Cell-Free System for Assembly of Transcriptionally Repressed
Chromatin from Drosophila Embryos
PETER B. BECKERt AND eARL WU ·
Laboratory 0/ Biochemistry, N ational Cancer Institute, Building 37,
Room 4C·09, Bethesda, Maryland 20892
Received 9 December 199I/Accepled 26 Feb rUII)' 1992
We descrtbe a cell·free system, derlved (rom preblastoderm DrosophÜIJ embryos. ror tbe efficient assembly
of eloned DNA into chromatin. Tbe cbromatin assembly system ulilizes cndogcDous eure histones aod assembly
racton aod yields loog arrays of regularly spaced nucleosomes with a repeat Icngtb of ISO bp. Tbc 8nembly
system is also eapable of C1lmplementary-strand DNA syotbesis accompanied by rapid nucleosome formation
wben tbe startiog template is single-stranded circl!lar DNA. Chromatin assembled wilb tbe preblastodenn
embryo exlrae' is natunlly deficienl in histone 91. but exogeoous "1 eIn be incorporated during nucloosome
assembly in vitro. Regular spaeing or nucieosomH witb or witbout bistone Hl i5 sulfieient to maximally repress
transtription from hsp70 and rusbi tarazu gene promoters. Tbe Drorophilo assembly system sbould be
particularty useful for in vitrn studies or cbromatln assembly durlng DNA synlbesis and ror elueidating the
action or transcription factors in the conte'" or native chromatjn.
Transc ription in cukaryotes is generally associated with
loeal changes of the regular c hromatin struclure and composilion (see reference 14 for a review). While "active chromatin" as a lechnical lerm has been in existence for many
years, the exact nalure and determi nants of these activated
regions are largely unknown (28, 43). In general, sites of
Iranscriplion appear 10 be less condensed and thus more
accessible than the tightly packed bulk chromalin . Probing
the accessibility of sequences with nueleases in isolated
nuelei has yielded a wealth of infonna lion aboul Ihe loea!ization of regulatory siles, such as enhancers or promoters,
and their potential or actual activity.
While the corrcla tions between Iranscriplion and accompanying chromatin features are weil established, it has only
recently been appreciated that structural components of
chromatin can be used to regulate gene activity in a highly
specific manner (16, 17, 42, 45). lt has been shown that
nueleosomes that are positioned in vivo with regard to the
underlying DNA sequence can influence transcription and
rcplication by modulating the accessibility of regulatory sites
for the binding of protein factars in vivo. Conversely,
DNA·binding proteins have been described that aetivate or
repress transcription by influencing the position of nueleosomes in the vicinity of important regulatory sequences (10.
19,31,37,46). To elucidate these phenomena at the molee·
ular level, an in vitro reconstruction of the mutual interde·
pendenee of transc riptional regulators and struetural com·
ponents of chromatin unde r physiological conditions is
necessary.
We are interested in general aspeets of transc riptional
regulation by RN A polyme rase 11 in the eontext of chroma·
tin, using the fru it Hy Drosophila me{anogasrer as a model
system. Apriori, Drosophila embryos should be an exccllenl
souree of faelors required for nueleosome asse mbly: during
the initial stages o( embryonic development, the fly genome
is replieated and paekaged into chromatin onee every 9 min
• Corrcsponding author.
t Presen! address: Ge ne Expression Program, European Molec·
ular Biology LaboralOry, Meyerhofstrasse 1, D·6900 Heide lbcrg.
Ge rmany.
by using a matemal pool of histones and assembly faetors
(11). A eell·free system for nueleosome assembly derived
from early Drosophila embryos has been reported previ·
ously (25). However, this proecdure has nol been adopted by
workers in the field, and our aUempts to creatc an cfficient
assembly extracl according to the published procedure were
not s uccessful. Here we report the development of a stable
and very effieient eell-free system for nucleosome assembly
using Drosophila embryos. The system is also able to carry
out a DNA synthesis reaetion that mimics lagging·strand
replieation. We find that assembly of nucleosomes on hsp70
and rushi larazu DNA templates is suffieient to maximally
rcpress transcription in vitra, wilh or without the eoneomi·
tant assembly of histone HL
MATERIALS AND METHODS
Exlntct prepanttion . Drosophila embryos 0 to 100 min
afte r egg laying were rinsed in water and allowed to senle
into embryo wash buffe r (0.7% NaCl, 0.05% Triton X-IOD)
on iee to arrest funher developmenl. Afler four to five
sueeessive eolleetions, the pooled harvest was deehorionated. Tbe wash buffer was deeanted and replaeed with wash
buffer at room temperature, and the volume was adjusted 10
200 ml. Afler Ihe addition of 200 ml of Chlorox bleach, the
embryos were stirred vigorously for 90 s, poured back into
the colleetion sieve, and rinsed extensively with tap water.
They were then allowed to settle in 1 liter of wash buffer for
aboul 2 min, after which the supernatant (contai ning the
chorions) was aspirated off. Four more seitlings were performcd : one in wash buffer, IwO in 0.7% NaCl, and one in
exlracl buffer (10 mM N-2-hydroxyelhylpiperazine·N'-2ethanesulfonie acid (HEPES, pH 7.6), 10 mM KCI, 1.5 mM
MgCI2 , 0.5 mM ethylene glycol.bis(ß·aminoethyl ethcr)N,N,N',N' -tetraaeetie acid (EGTA), 10% glyeerol, 10 mM
ß·glycerophosphate; 1 mM dithiothreitol and 0.2 mM phen.
ylmethylsulfony l fluoride, added freshly) at 4°C. The em·
bryos in extraet buffer were settled in a 6O·ml glass homog·
enizer on iee for abou t 15 min, and the volurne of the packed
embryos was estimated (- 10 to 15 ml). The supematan t was
aspi rated, and the embryos were homogcnized by six com·
plete strokes with a Teflon pestle connected to a drill press.
