehringer) in final volume of 30 B1 of reaction buffer s u p p l i e d by the manufacturer. After 3 h incubation at 37°C, the oligomers were ethanol precipitated in the presence of 20 ~g E. coli t-RNA (Boehringer). Finally, the d r i e d pellets was dissolved in tris-EDTA buffer. Prior to in situ hybridization, the cells were digested with RNase (100 ~g/ml PBS). These were, then, denatured in 70% formamide/2 x SSC at 70°C for 2 minutes and dehydrated in chilled methanol series. Five ~1 of the hybridization mixture (2 × SSC, 30% formamide, 0.5 ng/~l labeled oligomer DNA, 0.5 ~g/~l E. coli t-RNA) was a p p l i e d to the slides under a cover slip and sealed with rubber cement. After allowing overnight hybridization at 37°C, the slides were washed in 0.1 M NaHCO3/0.05% Tween 20 at 37°C. Biotin was detected using FITC avidin [Sigma) w i t h one r o u n d of signal amplification. Chromosomes were counterstained with a 1 ~l/ml of p r o p i d i u m i o d i d e in antifade (DABCO) solution. To detect DNA synthesis in the centromere region, async h r o n o u s l y growing cells were exposed to 10 .6 M 5-brom o d e o x y u r i d i n e (BrdU) during the last 5 hours of culturing. For the detection of the signal, the cells were treated w i t h anti-BrdU antibody for 30 minutes at 37°C. Followed
33 by washing in Dulbecco's phosphate buffer, the cells were reincubated for a d d i t i o n a l 30 minutes at 37°C in goat antimouse secondary antibody tagged with FITC. Chromosomes were counterstained w i t h 1:1000 dilution of e t h i d i u m bromide. Those centromeres that replicated before the a d d i t i o n of BrdU w o u l d not show the label; those replicating after the a d d i t i o n of the t h y m i d i n e analog w o u l d be labeled. This approach, however, would not d i s t i n g u i s h between all those centromeres that replicated before the a d d i t i o n of BrdU or between those that replicated after the a d d i t i o n of BrdU. Nonetheless, this treatment was found suitable for this study, i.e., for distinguishing between a sufficient n u m b e r of centromeres that replicated at different times. The fluorescence-tagged cells were studied using a Zeiss epifluorescent apparatus fitted with filter no. 16 (450-480 A range}. Fluorescent intensity of the minor satellite hybridized to the octacentric chromosome was studied using a confocal microscope, MRC 600 Confocal Scanning Laser Microscope (BioRad) attached to a Nikon Diaphot inverted microscope. Sequence of centromere separation was studied only for inactive centromeres, both on the heptacentrics as well as the dicentrics. All inactive centromeres in a given cell were analyzed for their having separated or not as seen u n d e r the light microscope. A line drawing representing the various centromeres along the length of the heptacentric was used to mark the positions of various centromeres and to designate whether a particular centromere in that chromosome h a d separated ( + ) or not ( - ). For BRdU-labeled cells, similar line drawings were u s e d to denote the presence/absence of the label. Photographs were made on black-and-white {Kodak TechPan, ASA 400, 8-10 second exposure) or color film (Kodak Ecktachrome, ASA 400, 45 second exposure). Giemsa-stained cells were photographed on a TechPan, 400 ASA, film using automatic exposure.
RESULTS
Location of Centromeres The L929 cell line has one heptacentric c h r o m o s o m e and one to three dicentrics. The heptacentric c h r o m o s o m e has three inactive centromeres in one arm and two in the other; the m i d d l e centromere is c o m p o s e d of two units and is the
Figure 2 Part of a mouse brain tumor cell showing the heptacentric chromosome and a dicentric chromosome. The DAPI-stained cell (a) shows heterochromatin blocks associated with inactive (small arrows) as well as active (large arrows) centromeres. Note the presence of two heterochromatin blocks in the heptacentric (open arrows). These are not associated with any centromere (cf Fig. 1). Upon in situ hybridization with the 42-mer partially homologous to the mouse minor satellite (b), all centromeres on the heptacentric and active centromere in dicentric exhibit the signal. The inactive centromere on the dicentric (small arrow in b) and non-centromeric heterochromatin blocks in the heptacentric do not show minor satellite DNA. F i g u r e 3 Confocal microscopy of the heptacentric chromosome reveals the intensity of the fluorescence due to FITC label used to detect the minor satellite DNA. Note the highest amount in the centromere in the middle, the active centromere. The next higher intensity is present in the regions of centromeres 1 and 5, which carry about equivalent amount. The lowest intensity is found in numbers 2 and 4. The peak points of the intensities for inactive centromeres are numbered; the active centromere is identified with an arrow. Other points of fluorescent intensity represent background fluorescence at various sites along the length of the heptacentric, x-axis = chromosome length marking the position of the various centromeres; y-axis = fluorescence intensity in relative arbitrary units.
