Báo cáo khoa học: Nucleosome positioning in relation to nucleosome spacing and DNA sequence-specific binding of a protein doc

Nucleosome positioning in relation to nucleosome spacing
and DNA sequence-specific binding of a protein
Rama-Haritha Pusarla*, Vinesh Vinayachandran* and Purnima Bhargava
Centre for Cellular & Molecular Biology, Hyderabad, India
Multifold compaction of DNA due to the presence of
nucleosomes on natural templates of eukaryotic RNA
polymerases results in transcriptional repression. Sev-
eral studies have established that histones and nucleo-
somes play an active role in regulating gene expression
in eukaryotic cells [1,2]. Gene-specific, localized config-
urations of in vivo chromatin in various genome
regions are found due to precise positioning of nucleo-
somes over the underlying DNA stretches [3,4]. Both
trans-acting factors and DNA sequences can determine
where histones will occupy the bound DNA. A nucleo-
some is positioned translationally when its histone–
DNA contacts are restricted to an identifiable stretch
of DNA, giving clear boundary zones. Further preci-
sion of the positioning can be achieved by restricting
the rotation of DNA over the histone octamer surface
(rotational setting), resulting in a defined phase ⁄ orien-
tation of a particular base pair with respect to
histones. Nucleosomes on certain constitutively active
genes can be excluded due to rapid and tight associ-
ation of trans-acting factors with promoter elements
during the replication-coupled assembly of chromatin
in vivo [5]. On other genes, they are removed or
reshuffled through several chromatin-remodeling and
Keywords
chromatin assembly; ionic strength;
nucleosome positioning; nucleosome

spacing; protein boundary
Correspondence
P. Bhargava, Centre for Cellular & Molecular
Biology, Uppal Road, Hyderabad-500007,
India
Fax: +91 40 27160591
Tel: +91 40 27192603
E-mail:
*These authors contributed equally to this
work
(Received 12 October 2006, revised 2
March 2007, accepted 7 March 2007)
doi:10.1111/j.1742-4658.2007.05775.x
Nucleosome positioning is an important mechanism for the regulation of
eukaryotic gene expression. Folding of the chromatin fiber can influence
nucleosome positioning, whereas similar electrostatic mechanisms govern
the nucleosome repeat length and chromatin fiber folding in vitro. The
position of the nucleosomes is directed either by the DNA sequence or by
the boundaries created due to the binding of certain trans-acting factors to
their target sites in the DNA. Increasing ionic strength results in an
increase in nucleosome spacing on the chromatin assembled by the S-190
extract of Drosophila embryos. In this study, a mutant lac repressor protein
R3 was used to find the mechanisms of nucleosome positioning on a plas-
mid with three R3-binding sites. With increasing ionic strength in the pres-
ence of R3, the number of positioned nucleosomes in the chromatin
decreased, whereas the internucleosomal spacings of the positioned
nucleosomes in a single register did not change. The number of the posi-
tioned nucleosomes in the chromatin assembled in vitro over different plas-
mid DNAs with 1–3 lac operators changed with the relative position and
number of the R3-binding sites. We found that in the presence of R3,

nucleosomes were positioned in the salt gradient method of the chromatin
assembly, even in the absence of a nucleosome-positioning sequence. Our
results show that nucleosome-positioning mechanisms are dominant, as the
nucleosomes can be positioned even in the absence of regular spacing
mechanisms. The protein-generated boundaries are more effective when
more than one binding site is present with a minimum distance of
 165 bp, greater than the nucleosome core DNA length, between them.
Abbreviations
IEL, indirect end-labeling; IPTG, isopropyl thio-b-
D-galactoside; MNase, micrococcal nuclease; NRL, nucleosome repeat length.
2396 FEBS Journal 274 (2007) 2396–2410 ª 2007 The Authors Journal compilation ª 2007 FEBS
chromatin-modifying mechanisms. Thus, cells use
nucleosome positioning as a mechanism to include or
exclude the binding sites of trans-acting factors from
accessible chromatin regions by passively restricting
the position of nucleosomes therein [6,7]. It is now well
established that a nucleosome is not necessarily repres-
sive; rather, it can facilitate the activation of genes.
Specific positioning of the nucleosomes allows the
transcriptional machinery to work effectively in a chro-
matin environment. Folding of DNA by the histones
and positioned nucleosomes can bring two widely sep-
arated regulatory elements into juxtaposition in space
[8–10] or even orient the bound factors for productive
interactions.
Nucleosomal repeat length (NRL) is characteristic
for a species, suggesting that it is a regulated feature
of the chromatin [11]. Uniform spacing of nucleosomes
is proposed to promote the higher-order folding of
chromatin [12]. Chromatin folding, in turn, is reported