All further manipulations were carried out at 4~C. The
homogenate was supplemented with additional 5 mM MgCl 2
from alM MgCi 2 stock solution and quickly mixed (final
MgCI 2 concenlration, 6.5 mM). When nucleosome assembly
was assayed at low MgCI 2 concentrations, no MgCl 2 was
added to the embryo homogenale (final MgCi 2 concenlralion, l.5 mM). Nuclei were pelleled by centrifugalion for 5
min at 5,000 rpm in a JA14 rotor (Beckman). The supemalant was clarified by centrifugation for 2 h at 40,000 rpm
(150,000 x g) in an SW 50.1 rotor (Beckman) and collected
with a syringe by puncturing the tubes just above the solid
pellet, avoiding Ihe floating layer of lipid. Aliquots (300 to
500 ...,1) were frozen in liquid nitrogen. Protein concenlrations
were determined with the Bradford assay, using bovine
serum albumin as the Slandardj they were usually between
20 and 30 mg/ml.
AI the last seitling step in Ihe extract buffer, not all
embryos seille bccause of changes in osmolarity. In addilion, variable degrees of settling will resull in corresponding
variable final exlract protein concentrations. To standardize
Ihe yields of the extract preparation, an alternative procedure was employed. After the seitling in 0.7% NaCi, the
embryo suspension was poured onto a filter paper on a
Büchner funnel, filtcred under vacuum, and washed twice
with 100 ml of extract buffer. The damp embryos were
resuspended in 1 ml of cold extraci buffer per g of embryos.
Homogenization and cenlrifugation were performed as described above. When assayed by micrococcal nuclcase digestion, this procedure yielded extracts which assembled
chromatin with an efficiency identical 10 that of the one
described above. It did, however, contain protein aggregates
(which did not disperse in 0.05% Nonidet P40) that cosedimented with the minichromosomes during sucrose gradient
purification and masked the analysis of histone assembly.
Extracts of this kind were therefore not used routinely.
Asscmbly reaction. In a standard assembly reaction, l.5 ILg
of plasmid DNA was incubaled at26~C wilh 3 mg of embryo
extraci protein and the following reagents (which were
added from a 10x premix): 30 mM crealine phosphale, 3 mM
MgCi 2, 3 mM ATP (pH 8), O.l...,g of creatine phosphokinase
(type 1; Sigma) per ml, and 1 mM dithiothreilol. The volume
was adjusted to 200 ILI wilh exlract buffer. The final conduc·
tivity in the asscmbly reactions is equivalcnt to 65 mM KCi.
The protein concentration for optimal assembly was determined for every cxlract. using the micrococcal nuclease
assay. Vnless stated olherwise, the MgCI 2 concenlration in
the assembly was adjusted 10 6.5 10 7 mM, taking into
account the MgCl 2 that had been added during the extract
preparation. Reactions with exogenous histone Hl were
mixed with the histone prior to the addition of DNA.
Supercoiling analysis. At various times during the reaction,
40 ...,1 of assembly mixture was removed and 10 ILI of 5 x stop
mix (2.5% Sarkosyl, 100 mM EDTA) and 1...,1 ofDNase·free
RNase (Boehringer Mannheim) were addedj the mixture was
then incubated for 15 min at 37"C. Then 6.5 Il-l each of 2%
sodium dodecyl sulfate (SDS) and lo.mg/ml proteinasc K
were added and again incubated for 30 min at 37"C. The
DNA was precipitated with 2 volumes of ethanol after the
addition of 1 "I of glycogen (10 "g) and 45 "I of 7.5 M
ammonium acelate. After centrifugation, Ihe DNA was
washed with 80% ethanol, dried under vacuum, and finally
dissolved in 8 Il-I of TE (10 mM Tris HCI (pH 7.51, 1 mM
EDTA). DNA was eleclrophoresed on a 1.2% agarosc gel
with Tris-glycine buffer (Iacking ethidium bromide) for 15 h
at20V.
Microc:occal nucleasc analysis. To a standard assembly
reaction, 6 "I of 0.1 M CaCI 2 was added and quickly mixed.
A 4o.Il-I portion of the mixture was removed and treated the
same as for supercoiling analysis. To the remaining reaction,
5 ...,1 of micrococcal nuclease (Boehringer Mannheimj 50
V/...,I in extract buffer) was addcd. After 0.5, 2, and 8 min al
room temperature, 40 ...,1 was again removed and the digestion was terminated. An RNase treatment for up to 1 h was
followed by ovemight proteinase K digestion. Gel electrophoresis of the micrococcal nuclease digestion products was
as described in detail by Shimamura el al. (35). Micrococcal
nuclease analysis of Drosophila embryo chromatin was
performed as described by Wu el al. (50).
DNA templates. In the experiments presented here, the
plasmid phsp706185 (3,443 bp), which containshsp70 (Iocus
87A) gene sequences between theXhol site al -185 and the
Acci site at +300 (+ 1 is the transcriptional start), was used.
The Xhol-Acci fragment was filled in with T4 polymerase
and cloned inlo the Hincll site of pBluescript SK M13+
(Stratagene). Single-stranded DNA was purified from phage
particles after superinfection of a bacterial culture containing
the plasmid wilh helper phage VCSM13 (Stratagene) according 10 standard procedures. A variant of phsp706185 (hsp70
minigene) was constructed by deleting an Alul fragment
(+41 10 +71) as described previously (6). The plasmid
carrying the fushi tarazu promoter (-950 to +151) was as
describcd previously (6).
Replication aS5lllY. Conditions for the replicalion assay
were similar 10 those for standard nucleosome assembly
reactions, excepl that 750 ng of single-stranded DNA instead
of double-stranded plasmid DNA was used. For "uniform
labeling," 1 Il-l of (a_ 32 PJdcrP at 2,000 to 3,000 CiJmmol
(NEN) was added. For site-specific labeling, the singlestranded DNA was first annealed with a fivefold molar
excess of a np-end-Iabeled primer (hsp70 nucleotides -185
to - 153) in 20 "I of 100 mM NaCI-S mM MgCi 2• The mixture
was incubated for 5 min at 75~C, 10 min at 37"C, and 5 min
at 26°C; the extract and asscmbly components (without
radiolabeled dcrP) were then added.
Histone purifications. Core histones as standards for gel
eleclrophoresis were purified flOm the chromatin of 0- to
2O-h-old embryos according 10 Ihe melhod of Simon and
Felsenfeld (36). They were identified by their migration
behavior on SOS gels (44). Hl was purified trom nudei of 010 12-h embryos by the procedure of Croslon et al. (8). The
peak fraction of the phenyl-Sepharose column was concentrated four- to fivefold by vacuum dialysis in a mini-collodion
bag (Schleicher & Schuell).