34
B . K . Vig and N. Paweletz
functional centromere (Fig. 1). As previously suggested [9], this heptacentric chromosome is derived from an octacentric that appeared to be an isochromosome. The dicentrics have the active centromere located at one terminus and the inactive centromere about the middle of the chromosome (Fig. 1). All centromeres, active or not, are flanked by pericentric heterochromatin. Starting with the end of the long arm of the heptacentric chromosome, we shall designate the inactive centromeres as numbers 1, 2, 3 (in the long arm) and 4, 5 (in the short arm). This places centromeres 3 and 4 proximal to the active centromere, which is not numbered in this communication.
Sequence of Centromere Separation: Presence versus Absence of Satellite DNA The initial study dealt with comparing the frequency of separated centromeres carrying the satellite DNA with those that lacked it. The centromeres carrying satellite DNA should separate later than those that do not have this component, by virtue of late replication of this DNA. If satellite DNA is responsible for holding the active centromeres together it is expected to do so in the inactive centromeres also. We scored 134 cells from four different fixations over a 3-month period. There appeared to be no consensus that the centromeres bearing satellite DNA separated with frequency lower than those not carrying the satellite DNA. Even though these studies were not carried out with synchronous populations, our sample indicated that about as many inactive centromeres in the heptacentric had separated (58.85%) as had those in the dicentrics (60%). The data (Table 1) show the frequency of separated and unseparated inactive centromeres in the dicentric when none, 1, 2, 3, 4, or all 5 inactive centromeres in the heptacentric had separated. It is clear that some inactive centromeres in the dicentrics had separated when no centromere in the heptacentric had yet separated. Conversely, in some cells all inactive centromeres in the heptacentric had separated but some inactive centromeres in the dicentric had not yet separated (Fig. 4). In summary, the data show that the presence of minor satellite of mouse has no preferential control over the timing of separation of inactive centromeres.
Location of Minor Satellite In order to determine the distribution of minor satellite, in situ hybridization was carried out with the 42-mer oligonucleotide. In the octacentric, the distribution of the signal coincides with all centromeres, active as well as inactive. At occasions the central, active centromere expressed two coalescing or coinciding signals, indicating the presence of two units harboring the minor satellite. In dicentrics, however, the signal was confined only to the region of the active centromere with no indication of its being present at the site of the inactive centromere (Fig. 2). These results are similar to what has been observed using the entire 120-bp minor satellite segment. Relative Quantity of Minor Satellite In order to estimate the relative quantity of the minor satellite at various centromeres, we used the FITC-generated fluorescent intensity of the centromeric regions after in situ hybridization with the entire 120-bp repeat. The confocal microscopy used for this purpose assumes that the compaction of the various inactive centromeres is more or less equal. Under the light microscope, it seems to be so. Figure 3 shows that the highest amount of satellite DNA is present in the region of the active centromere. Centromeres 1, 3, and 5 carry lesser but, essentially, about similar quantities. The other two centromeres, numbers 2 and 4, carry the least amount of minor satellite but may be considered near equal in their intensity. These similarities and differences made it possible to study whether the timing of separation is correlated with the quantity of minor satellite. Table 1
Separation of Centromeres with Varying Quantifies of Satellite DNA If inactive centromeres with or without satellite DNA do not express preference in relative timing of separation, then do the centromeres carrying satellite DNA follow any hierarchy of separation? We addressed this question by studying the sequence of separation of the inactive centromeres in the heptacentric chromosome. Because centromeres I and 5 are thought to be the replicates of the same original centromere due to isochromosome formation, we compared the separation of these two centromeres. Among 72 cells analyzed, 50 showed the separation of centromere 5 having occurred. Centromere I had separated only in 36 cells. Centromere 2 ap-
Frequency of inactive centromeres expressing separation with regards to the presence of minor satellite DNA (heptacentrics) or its absence (dicentrics) Inactive centrorneres in the Heptacentric
Dicentric
Cells analyzed
No. separated per heptacentric
Total separated/unseparated
Total separated/unseparated
20 13 19 27 14 41 Total 134
0 1 2 3 4 5
0/100 13/52 38/57 81/54 56/14 205/0
37/14 21/7 28/13 42/24 19/8 33/89
393/277 (58.7%)
180/120 (60%)
Sequence of Centromere Separation
35