to influence nucleosome positioning [13]. The presence
of positioned nucleosomes at defined and limited loca-
tions may result in disruption of the uniformity in spa-
cing. However, it is not known whether nucleosome
spacing influences nucleosome positioning. The
observed longer repeat lengths on inactive genes as
compared to those observed on transcribed sequences
[14,15] suggest that the folding of the 10 nm chromatin
with beads on a string into a higher-order structure
requires a minimum spacing to be maintained between
core particles. Thus, nucleosome spacing and position-
ing appear to be correlated.
Ionic strength is reported to influence nucleosome
conformations [16,17] as well as their spacings [18].
Within an array of positioned nucleosomes, the ionic
strength effect dominates the sequence effect [19]. It
also influences chromatin folding, presumably by
modulating H1 association as well as interparticle
interactions [20,21]. Of the two methods of chromatin
assembly in vitro [22], the salt gradient dialysis method
deposits nucleosomes in a random fashion, and has
been useful for checking the ability of various DNA
sequences to position nucleosomes in vitro.ADro-
sophila embryonic extract, in contrast [23], can deposit
nucleosomes with regular spacing in a sequence-inde-
pendent manner in the presence of ATP. The in vitro
chromatin assembly carried out by cellular ⁄ nuclear
extracts, giving uniformly spaced nucleosomes, is affec-
ted by parameters such as ionic strength, concentration
of linker histones, protein phosphorylation, and the

presence of core histone tails [18,24]. Spacing of nucle-
osomes is also influenced by DNA topology or histone
variants [19,25]. Using this system, binding of a
mutant lac repressor R3 (a small sequence-specific
DNA-binding prokaryotic protein) to its two sites,
183 bp apart, was shown to result in at least five trans-
lationally positioned nucleosomes in a single register
on a plasmid DNA [26]. In general, boundaries gener-
ated by proteins binding to the DNA restrict the
randomization of nucleosome positions [27]. It was
predicted that the nucleosomes close to a boundary
would be precisely positioned, whereas this precision
would decrease with distance from the boundary [28].
We have analyzed the effect of changing the ionic
strength and the number and spacing of the binding
sites on the protein-generated boundary for nucleo-
some positioning in both assembly systems. We have
found that the number of positioned nucleosomes in a
single register changes with the ionic strength of the
medium, although the spacing between them does not
change. The range of positioning effects of DNA
sequence-specific binding of a protein to chromatin
depends on the number of sites and the distance
between them.
Results
All the chromatin assemblies were constructed using
rat liver core histones and plasmid DNAs schematical-
ly depicted in the Fig. 1, in the presence or absence of
the R3 protein.
Number of positioned nucleosomes changes

with ionic strength
R3 is a mutant lac repressor protein that binds the lac
operator as a dimer with the same affinity as that of
the wild-type lac repressor but fails to tetramerize [29].
Binding of R3 protein to a plasmid DNA pU6LNS
(two lac operators at a distance of 183 bp) was reported
A
B
Fig. 1. Relative positions of the three lac operator sites in the plas-
mid DNAs. (A) Diagrammatic representation of the plasmid con-
structs with different numbers of lac operators. The solid
rectangular boxes and L1, L2 and L3 denote the first, second and
third lac operator sites. The distance in bp between each site is
given. (B) Schematic map of plasmid d35 with the shortest dis-
tance between L1 and L2.
R H. Pusarla et al. Mechanisms of nucleosome positioning
FEBS Journal 274 (2007) 2396–2410 ª 2007 The Authors Journal compilation ª 2007 FEBS 2397
to result in the positioning of an array of five nucleo-
somes in a single register [26]. As R3 can bind to the
naked DNA at salt concentrations as high as 0.4 m
(not shown), we looked at the effect of increasing ionic
strength on nucleosome positioning due to R3 binding
on the plasmid pU6lac3 by using the indirect end-
labeling (IEL) method of chromatin structure analysis
(Fig. 2). The micrococcal nuclease (MNase) digestion
pattern of the naked DNA in the IEL analysis
(Fig. 2A, lane 1; Fig. 2B, lane 4; Fig. 2C, lanes 5 and
6) did not change with the binding of R3 (Fig. 2A,
lane 2; Fig. 2B, lane 5; Fig. 2C, lanes 4 and 7), as
small footprints could not be resolved in the agarose

