The EMBO Journal vol.10 no.5 pp.1225- 1236, 1991
The maternally expressed Drosophila gene encoding the
chromatin-binding protein BJ1 is a homolog of the
vertebrate gene Regulator of Chromatin Condensation,
Max-Planck-Institut fur Entwicklungsbiologie, Abt.
Communicated by C.Nusslein-Volhard
1 and 3, D-7400
Using monoclonal antibodies I have identified a nuclear
protein ofDrosophila, BJ1 (Mr
gene. Biochemical analysis demonstrates that the BJ1
protein is associated with nucleosomes and is released
from chromatin by agents which intercalate into DNA,
as previously shown for the high mobility group proteins
(HMGs). On polytene chromosomes the protein
localized in all bands, with no preference for particular
loci. Both the BJ1 protein and in particular the BJ1
mRNA are strongly expressed maternally. In early
embryos all nuclei contain equal amounts of BJ1. During
neuroblast formation, BJ1 mRNA becomes restricted to
cells of the central nervous system, and higher protein
levels are found in the nuclei of this tissue. In late
the mRNA almost completely
disappears, but significant amounts ofBJ1 protein persist
until morphogenesis. The BJ1 gene encodes a 547 amino
acid polypeptide featuring two different types of internal
repeats. The sequence from amino acids 46 to 417
containing seven repeats ofthe firt type has been highly
conserved in evolution. 45% of the amino acids in this
region are conserved in seven si'milar tandem repeats of
the human gene Regulator of Chromatin Condensation,
RCCI. The phenotype of a cell line carrying a mutation
of RCCI suggested a main function for this gene in cell
cycle control. A yeast gene, SRMI/PRP20, also contains
these repeats and shows 30% amino acid identity to BJ1
in this region. Mutations in this gene perturb mRNA
metabolism, disrupt nuclear structure and alter the signal
for the mating pheromone.
Complementation experiments argue for a common
function ofthese genes in the different species. I propose
that their gene products bind to the chromatin to establish
or maintain a proper higher order structure as a
prerequisite for a regulated gene expression. Disruption
of this structure could cause both mis-expression and
default repression of genes, which might explain the
pleiotropic phenotypes of the mutants.
Key words: cell cycle control/chromatin structure/non-
68 kd), and isolated its
In nuclei of eukaryotic cells the DNA is complexed with
proteins to form a compact structure, the chromatin. The
basic unit of the chromatin is the nucleosome, which contains
an octameric core of four different histones with
ofDNA wrapped around them. The nucleosomes can form
Oxford University Press
higher order structures, resulting in variable degrees of
chromatin condensation along the interphase chromosomes.
It appears that not only the presence of specific transcription
factors (activators or repressors), but also the degree of
chromatin condensation may play a decisive role in the
expression or repression of a gene (Weinstraub and
Groudine, 1976; Wu et al., 1979; reviewed by Felsenfeld
and McGhee, 1986; Widom, 1989). Thus, inactivation of
genes can occur when these are brought into the vicinity of
rearrangements (reviewed by Henikoff, 1990). In normal
development particular states of expression can be stably
maintained even after the disappearance of regulatory
proteins that are required to establish expression or
repression initially. It has been proposed that this imprinting
ofgenetic activity occurs via the chromatin structure (Paro,
Many data have accumulated pointing to the role of both
the histones and the non-histone proteins in determining the
higher order chromatin structure of genomic regions. In
particular, the presence of histones HI, Hio, or H5, which
bind to the linker regions between the nucleosomes, can
induce assembly ofmore condensed structures and repression
of genes (Schlissel and Brown, 1984; Sun et al., 1989;
Wolffe, 1989). Of the non-histone proteins influencing
chromatin structure, the class of the moderately abundant
high mobility group proteins (HMGs) has been studied in
most detail. The term HMG proteins has been operationally
defined according to biochemical properties that facilitate
isolation of these proteins (reviewed in Johns, 1983). While
HMG14 and HMG17 are preferentially associated with
decondensed, transcriptionally active chromatin (Weisbrod
et al., 1980; Dorbic and Wittig, 1987), other HMGs such
as the a-protein are thought to induce specific positioning
of nucleosomes and higher order chromatin structures
(Strauss and Varshavsky, 1984). However, the function of
these proteins is far from being clear (Einck and Bustin,
1985). Presumably there exist additional proteins influencing
chromatin structure which have not been isolated because
of their lower abundance or because their biochemical
properties differ from those of the HMGs.
Monoclonal antibodies have been successfully used to
dissect the complexity of nuclear protein fractions and to
analyze the functions of their individual components
(Saumweber et al., 1980. Dreyer et al., 1981; Kuo et al.,
1982; Lacroix et al., 1985; Garzino et al., 1987). In
Drosophila, this approach has allowed the identification of
nuclear proteins involved in gene expression (Risau et al.,
Frasch and Saumweber,
condensation (James and Elgin, 1986; Eissenberg etal.,
1990), and in structural functions of the nucleus (Fuchs
et al., 1983; Frasch et al ., 1988). In such an immunological
screen, I obtained monoclonal antibodies against39 different
nuclear proteins (Frasch,
obtained class of antibodies recognizeda 68 kdantigenwith
1985). The most frequently
moderate abundance in embryonic nuclei, called BJ1. Here,
I describe the chromatin binding of BJ1, its expression
pattern, and the cloning and sequencing of the BJ1 gene.
The sequence of BJ1 turned out to be strongly conserved
in evolution, and genes coding for related proteins are found
from yeast to humans. A homologous gene in vertebrates,
Regulator ofChromatin Condensation (RCCJ), was studied
most extensively, and its main function appears to be in cell
cycle control (reviewed in Nishimoto, 1988). In a hamster
cell line with a temperature sensitive mutation for RCCJ,
premature initiation of mitosis occurs when the cells are
shifted to the restrictive temperature during S phase (Ajiro
et al., 1983; H.Nishitani, M.Ohtsubo, K.Yamashita, H.Iida,
J.Pines, H.Yasuda, Y.Shibata, T.Hunter and T.Nishimoto,
submitted). However, temperature shifts inGIphase result
in a different phenotype, and cause alterations in gene
expression (Nishimoto et al., 1981). In yeast, the phenotypes
ofmutations in a homologous gene, called SRMI (Clark and
Sprague, 1989) or PRP20 (Aebi et al., 1990), are also com-
patible with the idea that the gene product is required for
normal gene expression and chromatin structure. The pro-
perties of the Drosophila BJ1 protein provide important clues
to the function(s) of these proteins in the nucleus and may
explain the pleiotropic phenotypes observed in these mutants.
Monoclonal antibodies against BJ 1
Monoclonal antibodies were produced against protein
fractions from Drosophila embryonic nuclei. Forty one
independent hybridoma clones obtained from one particular
protein fraction (see Materials and methods) secreted anti-
bodies against nuclear proteins. On Western blots, 29 of
these antibodies recognized an antigen with an electrophoretic
mobility of 68 kd present in total nuclear proteins from
embryos (Figure IA, lane 3) and from Drosophila tissue
culture cells (KC cells, Figure IA, lanes 4-7). Several lines
ofevidence suggested that the antibodies recognized at least
four different epitopes in this 68 kd polypeptide, which I
call BJ1. The four classes of antibodies were represented
by the antibodies BjlO, Bj43, Bj59, and Bj7O. These
antibodies reacted with different peptides after partial
digestion ofBJ1 with V8 protease (data not shown; Cleveland
et al., 1977). BJ1/3Gal fusion proteins of different length
allowed the partial mapping of the binding regions of these
antibodies within the BJ1 polypeptide. The antibodies also
differed in terms of their crossreaction with proteins from
other species (see below). Furthermore, with nuclear proteins
from Kc cells they recognized minor bands in addition to
the 68 kd band in four different patterns (Figure IA).