Plasmid chromatin purißcation. A total of 1.5 ...,g of singlestranded DNA or 3 j.Lg ofdouble-slranded plasmid DNA was
assembled for 6 h under standard conditions. The synthesis
of the complementa~ strand was followed by thc introduclion of 3 ...,1 of la- 2PJdcrP 10 the sampies containing
single-stranded DNA. Purification of chromatin was
achieved by centrifugation Ihrough a 15 to 30% sucrose
gradient followed by pelleting Ihrough a 30% sucrose cushion exactly as described previously (35). The chromatincontaining fractions were identified by the incorporation of
radioactivity. Silver staining of the histone gels was done as
described by Wray et al. (49).
Transcription or asscmbled templaces. Drosophila embryo
transcription extracts were prepared and coupled assemblytranscription reactions were processed as described previously (6), wilh the following modifications. The chromatin
assembty reaction was scaled down to 50 j.LI in proportion to
the input amount ofDNA template (ISO ng each of thehsplO
KCl:
nc, rel .....
sc +
10 30 60 ISO 360 110 30 60 ISO 360
I
10 30 60 ISO 360
FIG. 1. Chromatin as.sc:mbly analyzed by DNA supcrcoiling analysis. Supcrcoiled plasmid (1.5 ~g) was incubated under standard
cooditions with 2 mg of eXlraet protein in the presence of 6.5 mM MgCl l and the indicated concentrations of KCI. At various times, aliquots
were removed from th e reaction. DNA was deproteinized, elcctrophoresed on a 1.2% agarose gel in Tris·glycioe bulfer, aod staioed with
ethidium bromide. Comparison with DNA prepared from a reaclion poinl al time zero showed no sigoificant loss of input DNA during Ih e
reaction (dala nOI shown). The positions in the gel of oicked (nc), relaxed closed (re i), and supc rcoiled (sc) plasmids are marked by arrows.
and ftz plasmids). To 5 fLl of Ihe assembled chromalin
reaction was added 2 fLl of 0.5-JLg/fLl pUC veclor DNA and
18 ng of hsp70 minigene in 9 JLI of HEMG (25 mM HEPES
[pH 7.61, 0.1 mM EDTA, 12.5 mM MgCI 2 • 10% glycerol).
Four microliters of unshocked embryo transcription extract
(6) and 5 fLl of the following components (final concentrations) were added: 500 JLM (each) ATP, GTP, CTP, and
UTP; 4 mM creatine phosphate; 10 ng of creatine phosphokinase; 5 U of RNasin; 4 mM dithiothreitol. Thc 25·).1.1 reaetion
mixture was then incubaled for 20 min at 26°C. RNA
produelS were purified and analyzed by primer extension (6).
Quantitation of the radioactivity was performed on dried
polyacrylamide gel slices by liquid scintillation counting.
RESULTS
In vitro assembly of nudeosomes witb rqular spaeing on
plasmld DNA. In order 10 minimize Ihe dilution of cytoplas·
mic components and 10 reduee the extraetion of nuelear
proteins, the procedure for preparing the Drosophila nuele·
osome assembly extract involves the homogenization of
staged. preblastoderm embryos in a smalJ volume of low-sall
extraclion buffer. Nuelei are pelleted by centrifugation. and
the cytosolic supernatant is elarified by centrifugation al
150,000 x g (S-150). This cmde S·150 extract contains the
major components necessary to assemble nueleosomes on
microgram quantities of plasmid DNA. When supplemented
with ATP, MgCl 2, and an energy·regenerating system (3,
35), the assembly of regularly spaced nueleosomes is
achieved.
The ability of the Drosophila extract 10 assemble nueleo·
somes was inilially assayed by DNA supercoiling analysis
(13, 23, 29). The winding of DNA around a nueleosomal core
introduces about one positive superhelical turn in the DNA
(13), which is rapidly relaxed by topoisomerase I activity,
which is abundant in cell extracts. Deproteinization of the
reconstituted plasmid chromatin results in the acquisition of
one negalive superhelical turn far eaeh nuelcosome assem·
bled (21). Figure 1 shows the time course of plasmid supercoiling perfonned with Drosophila embryo extracts in Ihe
presence of increasing concentrations of KCI. Upon incubation with the S-150 extract (10 min), the negatively supercoiled plasmid was initially relaxed by topoisomerase I
activity. Further incubation resulted in Ihe reintroduction of
DNA supercoils, suggesling the assembly ofnueleosomal or
nueleosomelike structures. DNA supercoiling was found 10
be inhibited by the presence of KCI concentrations greater
than that already present in a standard reaetion (the standard
sah concentration is equivalenl in conductivity to 65 mM
KCI). $inee the supercoiling assay was performed with a
crude eXlract, it is difficult to ascribe the inhibitory effeci of
increased KCI 10 a spccific component of the reaction. The
effect of KCI concenlrations lower than 65 mM was nOI
analyzed, as the assay was constrained by the salt contribu·
lion of {he extracl.
To assess whether the supercoiling of the plasmid was due
10 the assembly of nueleosomes, we subjeeted Ihe assembled
plasmid to digestion wilh micrococcal nuelease. Micrococcal
nuelease eleaves chromatin in the linker DNA between the
nueleosomes (27). A limit digest of chromatin typically
creates DNA fragments of the size protected by a nueleosome core (146 nueleolides), while partial digests result in a
ladder of fragments corresponding to oligonueleosome-sized
DNAs. As s hown in Fig. 2A. digestion of the assembled
chromatin resulted in extremely well·resolved mono· and
oligonueleosome·sized DNA fragments. A densitometer
tracing revealed 19 distincI bands (Fig. 28), suggesling the
assembly of a maximum of 19 nueleosomes on somc of the
plasmid templates. On these templates, the average nueleosomal repeal length was ealculated to be 181 bp (3,443
bp/19). Measurement of the sizes of oligonueleosomal fragments also indicated the spacing of nueleosomes at - 180·bp
intervals. We have made over 12 extracl preparations by Ihis
procedure with similar results and hence have adopted this
as the s tandard prolocol. Thc resolution of the oligonueleosomal DNA bands over the background smear was less
apparent when the levels of MgCl 2 were decreased below 6.5
mM (Fig. 2C). For reasons which are unelear, extracts that
were prepared by adjusting Ihe embryo homogenate 10 6.5
mM MgCI 2 prior to (Fig. 2A) rather than afler (Fig. 2C) the
$·150 centrifugation were superior in their potential to ereate
extended nueleosomal arrays.