gels. The digestion pattern did not change even with
the deposition of histones (Fig. 2A, lanes 3, 5 and 6;
Fig. 2B, lanes 1, 6 and 7; Fig. 2C, lanes 2, 3 and 8) at
every ionic strength, suggesting there are no preferred
locations for nucleosome assembly on the plasmid.
However, in the presence of R3, in a single register,
five positioned nucleosomes at a salt concentration of
50–90 mm (Fig. 2A, lanes 4, 7 and 8; Fig. 2B, lanes 2
and 3) and three positioned nucleosomes at a salt con-
centration of 110 mm (Fig. 2B, lane 8) could be seen.
In comparison to this, two positioned nucleosomes
could be seen even at a salt concentration of 130 mm,
whereas none were seen at a salt concenration of
150 mm (Fig. 2C, lanes 1 and 9). These results show
that with increasing ionic strength, fewer positioned
nucleosomes are aligned in a single register.
As positioned nucleosomes are seen in Fig. 2 only in
the presence of R3, we used DNaseI footprinting to
confirm the specific binding of R3 at L1 and L2 at all
the salt concentrations. As no positioned nucleosomes
were seen at a monovalent salt level of 150 mm in the
IEL analysis of Fig. 2C, no protection was seen
between L1 and L2 on the DNA subjected to chroma-
tin assembly in the representative gel at this ionic
strength (Fig. 3A). Chromatin assembly in our system,
as judged by the generation of MNase-resistant nucleo-
somal ladders, was found to be normal up to 110 mm
salt, whereas only a few nucleosomal bands could be
seen at higher salt levels (not shown). The binding of
R3 did not disrupt the nucleosomal ladders close to

the three distinct sites located at different distances
from each other (Fig. 3B). Therefore, the loss of posi-
tioned nucleosomes at higher ionic strengths is not due
AB C
Fig. 2. Nucleosome positioning in the presence of R3 at different ionic strengths. IEL analysis of the chromatin structure of plasmid pU6lac3
assembled with S-190 extract in the absence or presence of R3 at various salt concentrations. Nucleosome positions are numbered and
marked with ellipses, and arrowheads indicate the positions of lac operators, marked L1–L3. The total number of the positioned nucleo-
somes in a single register is given under each salt level. The 5¢-end of the radiolabeled oligonucleotide probe (arrowhead) hybridized 709 bp
downstream of the third lac operator, L3, as shown at the bottom of the figure. (A) Structure analysis of the chromatin assembled at 50 or
70 m
M salt. Lanes 1 and 2 show naked DNA digestion patterns, and lanes 3–8 represent chromatin. R3 is absent in lanes 1, 3, 5 and 6,
whereas chromatin was assembled at 50 m
M salt for lanes 1–4, and at 70 mM salt for lanes 5–8. (B) IEL analysis for the chromatin assem-
bled at 90 and 110 m
M salt concentrations. Naked DNA (lanes 4 and 5) and chromatin assembled in the absence or presence of R3 protein
are shown. R3 was added to lanes 2, 3, 5 and 8. (C) IEL analysis for the chromatin assembled at 130 and 150 m
M salt. Naked DNA
(lanes 4–7) and chromatin assembled in the absence or presence of R3 protein are shown. R3 was added to lanes 1, 4, 7 and 9.
Mechanisms of nucleosome positioning R H. Pusarla et al.
2398 FEBS Journal 274 (2007) 2396–2410 ª 2007 The Authors Journal compilation ª 2007 FEBS
to the absence of R3 binding, but is probably related
to the reported effects of the ionic strength on the bulk
chromatin properties. As an increase in the ionic
strength is reported to increase the NRL [18], the total
number of uniformly spaced nucleosomes on a plasmid
DNA may decrease, which would result in a change in
the topological state of the plasmid. Therefore, we
confirmed the ionic strength effects on nucleosomal
density on plasmid DNA by the one-dimensional
supercoiling assay in the presence of chloroquin. The