Interestingly, the 60 kd band recognized by the Bj7O
antibody (Figure lA, lane 7) appears to be identical to the
nuclear protein DI described previously (Alfageme et al.,
1980; Levinger and Varshavsky, 1982a,b). The Dl protein
is quantitatively extracted from nuclei with 5% perchloric
acid (PCA), together with only few other proteins including
histone Hi and A13 (Bassuk and Mayfield, 1982. Figure
iB, lanes 2 and 4). In a Western blot with the Bj7O antibody,
a 60 kd band corresponding to DI was stained with the PCA
extract, whereas BJ1 (68 kd) remained in the insoluble
fraction (Figure 1B, lanes 6 and 7). Thus, the nuclear
proteins BJ1 and D1 ofDrosophila are antigenically related.
Two classes of Bj antibodies also crossreacted with
proteins from species that are very distantly related to
2 3 4 5 6
Fig. 1. Characterization of BJ1 and related proteins by Western
blotting. A. Western analysis of Drosophila nuclear proteins with four
different antibodies against BJ1. Lanes 1 and 2:total nuclear proteins
from embyros or KC cells, respectively, stained with Coomassie
brilliant blue. Lane 3: Western blot of embryonic nuclear proteins, as
in lane 1, with Bj43 antibody. Lanes 4-6: Western blot of Kc cell
nuclear proteins, as in lane 2, with the antibodies Bj43 (lane 4), BjlO
(lane 5), Bj59 (lane 6) and Bj7O (lane 7) B. Extraction and Western
analysis of DI protein. Lanes 1-4: Coomassie stainings of total
nuclear proteins form KCcells (lane 1), 5% perchloric acid extract
enriched for Dl (lane 2 with 3Atgof protein loaded, lane 4 with
30Ag),and residual nuclear proteins after PCA extraction (lane 3).
Lanes 5-6: Western analysis of immobilized proteins as in lanes 1-3
with Bj7O antibody. C. Cross reaction of BJ1 antibodies with proteins
from nuclei of embryonic chicken (lanes 1 and 4), with proteins from
calf liver nuclei (lanes 2 and 5) and with proteins from total yeast cells
(lanes 3 and 6). Lanes 1-3 were stained with Bj59 antibody, lanes
4-6 with Bj7O antibody.
Drosophila. On Western blots with nuclear proteins from
embryonic chicken brains and calf liver, bands of55-64 kd
were stained with the Bj59 class, and two bands of-80 kd
and 90 kd with the Bj7O antibody class (Figure IC, lanes
1, 2, 4 and 5). These antibodies exclusively stained the nuclei
on tissue sections from chicken embryos (data not shown).
On Western blots with total proteins from yeast, two bands
were recognized by the Bj59 antibody class (69 kd and
84 kd), and a weak band by the Bj7O antibody class (40 kd).
These results suggest that the epitopes recognized by these
antibodies have been conserved in several nuclear proteins
Chromatin-binding of BJ1
Since BJ1 was found to be antigenically related to the HMG-
like protein Dl, I examined if it shared any biochemical
properties with HMG proteins. In particular, I tested whether
BJ1 (and Dl) is specificallv released from nuclei upon
alteration ofthe DNA conformation by intercalation. 'Elutive
intercalation' has been used previously to release HMG-like
proteins and other DNA-binding proteins from whole nuclei.
The composition of the eluted protein fraction depended on
the nature and concentration of the intercalating agent used,
and on the ionic strength (Schroter et al., 1985, 1987;
Schulman et al., 1987).
When nuclei from Kc cells were treated with 7.5 mM
ethidium bromide and 50 mM NaCl, the eluted protein
contained one major band with a mobility of 68 kd (Figure
2A, lane 3). This band corresponded to the BJl protein, as
shown by western analysis (Figure 2B, lane 3) and by co-
purification of the antigenic activity with the 68 kd band,
when the extract was further fractionated on an FPLC
MonoS column (data not shown). At a slightly higher ionic
100 mM NaCl, BJI was eluted even more
Expression, localization and chromatin-binding of BJ 1 protein
Fig. 2. Chromatin-binding and elution of BJ1. A-C. Intercalative elution of BJl and Dl from nuclei. A. Lanes 1-4: Coomassie staining of KC
nuclear proteins (lane 1), proteins eluted with buffer X2, 7.5 mM ethidium bromide (EtBr, lane 2), proteins eluted with buffer X2, 7.5 mM EtBr,
50 mM NaCl (lane 3), and proteins eluted with buffer X2, 7.5 mM EtBr, 100 mM NaCl (lane 4). B. Western analysis of immobilized proteins as in
A, lanes 1-4, with Bj43 antibody (antiBJ1). C. Western analysis as in B, but with BmlO antibody (antiDl, Frasch et al., 1986). D. Salt extraction
of BJI and DI from nuclei. Lanes 1-7: Coomassie blue stained proteins from total Kc cell nuclei (lane 1), proteins successively extracted with
buffer X2, 0 mM NaCl (lane 2), 50 mM NaCl (lane 3),
NaCl (lane 7). E. Western analysis of proteins as in D with Bj43 antibody (antiBJI; upper bands at 68 kd) and, in a second step, with BmlO
antibody (antiDl; lower bands at 60 kd). F-H. Sedimentation profiles of antigens on chromatin gradients. Top fractions of the 10-30% sucrose
gradient are to the left and bottom fractions are to the right. Chromatin separation was recorded measuring OD254 (ordinate scale right); the antigen
distribution was analyzed by ELISA (left scale). F. Analysis of gradient fractions with Bj43 antibody (antiBJI, 0), G. Analysis of gradient fractions
with BX65 (anti-histone H2A, 0) with P11 (antiRNP,
NaCl (0), 250 mM NaCl (U), and 450 mM NaCl. (O). The S values were determined using a Fortran program written by R.Burberg.
100 mM NaCl (lane 4), 250 mM NaCl (lane 5), 300 mM NaCl (lane 6), and 350 mM
). H. Sedimentation profiles of chromatin and BJ1 in sucrose gradients containing 125 mM
efficiently by ethidium bromide (Figure 2A, B, lane 4). As
judged by Western analysis, >70% of the BJ
eluted from the nuclei under these conditions. At 7.5 mM
ethidium bromide and 100 mM NaCl, a second prominent
band at 60 kd appeared in the eluate (Figure 2A, lane 4).
Western analysis with a monoclonal antibody specific for
DI (Frasch et al., 1986; Figure 2C, lane 4) and FPLC
fractionation demonstrated that this band corresponded to
the Dl protein. In the absence of ethidium bromide, neither
BJ1 nor DI were released under these ionic strengths (Figure
2D,E, lanes 3,4). BJ1 was only released, among many other
proteins, at 300 mM NaCl, and Dl at 350 mM NaCl (Figure
2D,E, lanes 6 and 7). The specific elution of BJ1 upon
treatment of nuclei with ethidium bromide, similar to that
of the DNA-binding protein Dl (Levingerand Varshavsky,
1982a,b, Levinger, 1985), indicates that BJ1 is also a DNA-
In a different assay, I tested whether BJ 1 was associated
with nucleosomes. After RNase A treatment of KC nuclei
and limited digestion with micrococcal nuclease, soluble
chromatin was extracted under low salt conditions and size-
fractionated on a sucrose gradient. The sedimentation profile
of nucleosomes and BJ1
antigen were compared. The
gradient fractions were tested for BJI by a solid phase
nucleosomes was monitored through their DNA absorbance.