DNA syntbesis and Duc1eo$Ome assembly. In vivo nueleosome assembly occurs naturallyon newly replicaled DNA.
To test far the presence of similar aetivities in our extracts,
we substiluted the plasmid DNA with single·stranded eircu·
lar DNA and added [a_32P1dCTP as aprecursor far DNA
synthesis. Upon incubation with the extract, the comple·
mentary strand was synthesized and the bulk of the labeled
material was recovered in the fonn of supcrcoiled DNA (Fig.
3A). The assembly extracI apparently contains DNA pri.
mase, polymerasc(s), accessory factoTs, RNase H . and DNA
ligase activities necessary to perform lagging-strand DNA
A
MN • .., (mln(
1.5 mM
B
0.528 1'.1
19 18
17 16
15 14
nc, Tel
13
11 12
10
...
.
"
5
4
=
'"
m
'"
3
2
,%
m
1
6.5 mM
n
:-t
-- -l --
I- - -
• ---
..
"
"
~.
9
8
7
6
•
2.5 m M
"
"
•"
•
•
•
•
-·-
•"
•
•
•
•
"
•
•
•
--I, -- •
,,- I."
1
•
•
'r•
'ta
...
"•
•
FlO. 2. Micrococcal nuclease analysis of reconstilUted chromatin. (A) Chromatin as.sembled for 6 h under standard conditions was
digested with micrococcal nudease for Ihe indicated limes. The final Mg0 2 concentration during the assembly reaction was 7 mM. Purified
DNA fragments were resolved on 1.5% agarose gels and Slained with elhidium bromide as described by Shimamura et al. (35). Lane M shows
size markers consisting of 123·bp multimers (Bethesda Research Laboratories). (8) Densilometer Iraclng of Ihe O.5-min microc:occal nudease
digest shown in panel A. The arrow indicates the direction of e1ectrophoresis. (C) Assembly reactions containing the indicaled ooncentrations
of Mg01 were ineubated at 260C for 6 h. The reoonstitutcd chromatin was digested with mierococcal nudease for the indicaled limes. The
positions in the gel of nieked (ne), reJ.ued dosed (rel), linear (Iin). and superooilcd (sc) plasmids are marked by arrows.
synthesis. The level of ONA synthesis was found to be
stimulated by the presence of MgCl z up to 6.5 mM.
To confirm that the supercoils in the newly synthcsized
DNA originate from thc assembly of nudcosomes, mierococcal nudease digestion assays wcre pcrformed. First,
DNA synthesis was traced by the incorporation of radiolabcled dcrP as describcd above, in areaction that relied on
endogenous primcrs in the assembly extract. This uniform
labeling (Fig. 38, right panel) displayed the overall chromatin assembly on the plasmid . Altematively, a specific 32p_
end-Iabeled primer was annealed to the single-stranded
tcmplate befoTe the addition of the extract. In this reaction,
the labeled dCfP was omitted, and the micrococcal nudease
analysis thcrefore reHects chromatin assembly around the
site where the primer annealed (Fig. 38, Icft panel). In both
reactions with single-slranded lemplates, assembly of regularly spaced nucleosomes over a significanl fraction of the
te mplatcs was already evident by 30 min of assembly, as
gaugcd by the emergence of Ihe lad der of oligonucleosomes ized ONA fragments upon mierococcal nudease cleavage.
An increase in the duration of chromatin assembly (60 and
120 min) further cnhanced the ladder of DNA fragments,
indicating complete or near complete deposition of nucleosomes on the DNA template.
Histone content or in vitro assembJed Dudeosomes. The
regular spacing of thc mierococcal nuclease cleavage sites in
reconSlituted chromatin is indieative of the assembly of
nucleosomes. To extend this finding, we determined the
hislone composition of the reconstituted nucleosomes. Fully
assembled plasmid chromatin (with single- or doublestranded DNA as the starting material) was purified by
sucrose gradient centrifugation (35). The plasmid chromatin
proleins were analyzed by SOS-polyacrylamide gel electrophores is (SOS-PAGE) and silver slaining (Fig. 4). From the
reactions that contained either single-stranded (Iane 1) or
double-slranded (Iane 2) ONA, a full complement of core
histones of roughly equal stoichiometry could be recovered,
while areaction that was processed in parallel but lacked
DNA (Iane 3) displayed cosedimenting non histone proteins
that were nOI associated with chromatin. It should also be
noted that histone Hl was absent from the reconstiluted
chromalin, as no band of the expected s ize (-35 kDa; see
below) was observed. Preblastoderm Drosophila e mbryos
have been previously rcported to bc deficient in histone HI
(9). We have also observed an apparent histone variant in
reconstituted chromatin Ihat migrates above the hislone
H2B band on the gel. This variant is absent or poorly
represented among the laie e mbryo hisiones; the laie embryo histones include another apparent, minor hislone variant whieh migrates below the histone H2A position .
A
MgC~:
,
15 mM
65 mM
111204080
••
B
nC, tel _
lin _
--- -""
$Ue-speeific labeling
30'
60'
H
2
,----
102040l1li
···1
3
H
---
uniform labeling
120'
29kD -
•
2O k D -
----__
FIG. 4. Histone composition of reeonstiluted chromatin. Pro·
teins that oosedimented wilh assembled plasmids on a suerose
gradient were separated by SDS-PAGE and stained wilh sitver.
Nucleosome assembly reaetions contained either single·stranded
DNA (Iane 1), double-stranded DNA (Iane 2), or no DNA (Iane 3) as
the starting material. The positions of protein size Slandards are
indieated 10 the left. Lanes H: Marker core histones from chromatin
of 0- 10 2O-h embryos.
FIG. 3. (A) DNA synthesis in embryo extracts. Single-slranded,
circular DNA was incubaled in a standard assembly reaetion in Ihe
presence of MgCI 2 (concentrations as indicated) and [a· l2 PJdcrr.