p e a r e d to separate in between these two frequencies having separated in 46 cells. Centromeres I and 5 carry about equivalent amounts of m i n o r satellite as d e t e r m i n e d by fluorescent intensity of the FITC tag. One w o u l d expect these two centromeres to separate w i t h equal frequency; however, this was not so (~2 = 2.2, p < 0.20). This might indicate that equivalent amounts of the DNA u n d e r consideration do not guarantee equivalent frequencies of separation of the bearer centromeres. Apparently, the a m o u n t of satellite DNA has no bearing on the timing of separation of these two centromeres. The other pair of centromeres w i t h almost similar quantities of satellite DNA, viz, centromeres 2 and 4, however, separated in about equal frequency; 46 for 2 and 43 for 4 (~2 = 1.2, p ~ 0.7). Considering the relative frequencies of separation of the four centromeres (1, 2, 4, and 5), we can c o n c l u d e that larger quantities of the satellite DNA do not delay the separation of centromeres, e.g., n u m b e r 5, w h i c h carries a larger amount, separates more frequently. Besides, w h e n we analyzed the data for the two earlier separating centromeres, numbers 1 and 5 were represented more frequently than numbers 2 and 4 (Table 2). Centromere 3, with an intermediate a m o u n t of satellite DNA, was only once among the first two centromeres to separate but was present most frequently among those that separated last. Centromere 1, nonetheless, was present most frequently among the two that separated last. W h e n we s t u d i e d w h i c h of the two centromeres with equivalent quantities of satellite DNA separated earlier than the other, we found no preference for one over the other; either centromere could separate ahead of its counterpart (Fig. 5a,b). In our sample, 20 of 72 cells showed centromere 5 to have separated ahead of centromere 1. This figure is slightly h i g h e r than 16 of 72, the frequency w i t h w h i c h centromere 1 separated earlier t h a n centromere 5. Centromeres 2 and 4 again showed equivalent frequencies in separating ahead of each other. Overall, these data indicate that the timing of separation of inactive centromeres is not strictly d e p e n d e n t u p o n the quantity of satellite DNA.
Replication of Centromeres with Equivalent Amounts of Minor Satellite The timing of r e p l i c a t i o n of inactive centromeres that separate at different times s h o u l d also be different. We therefore
Table 2
looked for the replication pattern of the various inactive centromeres. Following the fact that n o r m a l l y heterochromatin (made up of satellite DNA) is the last segment of the chrom o s o m e to replicate, the centromeres that carry the m i n o r satellite s h o u l d replicate later than those that do not have this fraction. Besides, the centromeres associated w i t h larger quantities of satellite DNA s h o u l d replicate later t h a n their counterparts associated w i t h lesser amounts of repetitive DNA. However, there appeared to be no relationship between the presence of the m i n o r satellite and delayed replication of that centromere in the BrdU-labeled population. Several dicentrics replicated in the region of the inactive centromere before the equivalent segments on the heptacentric did. The reverse was also true: inactive centromeres in the heptacentric showed no sign of BrdU incorporation in several cells in w h i c h the dicentrics h a d incorporated the t h y m i d i n e analogue at the site of inactive centromeres. The situation with satellite DNA-bearing centromeres was also similar w h e t h e r these carried similar or different quantities of the satellite DNA. Not only d i d the timing of replication of, for example, centromeres I and 2 exhibit indifference for earlier or later replication but also the centromeres with equivalent satellite DNA, e.g., numbers I and 5 differed from each other in this respect. For instance, in some cells centromere 1 a p p e a r e d to have replicated ahead of centromere 5 (Fig. 6a). In other instances the reverse was true; centromere 5 showed no BrdU incorporation w h e n centromere 1 d i d so (Fig. 6b). Thus the replicating of inactive centremeres, unlike those of the active centromeres, does not dep e n d u p o n the presence or absence of the satellite DNA nor does the quantity of satellite DNA dictate its timing of replication. In brief, these studies indicate that 1) inactive centromere may or may not carry m i n o r satellite DNA in mouse L929 cells; yet all appear to constitute bona fide primary constrictions; 2) the inactive centromeres separate ahead of the active centromeres but fail to show any sequential hierarchy whether or not these carry m i n o r satellite DNA; 3) there is no relationship between the quantity of m i n o r satellite DNA and the p o s i t i o n of separation of an inactive centromere in the sequence; 4) inactive centromeres harboring similar quantities of satellite DNA may separate at different times; and 5) the inactive centromeres w i t h similar quantities of
C o m p a r i s o n b e t w e e n various satellite DNA-bearing inactive centromeres on 72 heptacentric c h r o m o s o m e s regarding their frequency of separation
Centromere number
1
2
3
Times separated unseparated Times separated among the first two the last two First to separate compared to that with equivalent amount of minor satellite DNA centromere 1 vs. 5 centromere 2 vs. 4
36 36
46 26
11 22
4 5
° .. indicates the position "ofthe active centromere.