Figure 2F shows that the nucleosomes were separated from
each other up to the 4-mers, followed by a faster sedimenting
peak containing longer nucleosomal chains. The profile of
BJl antigen closely paralleled that ofthe nucleosomes (Figure
2F) and was similar to the sedimentation profile of histone
H2A (Figure 2G). The fraction with the mono-nucleosomes,
however, contained less BJl. In contrast to BJ1, the RNA-
associated protein P1
remained in the top fractions of the
gradient which contained the RNA fragments (Figure 2G).
The extraction of BJ
NaCl) was dependent on the treatment of the nuclei with
micrococcal nuclease (Figure 2E). These results strongly
at low salt conditions (<300 mM
suggest that the BJ1 protein is associated with nucleosomal
BJ1 is released from nucleosomes by increased ionic
For the experiment shown
chromatin of larger nucleosomal chain length was separated
on sucrose gradients containing different concentrations of
NaCl. At 125 mM NaCl, BJ1 was found in the fractions
containing the bulk of the chromatin. As expected from the
in Figure 2H,
2281 TATATATTCGTTCGCAATACGAAGCCGTAGTAAATTTGTATTTGAACGCAAAAggaaa ctaaatgagatttatttacaaactaattaacttgattatatcatacacaaagatat acat cc
Fig. 3. Genomic and cDNA sequences of BJ1. The nucleotide sequence of a 2.7 kb genomic EcoRI fragment is shown. Sequences also present in
the GP18-1 cDNA (2172 bp) are written with capital letters. Above the genomic sequence, several polymorphic nucleotide exchanges in the GP18-1
cDNA sequence are shown in italics. All of them are in the third positions of the codons. The conceptual translation of the GP18-1 open reading
frame is shown below the nucleotide sequence. The arrows indicate the 5' ends of cDNAs mentioned in the results section. The underlined sequence
(382 bp)showsstrong similarityto asequenceof similarlengthin agenomicXphage (GP18.2, map position lOB on the X-chromosome; data not
shown), which was also isolated underhigh stringencyconditions with GP18 cDNA as aprobe. Thehomology did not extend outside this sequence
and there is no evidence for itbeing expressed.
Expression, localization and chromatin-binding of BJ1 protein
results with intact nuclei, BJ 1 was released from the
sedimentation coefficient (SW20) of
fractions. 250 mM NaCl also released BJ1 from the
nucleosomes. This concentration is slightly lower than the
one required for its release from intact nuclei (Figure 2E,H).
Interestingly, after release at 250 mM NaCl, BJ1 sediments
with a higher velocity than at 450 mM NaCl (SW20
From this result I conclude that BJ1 is bound to chromatin
as a protein-protein complex that is released at increased
ionic strength and falls apart at an even higher salt concentra-
tion. Presently I do not have any information about the
composition of this complex.
at 450 mM NaCl and sedimented with a
-7S in the top
Isolation and sequencing of the BJ1 gene
cDNAs carrying parts of the BJ1 open reading frame were
isolated from a Xgtl 1 expression library made from 0-2 h
embryonic mRNA (U.Rosenberg, unpublished). A mixture
of the antibodies Bj43, Bj46, Bj59 and Bj7O, recognizing
different epitopes, was used for this screen. Two of the
isolated X phages, GP18 and GP4, encoded fusion proteins
recognized by all four monoclonal antibodies. Since BJl was
the only common polypeptide recognized by all four
antibodies, these clones were expected to contain BJ1 cDNA
sequences. One X phage, GP5, gave a signal with the
antibody Bj46 only. Subsequent analysis showed that all three
inserts corresponded to the 3' end of the BJ1 transcript. The
GP18 insert was used to isolate longer cDNAs and genomic
clones. The 2.2 kb insert of the longest cDNA, GP18-1,
closely matched the length (2.2 kb) of the BJ1 transcript,
as determined by Northern analysis (Figure 5A). The entire
GP18-1 cDNA sequence was contained on a 2.7 kb EcoRI
fragment of genomic DNA (derived from the phage clone
GP18.8). The 2.7 kb ofgenomic DNA and the cDNAs GP5,
GP18 and GP18-1 were sequenced (Figure 3). Comparison
of the GP18-1 cDNA sequence with the genomic sequence
revealed the presence oftwo small introns (87 bp and 59 bp).
The sequence ofthe GP18-1 cDNA starts close to the EcoRI
site of the genomic sequence. A canonical polyadenylation
signal, AATAAA, was not found, but the poly(A) of the
cDNA is preceded by AT-rich sequences. A long open
reading frame extends from nucleotide 213 to nt 1853 of
the cDNA. The nucleotide sequence TAAA preceding the
first ATG closely matches the consensus sequence for
Drosophila translation start sites (Cavener, 1987). The ORF
ofthe cDNA encodes a polypeptide of547 amino acids. The
calculated molecular mass of 58.85 kd is lower than the
electrophoretic mobility of BJ1 (68 kd). This difference
might be due to its high content ofcharged amino acids (36
Asp, 39 Glu, 24 Arg, 54 Lys).
The GP18 cDNA starts with codon 486, and GP5 with
codon 499. Therefore the antibodies Bj43, Bj59, and Bj7O
appear to bind to epitopes in the region between amino acids
486 and 499, and Bj46 (Bj10 class) binds somewhere
between amino acids 499 and 547 (Figure 3, Figure 4A).
The BJl gene maps to 64F on the left arm of the third
Structure of the BJ1 protein and sequence
The primary structure of the BJ1 polypeptide is shown
schematically in Figure 4A. Internal sequence comparison
showed that a large portion of the BJ1 protein consists of
homologous, internally repeated domains. Seven tandem
copies of the first type of repeat (repeat lengths of -50
amino acids) are found in the region between amino acids
455 VDKQEQKEN LPAKASTSS
490 VDKQEQKEN LPAKASTSS
BJ1 repeat 1
ENVD Q KEI
478[EP QD NA
BJ1 repeat 2/3
520 V|EAEQElS A
BJ1 EK stretch
Fig. 4. Primary structure and sequence comparisons of the BJ1 protein sequence. A. Schematic diagram of the predicted BJ1 protein. The seven
repeats with weak internal homology (amino acids 46-417) are strongly homologous to similar repeats of the vertebrate RCCI genes. Three repeats
of a different type (amino acids 417-520) are highly homologousto each other. The C-terminal EK-stretch is glutamine- and lysine-rich. The
brackets indicate binding regions of different monoclonal antibodies. Because of the strong sequence similarities of the BJ1 repeats, the antibodies
may have multiple binding sites on each proteinmolecule. B. Sequence comparisonof the BJ1 repeats. Amino acids present in two or all three
repeats are boxed. C and D. Comparison ofsequencesfrom BJ1 withsequencesfrom nuclear proteins of other species, B4, N1/N2, and
nucleoplasmin are from Xenopus laevis (Smith etal., 1988; Kleinschmidt et al., 1986; Dingwall et al., 1987). Histone H1I3 is from
Strongylocentrotus purpuratus (Lai and Childs, 1988).