Aliquots of the reaetion were analyzed for DNA superooiling by
eleclropboresis and autoradiography afler the indicated periods of
ineubation. (8) MicrococcaJ nuelease analysis of newly synthesized
plasmid DNA. Single·slranded DNA (660 ng) was ineubated with 2.6
mg of embryo extraet in a 150-..,.1 reaetion mixture in the presenee of
7 mM Mg0 2• Left: a_ J2 P_labeled oligonucleotide primer was annealed to the template prior to the addition of the extraet. Radioactive dCfP was omitted from the reaetion. Right: DNA synthesis
relied on endogenous primen and was traeed wilh radioaetive
dCTP. After incubation for the indicated times, the reaetions were
adjusted 10 3 mM eao 2 and digested with 190 U of micrococcal
nuelease (Boehringer Mannheim) for 0, 2. and 8 min (arrows indicate increasing digestion). DNA was analyzed by gel eleclrophoresis
and aUlOradiography. The gel on the left was eX])O$ed Ig times
longer Ihan the gel on the right. The positions in the gel of nicked
(nc), relaxed closed (rei), and linear (Iin) plasmids are marked by
arrows.
locorporatioo or exogenous HI Into reronstituted chroma·
lin. Histone Hl bas been suggested to play an important ro le
in organizing cbromatin and, in particular, to contribute to
the inactive state of a gene (8, 24, 34, 46, 51). Since
chromatin reconstituted with the early embryo extract was
deficient in Hl, we have delermined whcrher exogenous
purified Hl could be incorporated into the chromatin.
Hl was purified trom 0- to 12·h Drosophila embryos to
near homogeneity (Fig. 5A, lane 2) and was introduced
into the assembly extract prior to the addition of the DNA.
As shown in Fig. 5A, incorporation of exogenous Hl is
evident by SDS-PAGE analysis of the histone content of
assembled chromatin purified by sucrose grad ient centrifugation. Histone Hl cosedimcnted with the rcconstituted
chromatin when the assembly was carried out with doublestranded plasmid DNA (Fig. 5A, lane 4) or singlc-stranded
DNA (Iane 3), but not in the absence of DNA (Iane 5).
Furthermorc, when chromatin reconstituted in the presence
of Hl was analyzed by micrococcal nuelease digestion,
an increased nueleosome repeat length was observed (Fig.
SB). A densitometer tracing (Fig. SC, upper panel) showed
that the nueleosome repeat length was inc reased from
180 bp in the absence of Hl to -197 bp when Hl was
present; this length is essentially identical to the in vive
repeat length for postblastodcnn chromatin (Fig. SC, lower
panel).
Transcriptional repression or DNA lemplates assembled in
cbromatin. In order to assess the transeriptional potential of
chromalin templates assembled in vitro, two DNA templates
carrying the hsp70 aod fushi tarazu gene promoters were
assembled in the preseoce or absence of exogenous histone
Hl and assayed by in vitro transcriptioo. A naked, hsp70
minigene template was also introduccd into the transcription
mixture as a free DNA control. As shown in Fig. 6, incrcasing nueleosomc assembly resulted in the progressive repression of transcription from beth hsp70 and fushi larazu
lemplates. By 60 min of chromatin assembly, when the DNA
templates we re assembled in a regular nucJeosomal array as
gauged by the appearance of a defined ladder of DNA
fragments after micrococcal nuelease digestion (dala not
2
3
-
A
36kO 29100 -
36kO 29kO -
4
--
5
6
MNase
[minI
-HI
1
- Hl
o2
832
1 1
+ Hl
o2
1
832
B
2<11:02()
C
kO -
--
liili
•
_
H'
_
H28
-H>A
H'
I" .... tro
•
'n
."'"
'"
- - -.. --
'"
'"
'"
FlG . 5. Incorporation of exogenous histone Hl inlO assembled chromatin. (Al Chromatin assembly reactions (400 ","I) that conlained eithe r
single·stranded DNA (lane J). double-stranded DNA (lane 4), or no DNA (lane 5) and 15 jl.1 of Hl we rt fractionated by sucrose gradient
centrifugat ion. Proteins thaI cosedimented will! the assembled minichromosomes (or thc equivalent tractions in thc reaction lacking DNA)
we rt pelleted through a sucrosc cushion, analyzed by SDS·PAGE, and sitver stained. Lanes 1 and 6, COTe histon es purified (rom thc chromatin
ofO- 10 20-h embryos; lane 2, histone Hl (5 jl.l). The slaining of Ihe histone H2A band in lanes 3 and 4 is low, oompared with that of the histone
H2B band. (B) Micrococcal nuclease analysis of 1.5 ..,.g of plasmid DNA, assembled as described in the Fig. 2A legend for 6 h in the absence
or presence of 12 ...1of Hl. Siu markers are as in Fig. 2A. (q Upper panel: densitometer traeing of the lanes displaying the 2·min digestion
produets of panel B. The tracings were aligned with respeet to the neighboring marker lanes 10 IHustrale the change in repeat length. Lower
panel: nuelei isolated from 6- to 18-h embryos were digested with micrococcal nuelease, and the digestion produets were analyzed on a 1.3%
aga rose gel. Densitometer !raeings of rwo lanes showing different extents of digestion are provided.
shown), cssenlially 10lal repression was observed whether
or not histone Hl was incorporated into Ihe c hromalin
tcmplale (Fig. 6A, lanes 5 and 11). Maximal repression was
obSerVed even in Ihe absence of full nucleosome assembly
(which is accomplished in 6 h). As in previous studies, Ihese
findings iQdicate Ihat a high, bul not necessarily maximal,
density of regu)arly spaced nucleosome core particles is
suffident 10 repress transcription in vitro. AI early time
poinls of Ihe asscmbly reaction (5, 10, and 20 min), when
nucleosome assembly is suboptimal. Ihere was a modest
increase in the level of repression that was dependenl on
exogenous histone Hl (Fig. 6A; compare lanes 2 to 4 with
lanes 8 to 10) (see Fig. 68 for quantitation). This requiremenl
for histone HI in transcriptional repression when nucleosomes are assembled in vitro at low density is very similar to
what was previously reported concerning the effeci of histone Hl on transcriplion by Xenopus RNA polymerase III
(34, 35).
DlSCUSSION
We have characlerized an in vilro nucleosome assembly
extract from preblastoderm Drosophila embryos. Early
Drosophila embryos replicate their genomes on average
once every 9 min (11) and assemble the newly synthesized
DNA into chromatin by using the maternal pool of histones
and assembly factors. They are Ihus an excellent source of
the components requircd fOT replication and chromatin assembly. Thc cxtraction procedurc we have employed was
adapted from recent s tudies of in vitro chromatin assembly
that used Xenopus oocytes as thc starting material (34, 35).