a
4
5
35 37
43 29
50 22
1 17
9 11
13 12
16
20 13
13
36
B . K . Vig a n d N. P a w e l e t z
F i g u r e 4 A mouse brain tumor cell showing the differences in the timing of separation of the inactive centromeres in the heptacentric chromosome (large arrow) carrying satellite DNA and the dicentrics (small arrows), w h i c h do not exhibit this DNA. Note that all inactive centromeres in the heptacentric have separated but some in the dicentrics are still held together. F i g u r e 5 Differential separation of centromeres 1 and 5 in the heptacentric. In (a) centromere 5 (large arrowhead) is held as one unsplit unit, whereas centromeres 1, 2, and 3 have separated. Centromere 4 is also unseparated (small
Sequence of Centromere Separation
m i n o r satellite DNA may replicate at different times in the cell cycle. DISCUSSION We have previously demonstrated that the centromeres in a given genome separate in a n o n r a n d o m , genetically controlled manner. This sequence appears to be controlled by, or d e p e n d e n t upon, the amount a n d base composition of the pericentric heterochromatin [7]. The inactive centromeres, however, separate before the cell reaches meta-anaphase stage, some time in prophase. But it is not k n o w n w h e t h e r the inactive centromeres also separate in a given sequence. This study reveals that the inactive centromeres do not follow the constraints defining the hierarchy that governs the sequence of separation of the active centromeres; the former neither separate in a specific sequence nor does the process appear to be controlled by the quantity of minor satellite DNA. The true cause for p r e m a t u r e separation of centromeres is still not clear; however, these do not associate w i t h s p i n d l e microtubules nor do they participate in migratory processes. This p h e n o m e n o n does raise the question whether active centromeres w o u l d also fail to function p r o p e r l y should, for some reason, they separate prematurely. The case of the hum a n X c h r o m o s o m e [2], however, points to this fact. One possible control for the sequential centromere separation may be the timing of replication of the centromere. It has been suggested that meta-anaphase separation of normal centromeres might reflect late replication of centromeres [1] followed by some sort of "maturation" before these separate into two sub-units. The inactive centromere appears to defy this concept of late replication. It has been shown that, in mouse, heterochromatin, and, hence, the satellite DNA associated w i t h active centromeres, replicates late. But this is not true for the inactive centromeres. Even t h o u g h the inactive centromeres are associated w i t h pericentric heterochromatin of similar base c o m p o s i t i o n as those found in the equivalent r e g i o n of the active centromeres, the pericentric heterochromatin in the former replicates in earlier part of S [8]. Even then it does not e x p l a i n why the early replicating satellite DNA w o u l d show no relationship between the timing of replication a n d the quantity of satellite DNA, as observed in the present study. W h e t h e r or not the inactive centromeres carry m i n o r satellite DNA appears to make no difference either in the timing of centromere separation or in the timing of replication of the centric region. Considering the constraints i m p o s e d on the active centromeres with regard to the quantity of
37 pericentric heterochromatin, we expected some sort of relationship between the quantity of centric/pericentric satellite DNA and the two parameters m e n t i o n e d above. In mouse, all chromosomes except the Y carry minor satellite DNA. The Y c h r o m o s o m e is also the first one in the genome to separate at its centromere [6]. If the centromere owes its property of being h e l d as one unit to the centromeric DNA, a property suggested for the m i n o r satellite DNA in mouse, then we expect the centromeres w i t h m i n o r satellite to separate later than those that do not carry this fraction. This, however, is not the case with inactive centromeres; these separate r a n d o m l y irrespective of the presence or absence of the m i n o r satellite fraction. Besides, the inactive centromeres even with similar quantity of satellite DNA separate at different times, e.g., centromere 1 versus 5. This would also be true of the total quantity of periceutric heterochromatin. Regardless of the a m o u n t of heterochromatin, w h i c h is constant for any given inactive centromere, these centromeres separate in a r a n d o m manner. Hence, the inactive centromeres do not obey the laws that govern the sequence of separation of active centromeres.