46 and 417. Most interestingly, this region displays a high
sequence homology to the gene RCCI
Chromatin Condensation) from humans, hamster and
Xenopus (Ohtsubo et al.,
1987; Uchida et al.,
Nishitani et al., 1990). A sequence comparison between the
homologous regions of BJ1 and RCC1 from these species
is shown in Figure
(Ohtsubo et al., 1991). While the homology ofthese repeats
of BJ1 among each other is rather low, 45% of the amino
acids in this region have been conserved between Drosophila
and humans. In yeast, a gene homologous to BJ1 called
SRMI (Clark and Sprague, 1989) or PRP20 (Vijayraghavan
et al., 1989; Aebi et al., 1990) has been characterized. The
amino acid identity between Drosophila and yeast in the
region ofthe seven repeats ('RCC1 repeats') is 30% (Figure
1, Ohtsubo et al., 1991). Thus, the sequence of the RCC1
repeat region has been highly conserved in evolution. Each
ofthese proteins has an N-terminal domain of40-50 amino
acids. These domains lack sequence similarities but in all
cases have a high content of charged amino acids.
Unlike the yeast and vertebrate homologues, BJ1 has a
C-terminal extension of 130 amino acids. Immediately
following the RCC1 repeats, BJ1 contains three additional
repeats of a second type ('BJ1 repeats', repeat length
amino acids, Figure 4A).
homologous to each other (Figure 4B), but are not related
to the RCC 1 repeats. A nuclear protein from Xenopus, B4,
contains similar, but shorter repeats in its C-terminal part
(Figure 4C; Smith et al., 1988). A second nuclear protein
from Xenopus, the histone-binding protein N1/N2, also
contains a stretch of21 amino acids with a similar sequence.
C-terminal to the BJ1 repeats there is a cluster of glutamic
acids, followed by a lysine-rich sequence ('EK-stretch',
Figure 4A). This part of BJI has some similarities to
sequences of nucleoplasmin from Xenopus, and to H1
histones (Figure 4D; Dingwall et al., 1987; Lai and Childs,
1 of the accompanying publication
Expression and subcellular localization of BJ1
Figure 5 shows a Northern and Western analysis of the
abundance of the BJ1 mRNA and protein throughout the
Drosophila life cycle. A strong signal from the BJ1 2.2 kb
mRNA was seen with 0-2 h embryonic RNA (Figure SA).
Since there is no transcription at this stage, this mRNA must
be provided to the embryo maternally. A signal ofthe same
intensity was obtained with 2-4 h old embryos. In later
stages ofembryonic development the abundance of the BJ 1
mRNA decreased drastically. In late embryos and in all
following stages of development, only a faint signal was
detected. Only adult females contain a level of BJ 1 mRNA
similar to that seen in early embryos. This is probably due
to transcription of BJI in the ovaries (see below).
In contrast to the mRNA, only a small amount of BJ1
protein was detected in 0-2 h old embryos (Figure SB).
The level of BJ1 protein increased up until 8-12 h of
embryonic development and slowly decreased again in later
stages of embryonic and larval development. A stronger
signal was obtained with young pupae and adult females.
The differences in the patterns of BJl mRNA and protein
can be explained by translational control and a higher stability
of the protein as compared with that of the mRNA.
The expression and localization ofBJ1 mRNA and protein
were studied in more detail by in situ hybridizations and
r- r r
Fig. 5. Developmental profiles of BJ1 mRNA and BJl protein
accumulation. A. Northern analysis of BJ1 mRNA accumulation. 5 jg
of total RNA from different stages were loaded in each lane. A
random-primed 1.3 kb SacI-XbaI fragment from the 5' part of the
GP18-1 cDNA detected a 2.2 kb RNA. B. Western analysis of the
BJI protein accumulation. 50 jg of total proteins from different stages
were loaded in each lane and tested with the Bj43 antibody. All
samples were collected at 22°C. Abbreviations: 0-2 to 12-24:
embryonic stages in hours after egg deposition; L1,L2,L3: first,
second and third larval instars; P1, P2: early and late pupal stages; F:
adult females; M: adult males.
antibody stainings of ovaries and embryos. Figure 6A-C
shows the expression of BJl mRNA during
development. In Drosophila, the oocyte develops from a cyst
containing 16 interconnected germ line cells. Fifteen ofthese
cells develop into nurse cells which provide the 16th
cell, the prospective oocyte, with RNA and protein. The
surrounding follicle cells are of somatic origin.
In contrast to most other maternal transcripts, the BJ1
mRNA is strongly enriched in the prospective oocyte of
young follicles, and only low levels are detected in the nurse
cells. Figure 6A shows a germarium with two cysts. The
BJ1 transcript appears to be localized to two of the 16 cells
ofeach cyst. From stage 1 to 9 of follicle development (King
et al., 1956), BJ1 mRNA is mainly found in the future
oocyte (Figure 6B,C). At stage 10, when the oocyte grows
by taking up yolk, strong expression starts in the nurse cells
Expression, localization and chromatin-binding of BJ11 protein
Fig. 6. BJl expression during oogenesis. A-C. Detection of the BJl mRNA in follicles of different stages by whole mount in situ hybridization
with a digoxigenin-labeled cDNA probe. A. Germarium representing the earliest stages of oogenesis with two cysts (Cl, C2). BJl mRNA is mainly
detected in one or two out of the 16 cells of each cyst. B. Follicles of stages 2 and 3. BJ1 mRNA accumulates most strongly in the prospective
oocytes. C. Same ovariole as in B, with more advanced stages of follicle development. In stages 4 and 6, high levels of BJI mRNA are found in the
oocytes, and only low levels in the nurse cells. In stage 10, the nurse cells also accumulate high levels of BJI transcript. D. Detection of the BJI
protein by immunofluorescence on frozen sections of follicles (stages 5, 6 and 8). BJI protein accumulates strongly in the oocyte nuclei. The nurse
cell nuclei are weakly stained. E. DNA staining with Hoechst dye of the section shown in D. OCN: oocyte nucleus.
(Figure 6C). At later stages this mRNA is transported into
the oocyte and in early embryos it is equally distributed. No
transcript was detected in follicle cells (Figure 6A-C).
The BJl protein is also strongly enriched in the oocytes
of follicles from early stages onwards. BJI is exclusively
localized in the nuclei (Figure 6D). In germinal vesicles,
BJl is detected in the nucleoplasm and even higher levels
are seen on the condensed chromatin. Nurse cell nuclei
contain less BJl, and in follicle cell nuclei the protein was
not detected (Figure 5D). The presence of BJl in very early
oocytes suggests that, in addition to its role as a maternal
store, the protein has a function in the oocyte nucleus itself.
In embryos undergoing the first cleavages, BJ1 protein
is concentrated in the nuclei and is probably also present
at a lower level in the cytoplasm (Figure 7A). From
blastoderm onwards BJ1 protein is exclusively nuclear during
interphase (Figure 7B). In elongated germ band embryos
(stage 10), BJl mRNA disappears first from the ectoderm
and then also from the mesoderm. High levels ofBJl mRNA
persist only in neuroblasts, and later in ganglion cells derived
from those, until stage 13 (Figure 7C,E). Also for the BJ1
protein, higher levels are detected in nuclei ofthe neuroblasts
and the cells of the CNS (Figure 7D,F). In contrast to the
mRNA, significant levels of BJ1 protein persist in all the
tissues until the end of embryogenesis.