The Drosophila assembly system is characterized by a
number of features that render il a highly useful allernative
for studies on chromatin assembly. (i) Microgram amounts
of plasmid DNA can be quantitatively assemblcd into chromatin carrying the maximal number of regularly s paced
nucleosomes. (ii) The nucleosomes assembled in viuo re-
A
• HI
-..
2
3
4
• H1
5 6
~-
7
8
9 10 11 12
---
hsp70
__ _
ftz
• • • - • • • • • Ji: • •
----
hsp70 (frec)
,.,B
,.s
,.,
_
ft:z·Hl
_
ttz .. Hl
••
'.2
",
"
"
,.,
,.s
c
.2
~
L
U
"C
•
I-
-0--
_
hsp70 ·Hl
hsp70 . Hl
,.,
,..
L
'.2
,., ,
"
"
Chromatin assembly {min}
"
FlG. 6. Transcriptional repression of DNA tcmplales assembled
in chromatin. (Al Plasmids containing the hsp70 and rush; tarazu
iftz) promoters wert assembled for O. S, 10,20. aod 60 min (Ianes 1
10 5 and 7 10 11) in the absence or prestoc!: of lIi510ne Hl, as
indic8lcd. After quenching the assembly ruelion wilh excess carrier
DNA, Im hsp70 minigene was introduccd u free DNA and thc
mixture was incubated in an in vitra transcriplion C"lr8CI. The RNA
producls dcrivcd from each template 3fe displayed by primer
extension (6). Thc transcription reactions in lanes 6 and 12 lackcd
plasmid chromatin templates and served as controls fOT subtracting
background radioactivity for signals dcrived from the chromatin
templates. (8) Graphical representation of the level of transcription
of the chromatin templales relalive 10 transcription of the free
template.
semble native core particles with regard to the stoichiometry
of core histone content and the spacing of the nucleosomes
on the DNA. (iii) As in the Xenopus system (1-4), the
Drosophj/a extraci can be uscd eilher to assemble nucleosomes onto double-slranded plasmid DNA or in conjunction
with DNA synthesis slarting from a single-stranded templale. Chromalin that does or docs not contain histone Hl
can be obtained. (iv) Large quantities of eXlracts with high
activities are easily and reproducibly prepared according 10
the standard prolocol. In particular, we have not encounlered Ihe seasonal variation in the activity of extracts prepared from frog oocyles. The eXlracts can be kept for many
months al -BOOC and can be Ihawed aod ffozen several times
withoul noticeable loss of activity.
There are several technical points important 10 Ihe overall
activity of the assembly extract Ihat should be noled. The
Drosophila embryos are homogenized in a low volume of
extraction buffer in order 10 minimize the dilution of cytoplasrnalic assembly componcnts and, presumably, 10 approach physiological assembly condilions. The extraci
buffer eonlains low salt, 10 mM KCI, which minimizes
extraction of nudear proteins inlo the cytosol. The extended
centrifugation at 150,000 x g in the presence of 5 mM
magnesium pellets ribosome subunits from the exiract thaI
would otherwise obscure the analysis of plasmid chromatin
proleins (35). The low ionic slrength of Ihe extraction buffer
also enables thc assembly reaclion 10 be carried out in the
presence of an energy regeneraling syslem (ATP aod creatine phosphale) and 7 mM MgCl 2 , which are necessary for
the assembly of long arrays of spaced nucleosomes. lmportanlly, the assembly relies enlirely on the endogeoous pool
of malernal histones and Iheir carrier proleins and assembly
faclors, Ihe native slale and sloichiometry of which is
maintained. This may be cruciaJ for the reconstitution of
long arrays of spaced nucleosomes on natural DNA sequences, a property lacking in chromatin reconstituted by
using polyanions as an assembly vehicle for exogenous
hisIones (e.g., see rcferences 26, 33, and 41 and references
therein).
The pool of componenls required for proper nucleosome
assembly should be highesl in the early embryos and should
be rapidly depleted as replication proceeds. Indeed, older
Drosophila embryos (4 to 6 h afler egg laying) yie1ded
significantly weaker extracts (dala not shown). Exlracis
from Xenopw.· oocyles, which also contain a maternal hislone pool, are similarly efficient in quantitatively assembling
large amounts of plasmid DNA inlo chromatin while maintaining a high degree of regular nucleosomal spacing (34. 35).
Tissue culture celJs may not be as rich in assembly components, which suggests why cell-free chromatin assembly
systems derived from mammalian cells appear 10 be less
potent (5, 15).
In agreement with previous reporls (2, 5, 30), e levated
concentrations of MgCI 2 are required 10 reconstitute extensive arrays of spaced nudeosomes in our celJ-free system
(Fig. 2). Interestingly, we have observed rapid supercoiling
of plasmid DNA in the exlract at prolein concentra!ions
insufficie n! for !he generation of spaced nudeosomes and at
a suboplimal MgCI 2 concentration (L5 mM; data not
shown). Whether Ihis supercoiling is a result of Ihe winding
of the DNA a roun d subnucleosomal particles, such as
H3-H4 tetramers, remains 10 be established. Such parlic1cs
a re likely to be intennediates in the assembly of nucleosome
cores (1, 12, 18, 26, 32, 39).
Thc lack of histone Hl in chromatin reconstituted by the
early embryo extract is not due to a deficiency of the
assembly machinery, since exogenous Hl can be efficiently
assembled, resulting in an increased linker Jength. The
nudeosomal spacing of chromatin assembled in the presence
of exogenous Hl is very similar to the natural spacing of bulk
chromatin in postblastoderm embryos. Hence, chromatin
assembled wilh exogenous histone Hl in vitro dosely apo
proximates the physiological structure. Although Hl assem·
bly is a determinanl of the average linker length between
core partides, il appears not to be required for regular
spacing of nudeosomes on the DNA itselL Such a function
may be dependent on other factors present in the extraCI,
analogous to those found in the Xenopus system and mammalian systems (32, 38). lt will be of interest to further
delermine at nudeotide resolution whether the specifie positions of nudeosomes assembled in vitro under physiological conditions are exactly equivalentto those found in vivo
or to those reconstituted by using other procedures, such as
salt dialysis.