This study was carried out while one of us (B.K.V.)was on sabbatical leave from the University of Nevada, Reno. B.K.V. is grateful to Alexander vonHumboldt Stiftung for the fellowship. We thank Dr. Harry Scherthan for synthesis of the 42-mer and for carrying out in situ hybridization; Mr. Burt Richards and Ms. Karen McCoy for help with generating Figure 3; Mrs. M. Enulescu for help in tissue culture; and Mr. Gerald Withers with photography. Dr. D. Schroeter read the manuscript carefully and made constructive suggestions, for which we are thankful.
REFERENCES 1. Broccoli D, Paweletz N, Vig BK (1989): Sequence of centromere separation: characterization of multicentric chromosomes in a rat cell line. Chromosoma 98:13-22. 2. Fitzgerald (1989): Aberrant chromatid separation and aneuploidy. In: Mechanisms of Chromosome Distribution and Aneuploidy, MA Resnick, Vig BK, eds. Alan R Liss, New York, pp. 104-110. 3. Joseph A, Mitchell AR, Miller OJ (1989): Organization of mouse satellite DNA at centromeres. Exp Cell Res 183:494-500. 4. Kipling D, Ackford H, Taylor BA, Cooke HJ (1991): Mouse minor satellite DNA genetically maps to the centromere and is physically linked to the acrocentric telomere. Genomes 11:235-241. 5. Scherthan H, Kohler M, von Malsch K, Schweizer D (1992): Nonisotopic chromosomal in situ hybridization with double labelled
arrowhead). In this cell two dicentrics (small arrowheads) have separated at their inactive centromeres whereas the third dicentric is still not resolved into sub-units at its inactive centromere (intermediate-sized arrowhead). In (b) centromere 1 (thick arrow) is held as one unit while centromeres 3 and 5, with about equivalent satellite DNA, have resolved into two units. Note that the two dicentrics in this cell show no evidence of separation at their inactive centromeres (small arrowhead). F i g u r e 6 Parts of two cells labeled with BrdU during the last 5 hours of culture. In (a) the label is present in the region of centromere 5 (large arrow) but not in that of centromere 1 (small arrow), indicating that centromere 5 had replicated before the addition of BrdU. The reverse is seen in (b), where centromere 1 (large arrow) replicated later than centromere 5 (small arrow).
38 DNA: signal amplification at the probe level. Cytogenet Cell Genet 60:4-7. 6. Vig BK (1982): Sequence of centromere separation: role of centromeric heterochromatin. Genetics 102:795-806. 7. Vig BK (1987): Sequence of centromere separation: a possible role for repetitive DNA. Mutagenesis 2:155-159. 8. Vig BK, Broccoli D (1988): Sequence of centromere separation: differential replication of pericentric heterochromatin in multicentric chromosomes. Chromosoma 96:311-317.
B. K, Vig a n d N. P a w e l e t z
Vig BK, Richards B (1992): Formation of primary constrictions and heteroehromatin in mouse does not require minor satellite DNA. Exp Cell Res 201:292-298. 10. Vig BK, Sternes K, Paweletz N (1991): Centromere structure and function in neoplasia. Cancer Genet Cytogenet 43:151-178. 11. Wong AKC, Rattner JB (1988): Sequence organization and cytological localization of the minor satellite of mouse. Nucleic Acid Res 16:11645-11661.