During mitosis, the BJ1 protein leaves the nuclei and
becomes evenly distributed throughout the cells. This was
best seen in fluorescent antibody stainings of embryos
post-gastrulation, where defined patches of cells divide
simultaneously (Foe, 1989). In these mitotic domains the
whole cells were stained (Figure 8A, arrow; Figure 8C),
whereas the neighboring, non-dividing cells showed nuclear
staining. High magnification views of cells progressing
through the cell cycle are shown in Figure 8D-H (Bj59
antibody stainings) and I-M (DNA stainings). In interphase
nuclei, BJ1 has a similar distribution to that of the chromatin,
but is not enriched in the centromeric regions that are brightly
stained with Hoechst dye (D, I). From early prophase to
early telophase, granular BJl staining is seen throughout the
cells (E,J; F,K) and BJI
chromosomes. During telophase BJl reaccumulates in the
nuclei (G,L). In Figure 7H, mitotic cells were stained with
the antibody Bj7O, which crossreacts with the nuclear protein
DI on Western blots (see above). In histological stainings,
a composite signal derived from both BJl and DI was also
seen with this antibody (Figure 8H). In contrast to BJ1, Dl
remains bound to the chromosomes during mitosis, being
most concentrated at the centromeres (H,M, arrows; Figure
2 in Frasch et al., 1986).
BJ1 protein was also detected in nuclei of larval tissues,
including the salivary glands.Thus it waspossibletoanalyze
the distribution of BJ1 on polytene chromosomes. Figure
8N shows that BJl is found in all bands, the more condensed
regions of the chromosomes. Developmental puffs, heat-
is not associated with the
Fig. 7. Expressionand localization of BJI in embryos. A. Whole mount staining of an early cleavage stage embryo with Bj59 antibodies. The nuclei
are more strongly stained (arrow heads). B. Embryo in syncytial blastoderm stage, stained as in A. BJI1 protein is only detected in the nuclei.
C. Whole mount in situ hybridization of a stage 10 embryo with a BJI cDNA probe, Ventral view, anterior is to the left. BJl mRNA is mainly
detected in the neuroblasts. D. Stage 10 embryo, antibody staining as in A; dorsal view, anterior is to the left. The nuclei of neuroblasts are
preferentially labeled (arrow heads). E. In situ hybridization of a stage 10 embryo as in C; cross section, showing BJI mRNA accumulation in the
neuroblasts. F. Stage 13 embryo, antibody staining as in A; anterior is to the left, ventral is down. Nuclei in the CNS are more stronglylabeled
than in ectodermal cells.
shockpuffsand interbands are negative.This stainingis very
similar to the one obtained with DNA dyes (Figure 8H).
A 69 kd antigenwith an identical distribution described by
Fleischmann et at. (1987) may also correspond to BJL.
InDrosophila embryosand in cultured embryonic cells, BJlI
is a relatively abundant nuclear protein. I estimate that the
amount of BJI protein inK,cell nuclei is 2-5% of that
of histone HI. Thus, the abundance of BJI1 is significantly
higher than that ofproteinswhich are known or presumed
to act as transcription factors (Parker and Topol, 1984;
Heberlein et at., 1985; Wu et at., 1987; Krause et at., 1988;
Perkins et at., 1988). The level of BJ1I in nuclei is more
comparablewith that ofhigh mobility group (HMG) proteins
from other species and with that of the HMG-like protein
DI of Drosophila (Johns, 1983; Alfageme et at., 1980;
Bassuk and Mayfield, 1982). BJI shares several other
propertieswith HMGproteins.Theproteinis extracted from
nuclei at 300 nmM NaCl, a salt concentration similar to that
requiredfor the release of HMGs (350 mM NaCI). BJ1I is
bound to nucleosomes and is eluted from nuclei upon
treatment withagentsthat intercalate into DNA, as has been
shown for several HMG proteins (Schroter et at.,' 1985;
Schulman et at., 1987). HMGs typicallycontain --40% of
acidic and basic amino acids. and the BJl protein also has
a highcontent of such residues (31 %). However, BJ1I does
not meet all requirements for an HMG protein because of
its rather high electrophoretic mobility (68 kd) and its
insolubilityinperchloricacid. In contrast to DI, it does not
contain any sequence motifs that are typical of HMGs
(Ashleyet at., 1989; Jantzen et at., 1990; Schulman et at.,
1991). Althoughone class of monoclonal antibodies reacted
with bothBJ1I and Dl, I did not detect anyobvious sequence
the two proteins. Perhaps
antibodies recognizeanon-contiguous epitopewith a similar
structure on both proteins.
A special feature of the BJI primary sequence is its
modular structure. The central 90% of the sequenceis built
Expression, localization and chromatin-binding of BJ1 protein
Fig. 8. Subcellular distribution of BJl protein during the cell cycle and its localization on polytene chromosomes. Indirect immunofluorescence with
Bj59 antibodies (A-G), Bj7O antibodies (H), and Bj86 antibodies (Bj43 class; N). A Antibody staining and B. DNA staining of ectodermal cells in
the head region of a stage 9 embryo. In the domains of mitotically active cells (cycle 14; arrows), BJ1 protein is distributed throughout the cells. C.
Stage 10 embryo. The whole cells are stained in segmentally repeated, mitotic domains (cycle 15). D-H, Antibody stainings and I-M, DNA-
stainings of embryonic cells during cycle 14. D, I. Interphase. BJl is strictly nuclear. E, J. Prophase. BJE is found in granules throughout the cells.
F, K. Anaphase. BJ1 is cytoplasmic and not associated with the chromosomes. G, L. Telophase. BJI redistributes to the daughter nuclei. H, M.
Prophase and telophase. Bj7O antibody detects BJI protein in the cytoplasm and Dl protein at the chromocenters (arrows). N. BJI antibody staining
of the bands on polytene chromosomes from third instar larvae. 0. DNA staining of the chromosomes from N.
from repeated units of
According to their sequence, they fall into two classes: seven
N-terminal repeats with low internal homology and a
different series ofthree C-terminal repeats, which are highly
homologous to one another. It is likely that the protein has
evolved from two ancestral sequences by duplication events,
with the C-terminal repeats ('BJ1 repeats') being of more
recent origin. The proteins from vertebrates and yeast that
are homologous to the seven N-terminal repeats ofBJ 1 lack
the 'BJ1 repeats' (Ohtsubo et al., 1987, 1991; Clark and
Sprague, 1989). However, monoclonal antibodies binding
to the BJ1 C-terminus react with other proteins from these
species, and motifs with similar sequences have been found
in nuclear proteins from Xenopus, particularly in the histone
HI-like protein B4 and the histone-binding proteins N1/N2
and nucleoplasmin. Although these sequences do not
-30-60 amino acids each.