Arccent study with chromatin templates reconstiluled
with core histones and histone Hl has implied a significant
role for histone Hl in Ihe repression of transcription by RNA
polymerase 11 (24). The present results with the hsp70 and
fushi tarazu promoters, and previous reconstitution sludies
with Xenopus oocyte eXlracts, indicate, however, thai the
optimal assembly of regularly spaced nudeosome eores is
sufficient to maximally repress Iranseription in vitro, even in
the absence of histone Hl. lt is probable thai Ihe different
effects obselVed with histone Hl are due to the different
procedures employed for core histone deposition, which
lead to differences in the spacing of reconstituted nudeosomes. It is also possible thai a requirement rOT repression
by histone HI is specific to the individual promoter sequences used in the separate studies.
The ability to assembte differenl types of chromatin unde r
near-replicative condilions in vitro opens avenues for studying the compelition between histone deposition during DNA
synthesis and the binding of factors that govern transcriptional activation. Moreover, Ihe preblasloderm cytoplasmic
extract itself does not support RNA polymerase 11 transcription (data nOI shown), in contrast to the nudear extracls
de rived from later embryos which are highly aclive for
transcription (20, 40). An extract for nucleosome reeonstitution that is free of interfering transcription initiation by
polymerase 11 should be a particularly useful alternative to
the assembly systems derived from Xenopus eggs and
oocytes for sludies on the interrelationships between transcriptional regulators and chromatin structure (6, 7, 22, 47,
48).
ACKNOWLEDGMENTS
We thank A. Wolffe for helpful suggestions and G. Wall for
excellent technical assistance with the trans.criplion experiments.
We also thank {wo referees for helpful comments.
P.B.B. was supporled by a fellowship from the Fogarty International Center.
REFERENCES
1. A1mouznl, G., O. J. Clark, M. Mkhall, and A. P. Wolffe. 1990.
Chromatin assembly on replicating DNA in vitro . Nucleic Acids
Res. 18:5767- 5774.
2. Almouznl, G., Ind M. M~chlli. 1988. Assembly of spaeed
cllromatin promotcd by DNA synthesis in e",tracts from XenopUS eggs. EMBO I. 7:665-672.
3. A1mouznl, G., and M. Mkhall. 1988. Assembly of spaced
chromatin: involvement of ATP and DNA topoisomerase activ.
ity. EMBO J. 7:4355-4365.
4. A1mouJ.nl, G., M_ Mkhati, Ind A. P. Wollre. 1990. Competition
S.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
16.
17.
18.
19.
20.
21.
22.
23.
24.
25.
26.
27.
between trans.cription complex assembly and chromatin assembly on replicating DNA. EMBO J. 9:573-582.
Banerjee, S., Ind C. R. Canlor. 1990. Nucleosom e assembly of
simian virus 40 DNA in a mammalian cell extract . Mol. Cell.
Biol. 10:2863-2873.
Becker, P. B., S. K. Rabindran, and C. Wu. 1991. Heat
shock-regulated trans.cription in virro from a reconstituted chro·
matin template. Proc. Natl. Acad. Sei. USA 88:4109-4113.
Cort h&y, B., P. Uonard, Ind W. Wahli. 1990. Trans.criptional
polentiation of the vitellogenin BI promoter by a combination of
both nucleosome assembly and trans.cription factors: an in vitro
disseclion. Mol. Cell. Biol. ID: 3926-3933.
Croston, G. E., L E. Kerripll, L M. Ura, D. R. Mlnhlk, I nd
J. T_ Kadonagl. 1991. Sequencc-spedlic anlirepression of his·
tone Hl-mediated inhibition of basal RNA polyme rase 11 lranscription. Seiencc 251:643-649.
EI,in, S. C. R., Ind L E. Hood. 1973. Chromosomal proteins of
Drosophila embryos. J. Biol Chem. 12:4984-4991.
Fischer, K., J. Schmitz, Ind W. Hön. 1990. Role of trans·
activating proteins in the gene ration of active chromatin at the
PH05 promoter in S. cenvisiae. EMBO J. 9:2523-2528.
Foe, V. E., Ind B. M. A1btrU. 1983. Sltldies of nuclear and
cytoplasmic behavior in the five mitotic cycles that precede
gastrulation in Drosophila embryogenesis. J . Cell. Sci. 61:3110.
Fotedlr, R., Ind J. M. Roberts. 1989. Multistep palhway for
replication-d ependent nucleosome assembly. Proc. Natl. Acad.
Sei. USA 86:6459-6463.
Gennond, J. E., P_ Hirt, M. Dudel, P_ Grou-Ikllard, Ind P.
Chambon. 1975. Folding the DNA double helix into chromatinlike structures from simian virus 40. Proc. Natl. Acad. Sei. USA
72:1843-1847.
Gross, O. S., Ind W. T. Glrrard. 1990. Nuclease hypersensitive
sites in chromatin. Annu. Rev. Biochem. 57:159-197.
Gruss, C., C. Gutitrnz, W. C. Rumins, M. L DePlmphllls, T.
Koller, and J. M. Sogo. 1990. Nucleosome assembly in mammalian cell extracts before and after DNA replication. EMSO J.
9:2911- 2922.
Hin, M., Ind M. Grunsteln. 1988. Nucleosomc loss activates
yeast downstream promoters in vivo. Cell 55:1137- 1145.
Han, M., U. Kirn, P. Karne, Ind M. Grunstdn. 1988. Depletion
of histon e H4 and nucleosomes activates the PH05 gene in
Saccharomyces cerevisiae. EMBO J. 7:2221- 2228.
HlnStR, J. C., K_ E. van Holde, Ind D. Lohr. 1991. The
mechanism of nucleosome assembly onlo oligomers of the sea
urchin 5S DNA positioning seque nce. J. Biol. Chem. 266:42764282.
Johnson, L M., P. S. Kayne, E. S. Kahn, Ind M. GllInstein.
1990. Genctic evidencc for In interaction be{Ween SIR3 and
histone H4 in the repression of the silent mating Iod in Saccharomyces cenvisiae. Proc. Natl. Acad. Sei. USA 87:6286-6290.
K.adonlp, J. T. 1990. Assembly and disassembly of the Drosophila RNA polymerase 11 complex during trans.cription.l. Biol.
Chem.265:2624--2631.
Keller, W. 1975. Determination of the number of superhelical
turns in simian virus 40 DNA by gel electrophoresis. Proc. Nltl.