correspond to any of the sequences known to interact with
histones, the glutamine-rich stretch near the C-terminus of
(Kleinschmidt and Seiter, 1988; Erard et al., 1988). The
function ofthe repeated structure is not yet clear. The yeast
nuclear protein nuc2+ also contains 10 internal repeats and
it has been proposed that these repeats form 'snap helices',
which associate with each other intra- and intermolecularly
(Hirano et al., 1990; Sikorsky et al., 1990).The DNA-
binding activity of nuc2+ resides in a domain outside the
repeats. Obviously, a protein with such a dual function would
be ideally suited to induce or stabilize higher order structures
of chromatin. Presently, it is not known whether the repeated
domains in BJ1 have a secondary structure similar to that
proposed from nuc2+ and related proteins. However, the
sedimentation behavior of BJ1 released from chromatin at
is a good candidate for a histone-binding region
250 mM NaCl argues for the occurrence ofprotein-protein
interactions. It is also not known if BJl binds to DNA
directly or through other proteins. The specific release of
BJ1 upon alteration of the DNA and chromatin structure
induced by intercalation argues for a direct binding. With
7.5 mM ethidium bromide and 50 mM NaCl, no othermajor
polypeptide is released that could mediate BJ1 binding to
DNA. If BJ1 is a DNA-binding protein, it must have a low
sequence specificity, since its distribution on polytene
chromosomes does not reveal any preference for particular
chromosomes is almost identical to that seen for histone HI
The sequence homology of BJ1 to the gene products of
RCCI from vertebrates and SRMJ/PRP20 from yeast
covering 70% of the BJ1 polypeptide is highly significant
and argues for a conserved function of these genes. The
RCCJ protein, which has been studied most extensively,
shares many other features with BJ1. Like BJ1, it is a nuclear
protein that diffuses into the cytoplasm during mitosis
(Ohtsubo et al., 1989). The RCC1 protein is also extracted
from nuclei at 300 mM NaCl (Ohtsubo et al., 1989) and
a protein likely to be identical to RCC1 is released from
nuclei by intercalating agents (Bischoff et al., 1990). There
is evidence for a DNA-binding affinity of RCC1 (Ohtsubo
et al., 1989). The patterns obtained with antibodies against
the Xenopus RCCI gene product (Nishitani et al., 1990) on
Western blots with Drosophila extracts and in histological
embryo stainings were identical to those obtained with the
BJ1 monoclonal antibodies (unpublished results). Most
importantly, the complementation of an RCCI mutation in
a mammalian cell line with the Drosophila BJ1 gene, and
also with the yeast gene, strongly argues for the idea that
these genes have equivalent functions in different species
(Ohtsubo et al., 1991). The functions of these genes are not
yet fully understood in any ofthe systems, vertebrates, yeast
or Drosophila. Since the genes have been studied in each
of these systems from a different perspective, we can now
combine this information to get a more complete image of
the cellular functions of their products.
Most information on the vertebrate RCCI gene was
obtained from the analysis of a hamster cell line, tsBN2,
carrying a temperature-sensitive mutation in this gene
(reviewed by Nishimoto, 1988). The phenotype of this cell
line at restrictive temperature suggested a function ofRCC 1
in the control of the cell cycle. In G2 phase of wild type
cells, a cascade of events involving newly synthesized
proteins and post-translational modifications results in the
activation of the M-phase- (or maturation-) promoting factor
(MPF) that triggers entry into mitosis. The active component
of the MPF complex is a protein kinase (p34Cdc2), which is
also present, but inactive, during S phase (reviewed by
Nurse, 1990). An unknown mechanism prevents activation
of MPF before completion ofDNA replication (Dasso and
Newport, 1990). This mechanisms appears to be disrupted
in the tsBN2 cells at restrictive temperature. The loss of
RCC1 function in these cells after a temperature shift during
S phase results in a premature activation of MPF and thus
entry into mitosis (Nishitani et al., 1991). This activation
requires synthesis of new proteins, and probably also
transcription ofmRNA(s) (Nishimoto et al., 1981; Nishitani
et al., 1991). Therefore in wild type, until completion of
replication during S phase, RCC1 seems to inhibit the
pattern on polytene
expression of one, or several, gene products which are
involved in MPF activation.
Thedevelopmental expression ofBJl would be consistent
with a function in mitoticregulation. Theembryoisprovided
with alarge amount of maternal BJl mRNA and also with
some maternal BJl protein. It is conceivable that these
maternal products arenecessary to allow the rapid nuclear
divisions which occur inearly embryos priortozygotic gene
expression. After cell divisions have ceased in theepidermis,
BJ1 mRNA isonlyfound in the nervoussystem, where cells
continue to divide. Thispattern ofexpression is similar to
that found for othergenes involved in mitoticcontrol, such
as thecyclins (Lehner andO'Farrell, 1990. Whitfield etal.,
1990) andstring (Edgar and O'Farrell, 1989). However,
unlike thecyclins and thestring product, the BJ1 protein
appearsto beverystable and asignificant amount of it seems
to persist until larval or pupal stages. BJ1 protein is also
found,at a lowerlevel, inpost-mitotic cells that are arrested
inGI, e.g. in the late embryonic ectoderm, or inpolytene
cells of thesalivary gland. One can assume that, in these
cells, BJ1 has a function different from regulation of cell
cycleevents. Similarly, tsBN2 cells mutant for RCCIdisplay
aphenotype which is not due to cell cycle defects, when
they are shifted to restricted temperatures during G1 phase
(Nishitani etal., 1991). Rather, these shifts lead to drastic
alterations ingene expression (Nishimoto et al., 1981). The
prp2Omutation in theRCCJ-like gene from yeast also results
in altered mRNA metabolism, and in perturbations of
chromatin structure. The steady state levels of several
transcripts decrease, while new transcripts seem to be
expressedfromcryptic promoters (Aebi et al., 1990). SRMI,
a(probably hypomorphic) allele ofPRP20, wasoriginally
isolated as asuppressor that restored mating ofpheromone-
receptorless mutants (Clark and Sprague, 1989). This
phenotype may be due to alterations in gene expression as
well. Such alterations could include over-expression of
components of the signal transduction pathway, or under-
expression of inhibitors, resulting in constitutive signal
Thus, effects on gene expression appear to be a common
theme in the phenotypes ofRCCIISRMIIPRP20 mutations.
As I have pointed out, the properties of BJ 1 do notsuggest
that it acts as a transcription factor which binds tospecific
regulatory sequences oftarget genes. Rather, most datapoint
to a role of BJ1 in the establishment or maintenance of a
proper higher order structure of the chromatin. I suggest
thatBJl, and the analogous genes in vertebrates andyeast,
influence or stabilize chromatin structure in a way that allows
correctregulation andexpression ofgenes during interphase
and Sphase, including those genes controlling activation of
MPF. The coupling of MPF activation to the completion
ofDNA replication could act through a separate mechanism,
which only requires the function of BJl/RCCI/PRP20 as
a prerequisite. However, it is also possible that the higher
order structure of chromatin and replication are more
intimately connected to each other. The silencing ofmating
type loci in yeast, where the establishment of a particular
chromatin structure depends on a replication event, could
be a precedent for such a mechanism (Miller andNasmyth,
1984; Kayne et al., 1988; Johnson et al.,
identification and analysis of proteins associated with
BJl/RCC1/PRP20 (Figure 2; Ohtsubo et al., 1989; Bischoff
et al., 1990) together with genetic and molecular studies in
Expression, localization and chromatin-binding of BJ1 protein
Drosophila and yeast should help to further clarify the rela-
tionship between chromatin structure, gene expression and
cell cycle control.
Materials and methods
Isolation of nuclei
10 g portions of 1- 15 h old embryos were dechorionated and homogenized
in 100 ml ofembryo buffer (60 mM KCI, 15 mM NaCl, 0.34 M sucrose,
1 mM DTE, 1 mM EDTA,
X-100, 15 mM Tris pH 6.4). The suspension was filtered through a nylon
mesh and layered onto step gradients of 3 ml 2.4 M sucrose and 7 ml 1.3 M
sucrose in embryo buffer. After centrifugation for 15 min at 6000 r.p.m.
in a Sorvall HB4 rotor, the pellets on top of the 2.4 M sucrose cushion
were homogenized and diluted in 150 ml of embryo buffer. The nuclei were
recentriftiged for 90 min at 25000 r.p.m. in a SW27 rotor on step gradients
of 1 ml 2.4 M sucrose and 10 ml 2.2 M sucrose in embryo buffer. The
nuclei were collected from the top phase of the 2.4 M sucrose cushion.
Nuclei from Kc cells were isolated as described previously (Risau et al.,
1 mM EGTA,
1 mM PMSF, 0.1 % Triton
Generation of monoclonal antibodies against nuclear proteins
The preparation of fractions of nuclear proteins used for immunization has
been described (Frasch and Saumweber, 1989). The Bj antibodies were
obtained with a fraction of proteins that were released from chromatin at
450 mM NaCl and that were contained in the flow through of a QAE
Sephadex column at 5 M urea, 150 mM NaCl, 0.5 mM DTE, 10 mM Tris
pH 8.3 (Augenlicht and Baserga, 1973). BjlO, Bj46, Bj59 and Bj70 are
IgG1, Bj43 is IgG3, and Bj86 isIgG2h.
Preparation and fractionation of soluble chromatin
The digestion of nuclei fromKccells with micrococcal nuclease and RNase
A, the fractionation of the extracted chromatin on sucrose gradients and
the analysis of the gradient fractions were done as described previously
(Frasch and Saumweber, 1989).
Protein extraction by 'elutive intercalation'
This procedure was performed according to Schroter et al.
4.5 x 108 nuclei from Kc cells were resuspended in 3 ml of X2 buffer
(1 mM Tris pH 7.0, 0.2 mM EDTA), 7.5 mM ethidium bromide, and NaCl
at concentrations of 0, 50 or 100 mM. Extraction was for 45 min at 0°C
with shaking. After centrifugation (10 min, 4000 g), the extracted proteins
were precipitated with 20% TCA, washed with ethanol, and boiled in SDS
gel loading buffer. The extract recovered from 5 x 107 nuclei (Coomassie
staining) or 1 x 107 nuclei (Western blots) was loaded in each lane.
Gel electrophoresis and Western blots
Electrophoresis on SDS-polyacrylamide gels and Western blots were done
as described (Frasch and Saumweber, 1989). The following proteins were
used as molecular weight standards: trypsin inhibitor (soybean, 20.1 kd),
carbonic anhydrase (29 kd), P11 antigen (Risau et al., 1983; 36 kd), alcohol
dehydrogenase (subunit, 39.8 kd), BSA (68 kd), S5 antigen (Risau et al.,
1983; 70 kd), ,B-galactosidase (E.coli, 116 kd), RNA polymerase (E.coli,
155 kd and 165 kd) and myosin (205 kd).
Isolation of cDNA clones
cDNA clones were isolated from a Xgtl 1 expression library as described
by Young and Davis (1983). The library was made from 0-2 h embryonic
RNA with a size-selection
(U.Rosenberg, unpublished). 240 000 phages were plated with a density
of 14 000 phages per plate (11 x 11 cm). After IPTG induction, a mixture
of the monoclonal antibodies Bj43, Bj46, BjS9 and Bj7O, binding to different
epitopes of the BJl protein, was used for filter incubation, followed by a
secondary, alkaline phosphatase conjugated antibody. After the staining
reaction, signals were obtained from 23 phages, 12 of which were further
purified. Three of these could be detected by all four antibodies, two were
only detected by Bj46, four were only detected by BjS9, and nine of them
showed cross-hybridization of their inserts.
cDNAs with longer inserts were isolated from a pNB40-plasmid library
prepared from 4-8 h embryonic RNA (Brown and Kafatos, 1988).
in the 500 bp range
cDNAs and genomic fragments were subcloned into pBluescript KS + and
SK+ (Stratagene). Deletions were created with appropriate restriction
enzymes and the sequencing reactions were carried out with M13 primers
or with primers complementary to insert sequences. Sequencing was
performed by the dideoxynucleotide sequencing method (Sanger et al., 1977)
using single stranded templates and a modified form of T7 polymerase
Northern blots were done as described previously (Dohrmann et al., 1990),
but using total RNA rather than poly(A)+ RNA.
In situ hybridization
In situ hybridizations to whole embryos and ovaries were carried out ac-
cording to the procedure ofTautz and Pfeifle (1989). Digoxigenin conjugated
dUTP and anti-digoxygenin antibodies were from Boehringer Mannheim.
Hybridization probes were prepared according to the protocol of Feinberg
and Vogelstein (1984), using a molar ratio ofdTTP:DigdUTP of 2:1. For
sections (10/Am),the embryos were embedded in Araldite after staining
(Leptin and Grunewald, 1990).
For antibody stainings ofwhole embryos (Dequin et al., 1984) and ovaries,
fixation was carried out for 20 min in a buffer containing 45 mM KCI,
15 mM NaCl,
13 mM MgCl2,10 mM K-phosphate pH 6.8, 4%
formaldehyde and 15% (v/v) of a saturated, aqueous solution ofpicric acid.
When picric acid was not included, the cytoplasmic BJI protein was not
retained. Frozen sections were made after the staining (Dequin et al., 1984).
Hoechst 33258 was used for DNA staining at 1 Stg/ml (in PBS). For HRP
stainings, the VECTASTAIN detection kit (Vector Laboratories) was used
with diaminobenzidine as a substrate.
Fixation, squashing and staining of polytene chromosomes was done as
described by Saumweber et al. (1980). Photographs were taken with Kodak
2415 technical pan film and developed with developer HCl 10 (Nomarski
optics) or D19 (fluorescence).
I gratefully acknowledge the contributions of Raymond Dequin at the
beginning of the project and would like to thank Harry Saumweber for his
advice. I appreciate the expert technical assistance ofMonika Wild and Meike
Muller. I would like to thank Friedrich Bonhoeffer and Janni Niisslein-
Volhard for their support and encouragement. I also thank Takeharu
Nishimoto for communicating unpublished results, and Christian Lehner
and Daniel St Johnston for critically reading the manuscript.
Aebi,M., Clark,M.W., Vijayraghavan,U. and Abelson,J. (1990) Mol. Gen.
Genet., 224, 72-80.
Alfageme,C.D.R., Rudkin,G.T. and Cohen,L.H. (1980) Chromosoma, 78,
Ajiro,K., Nishimoto,T. and Takahashi,T. (1983) J. Biol. Chem., 258,
Glover,C.V.C. (1989) J. Biol. Chem. 264, 8394-8401.
Augenlicht,L.H. and Baserga,R. (1973) Arch. Biochem. Biophys., 158,
Bassuk,J.A. and Mayfield,J.T. (1982) Biochemistry, 21, 1024-1027.
Bischoff,F.R., Maier,G., Tilz,G. and Ponstingl,H. (1990) Proc. Natl. Acad.
Sci. USA, 87, 8617-8621.
Brown,N.H. and Kafatos,F.C. (1988) J. Mol. Biol., 203, 425-437.
Cavener,D.R. (1987) Nucleic Acids Res., 15, 1353-1361.
Clark,K.L. and Sprague,J.F.,Jr (1989) Mol. Cell. Biol., 9, 2682-2694.
Cleveland,D.W. Fisher,S.G., Kirschner,M.W. and Laemmli,U.K. (1977)
J. Biol. Chem., 252, 1102-1106.
Dasso,M. and Newport,J.W. (1990) Cell, 61, 811-823.
Dequin,R., Saumweber,H. and Sedat,J.W. (1984)Dev. Biol., 104, 37-48.
Dingwall,C., Dilworth,S.M., Black,S.J., Kearsey,S.E., Cox,L.S. and
Laskey,R.A. (1987) EMBO J., 6, 69-74.
Dohrmann,C., Azpiazu,N., and Frasch,M. (1990) Genes Dev., 4,
Dorbic,T. and Wittig,B. (1987a) EMBO J., 6, 2393-2399.
Entwickungsmech. Org., 190, 197-207.
Edgar,B.A. and O'Farrell,P. (1989) Cell, 57, 177-187.
Einck,K. and Bustin,M. (1985) Exp. Cell Res., 156, 295-310.