Acad. Sci. USA 72:4876-4S80.
Laskey, R. A., Ind W. C . Eamshlw. 1980. Nucleosom e assembly. Nature (London) 286:763-767.
Laskty, R. A., A. D. MlIIs, I nd N. N. Morris. 1977. Assemblyof
SV40 chromatin in a eell·free system from Xenopus eggs. Cell
10:237_243.
Layboum, P. J., I nd J . T. Kadonlgl. 1991. Role of nucleosomal
cores and histone Hl in regulation of trans.cription by RNA
polymerase 11. Stienet 254:238-245.
Nelson, T., T. Hsieh, Ind O. Brutli" 1979. Extracts of Drosophila e mbryos mediate ch romatin assembly in vitro. Proc. Nat!.
Acad. Sei. USA 76:5510-5514.
Nelson, T., R. Wiepnd, Ind O. BrutllJ_ 1981. Ribonucleic acid
and other polyanions faci lilate chroma tin assembly in vitro.
Biochemistry 20:2595- 2601.
NolI, M., Ind R. D. Kombtra:. 1977. Action of micrococcal
nuclease on chromatin and the location of histone Hl. J. Mol.
Biol. 109:39J-404.
28. Ptder50D, D. S., Ind R_ T. Simpson. 1988. Structural and
reguiatory lIierarcllies in transcriptionally active cllromatin. ISI
Atlas Sei. Biocllem. 1:155-160.
29. Pulleybllnk,. D. E., M. SlI u ~, D. Tlng, J. Vinograd, Ind H. P.
Vosberg. 1975. Act ion of nicking-closing enzym e on supcrcoiled
and nonsupcrcoiled closed circular DNA: formation of BoItzman distribution of topological isomers. Proc. Natl. Acad. Sei.
USA 71:4280-4284.
30. Rodrlgue:r.-Clmpos, A., A. Shlmllllunl, Ind A. WOfftI. 1989.
Assembly and propcnies of chromatin containing histone Hl. J.
Mol. Biol. 209:13~ lSO.
31. Rnth, S. Y., A. Dei n, Ind R. T. Slmp$Oo. 1990. Yeast 012
repressor positions nucleosomes in TRPlJARSl chromatin.
Mol. Ce ll. Bio l. 10:2247-2260.
32. SIPP, M., Ind A. Wo~l. 1990. Purification and mechanism of
action of a nucleosome assembty factoT from Xenopus oocytes.
J . Biol. Chem. 2.65:9357-9365.
33. Sealy, L, R. R. Burgus, M. M. COllen, and R. Chllkty. 1989.
Purification of Xenopus tgg nucl eoplasmin and its use in chromatin assembly in vi/ra. Methods Enzymol. 170:612-630.
34. SbimlmunI, A., M. Sapp, A. Rodriguez-eampos, I od A. Worcel.
1989. Histone Hl represses transcription from minich romosornts assembled in vilro. Mol. Cdl. Biol. 9:5573-5584.
35. Shima mura, A., D. Tftmelbkk, aod A. Worul. 1988. Characterization of the repressed 5S DNA minichromosom es assem bled in vilTO with a high·speed supcrnatanl of X enopus laevis
oocytes. Mol. Cdl. Bio!. 8:4257-4269.
36. SIlIIon, R. H., Ind G. FtlstDkld. 1979. A new procedure for
purifying histone pairs H2A + H2B and H3 + H4 from
c hromatin using hydToxylapatite. Nucleic Acids Res. 6:689-696.
37. Simp$OD, R. T. 1991. Nuc leosome positioning: occurrence.
mechanisms, and functional consequences. Prog. Nucleic Acid
Res. Mol. Biol. 40:143-183.
38. Smitb, S., I nd B. Stillmln. 1989. Purification and cha racteriza tion of CAF-l, a human cell factor required for chromatin
assembly during DNA replication irr vitra. Cell 58:15-25.
39. Smitb, S., I nd B. Stlllmin. 1991. 5tepwise assembly of ch romatin during DNA replica tion in vitra. EMBO J . 10:971- 980.
40. ~I ~r, W. C., S. J. Poole, aod T . Knmberg. 1988. In vitra
transcript ion of the Drosophila engrailed ge ne. Genes Dev.
2:63-8l.
41. SIei n, A. 1989. Reconstitlltion of chromatin from purified components. Method s Enzymol. 110:585-603.
42. Slraka, C., Ind W. Hön. 1991. A functional role for nucleosomes in the repression of a yeast promoter. EMBO J . 10:361368.
43. Suren, J., Ind R. Chllkey. 1990. Tbe s tructure and assembly of
active chromatin. Trends Genet. 6:52-55.
44. von Holt, C., W. F. Bnlodt, H. J. Greyling, G. G. Und~y, J. D.
Ret!tf, J. D. A. Rodrlgues, S. $ehwager, and B. T. Sewell.
Isolation and characterization of histo nes. Methods Enzymol.
170:431-523.
45. WollI'e,A. P.I990. New approaches tochromatin function. New
Biol. 2:211-218.
46. WollI'e, A. P. 1989. Dominant and specific re pression of Xeno·
pus oocyte 55 RNA genes and satellitt 1 DNA by histone H1.
EMBO J. 8:527-537.
47. Workmao, J. L, R. G. ROotder, aod R. E_ Klngston. 1990. An
upstream transcription fac tor, U5F (MLTF), facilitates the
formation of pre-initiation complexes during in vitra chromatin
assembly. EMBQ J. 9:1299-1308.
48. Workma o. J. L. I. C. A. Taylor. a nd R. E. Klngston_ 1991.
ACliv3tion domains of s tably bound GAlA derivatives alleviate
repression of promote rs by nucleosomes. Cell 64:533-544.
49. Wray, W., T. Boulikas, Y. P. Wray, a nd R. Hlncock. 1981.
Silver s taining of proteins in polyacrylamide gels. Anal. Bioehern. 118:197-203.
50. Wu, C., P. M. BIogblm, K. J. Unk, R. Holmgre n, a nd S. R. C.
Elgin. 1979. Tbe chromatin structliTe of spccitic genes. I.
Evidence fOT high er order domains o f defined DNA sequence.
Cell 16:797-S06.
51. Zlatallova, J. 1990. His tone Hllnd the re gulation of transcriplion of eukaryotic genes. Trends Biochem. Sei. 15:273-276.