Eissenberg,J.C., James,T.C., Foster-Hartnett,D.M., Hartnett,T., Ngan,T.,
Elgin,S.R.C. (1990) Proc. Natl. Acad. Sci. USA, 87, 9923-9927.
Amalric,F. (1988) Eur. J. Biochem., 175, 525-530.
Feinberg,A.P.andVogelstein,B. (1984)Anal. Biochem., 137, 266-267.
Felsenfeld,G. and McGhee,J.D. (1986) Cell, 44, 375-377.
Fleischmann,G., Filipski,R.andElgin,S.C.R. (1987) Chromosoma, %,
Foe,V.E. (1989) Development, 107, 1-22.
Frasch,M. (1985) Charakterisieurung chromatinassozieerter Kernproteine
vonDrosophila melanogastermitHilfe monoklonaler Antikorper. Ph.D.
Thesis, Eberhardt-Karls-Universitat, Tubingen.
Frasch,M. and Saumweber,H. (1989) Chromosoma 97, 272-281.
Frasch,M., Glover,D.M. and Saumweber,H. (1986) J. Cell Sci., 82,
Frasch,M., Paddy,M.andSaumweber,H. (1988)J. CellSci., 90, 247-263.
Fuchs,J.-P., Giloh,H., Kuo,C.H., Saumweber,H. and Sedat,J. (1983)J. Cell
Sci., 64, 331-349.
Garzino,V., Moretti,C. and Pradel,J. (1987) Biol. Cell, 61, 5-14.
Heberlein,U., England,B.and Tjian,R. (1985) Cell, 41, 965-977.
Henikoff,S. (1990) Trends Genet., 6, 422-426.
Hirano,T., Kinoshita,N., Morikawa,K. andYanagida,M. (1990) Cell, 60,
James,T.C. andElgin,S.C.R. (1986) Mol. Cell. Biol., 6, 3862-3872.
Jantzen,H.-M., Admon,A. Bell,S.P. andTjian,R. (1990) Nature, 344,
Johns,E.W. (1983) (ed.)The HMG Chromosal Proteins. Academic Press,
Grunstein,M. (1988) Cell, 55, 27-39.
King,R.C,. Rubinson,A.C. andSmith,R.F. (1956) Growth, 20, 121-157.
Kleinschmidt,J.A. and Seiter,A. (1988)EMBO J., 7, 1605-1614.
Kleinschmidit,J,A., Dingwall,C., Maier,G. andFranke,W.W. (1986)EMBO
J., 5, 3547-3552.
Krause,H.M., Klemenz,R. andGehring,W.J. (1988) Genes Dev., 2,
Kuo,C.H., Gioh,H., Blumenthal,A.B. and Sedat,J.W. Exp. CellRes., 142,
Lacroix,J.C., Azzouz,R., Boucher,D., Abbadie,C., Pyne,C.K.
Charlemagne,J. (1985) Chromosoma, 92,69-80.
Lai,Z.-C. andChilds,G. (1988)Mol. Cell. Biol., 8, 1842-1844.
Lehner,C.F. andO'Farrell,P.H. (1990) Cell, 61, 535-547.
Leptin,M. and Grunewald,B. (1990) Development, 110, 73-84.
Levinger,L.andVarshavsky,A. (1982a) Cell, 28, 375-385.
Levinger,L.andVarshavsky,A. (1982b)Proc. Natl. Acad. Sci. USA, 79,
Levinger,L.F. (1985)J. Biol Chem., 260, 14311-14318.
Miller,A.M. andNasmyth,K.A. (1984) Nature, 312, 247-251.
Nishimoto,T. (1988) BioEssays, 9, 121-124.
Nishimoto,T., Eileen,E. and Basilico,C. (1978) Cell, 15, 475-483.
Nishimoto,T. Ishida,R., Ajiro,K., Yamamoto,S. and Takahashi,T. (1981)
J. Cell.Physiol., 109, 299-308.
Nishitani,H., Kobayashi,H., Ohtsubo,M. and Nishimoto,T. (1990) J.
Biochem., 107, 228-235.
Nurse,P. (1990) Nature, 344, 503-508.
Matsubara,K. and Nishimoto,T. (1987) Genes Dev., 1, 585-593.
Ohtsubo,M., Okazaki,H. and Nishimoto,T. (1989) J. Cell Biol., 109,
Ohtsubo,M., Yoshida,T., Seino,H., Nishitani,H., Clark,K.L., Sprague,G.,
Frasch,M. and Nishimoto,T. (1991) EMBOJ., 10, 1265-1273.
Parker,C.S. andTopol,J. (1984) Cell, 36, 357-369.
Paro,R. (1990) Trends Genet., 6, 416-421.
Perkins,K.K., Dailey,G.M.andTjian,R. (1988)EMBOJ., 7, 4265 -4273.
Risau,W., Symmons,P., Saumweber,H., and Frasch,M. (1983) Cell 33,
Sanger,F.S., Nicklen,S. andCoulson,A.R. (1977)Proc. Natl. Acad. Sci.
USA, 74, 5463-5474.
Saumweber,H., Symmons,P., Kabisch,R., Will,H. and Bonhoeffer,F.
(1980) Chromosoma, 80, 253-275.
Schlissel,M.S. and Brown,D.D. (1984) Cell, 37, 903-913.
Schroter,H., Maier,G., Ponstingl,H.andNordheim,A. (1985) EMBO J.,
Schroter,H., Shaw,P.E. andNordheim,A. (1987)Nucleic Acids Res., 24,
Schulman,I.G., Cook,R.G., Richman,R. andAllis,C.D. (1987)J. Cell.Biol.,
Schulman,I.G., Wang,T., Wu,M., Bowen,J., Cook,R.G., Gorovsky,M.A.
and Allis,C.D. (1991) Mol. Cell. Biol., 11, 166- 174.
Sikorsky,R.S., Gogushi,M.S., Goebl,M. and Hieter,P. (1990) Cell, 60,
Smith,R.C., Dworkin-Rastl,E. andDworkin,M.B. (1988) GenesDev., 2.,
Strauss,F. andVarshavsky,A. (1984) Cell, 37, 889-901.
Sun,J.-M., Wiaderkiewicz,R. andRuiz-Carillo,A.-R. (1989) Science, 245,
Tautz,D. and Pfeifle,C. (1989) Chromosoma, 98, 81-85.
Uchida,S., Sekiguchi,T., Nishitani,H., Miyauchi,K., Ohtsubo,M. and
NishimQto,T. (1990) Mol. Cell. Biol., 10, 577-584.
Vijayraghavan,U., Company,M. and Abelson,J. (1989) Genes Dev., 3,
Weintraub,H. and Groudine,M. (1976) Science, 193, 848-865.
Weisbrod,S., Groudine,M. andWeintraub,H. (1980) Cell, 19, 391-400.
Whitfield,W.G.F., Gonzalez,C., Maldonadao-Codina,G. and Glover,D.M.
(1990) EMBO J., 9, 2563-2572.
Widom,J. (1989) Annu. Rev. Biophys. Biophys. Chem., 18, 365-395.
Wolffe,A.P. (1989) EMBO J., 8, 527-537.
Wu,C., Wong,Y.-C. and Elgin,S.C.R. (1979) Cell, 16, 807-814.
Wu,C., Wilson,C., Walker,B., Dawid,I. Paisley,T., Zimarino,V. and
Ueda,K. (1987) Science, 238, 1247-1253.
Young,R.A. and Davis,R.W. (1983) Proc. Natl. Acad. Sci. USA, 80,
Received on January 30, 1991