plasmid in this assay was not relaxed prior to chroma-
tin assembly, as the S-190 extract is known to have
topoisomerase I activity, and assembly in this system
proceeds to completion. Therefore, under the condi-
tions of the gel run in Fig. 3C, all of the resolved
bands may be positively supercoiled topoisomers,
which differ by one linking number [30]. Comparison
of the profiles of the topoisomers (Fig. 3D) showed a
downward shift of the mean of the Gaussian distribu-
tion at different salt concentrations, denoting an
increase in the linking number of the DNA. As chloro-
quin introduces positive supercoils into the DNA, this
shift confirms that with increasing salt concentration,
there is a change in the superhelical density, which is
caused by a decrease in the number of nucleosomes
deposited over the plasmid DNA. In contrast to this,
the topoisomer profiles in the presence and absence of
R3 protein at different salt concentrations showed a
perfect overlap (Fig. 3E), suggesting that binding of
R3 at these ionic strengths did not change the nucleo-
some spacing further. Although the chromatin assem-
bly appeared to be better in the absence of R3
AB
C
E
D
Fig. 3. Influence of ionic strength and R3 binding on chromatin assembly. (A) Binding of R3 to its sites in the chromatin. High-resolution
DNaseI footprinting gel shows digestion profiles of the naked DNA (N) and chromatin assembly (C) in the presence and absence of the R3
protein. Chromatin was assembled with S-190 extract over plasmid pU6lac3 at a monovalent salt concentration of 150 m
M. Comparison of

the profiles of lanes 3 and 4 in the right-hand panel shows protection due to R3 at L1 and L2 (gray boxes) but not between them. The pri-
mer was located 93 bp upstream of L1. GATC shows the sequencing ladders generated by the same primer. (B) MNase-resistant nucleo-
some ladders from chromatin assemblies were resolved on 1.25% agarose gels, Southern transferred, and probed with a primer that
hybridizes to the top strand, 53 bp downstream of L1. (C) One-dimensional supercoiling assay of chromatin assembly. Topoisomers were
resolved on a 1% agarose gel with 15 l
M chloroquin present in the gel as well as the tank buffer. The arrow marks a band seen in every
lane, used as a reference. (C) Plasmid DNA control, which was not subjected to chromatin assembly. (D) Profiles of the topoisomer distribu-
tion of chromatin from lanes 2, 6, 10 and 12 in (C). The gray vertical line shows the alignment of peaks corresponding to the band marked
with an arrow in (C), and the dot marks the peak with the highest intensity in a profile. (E) Profiles of the topoisomer distribution in lanes 3
and 7 (chromatin with R3) are compared with those in lanes 4 and 8 (chromatin without R3) in (C).
R H. Pusarla et al. Mechanisms of nucleosome positioning
FEBS Journal 274 (2007) 2396–2410 ª 2007 The Authors Journal compilation ª 2007 FEBS 2399
(Fig. 3C), at 150 mm salt, the significantly different
superhelical density (Fig. 3C) and presence of only
sparse nucleosomes in an MNase ladder assay (not
shown) suggest that the absence of positioned nucleo-
somes in this case is due to inefficient chromatin assem-
bly, rather than the loss of the boundary effect of R3.
Spacing between the positioned nucleosomes
does not change with ionic strength
The bulk chromatin is reported to show an increase in
NRL with increasing ionic strength [18]. Although this
may influence the spacing even between the positioned
nucleosomes, the IEL analysis in Fig. 2 suggested that
the relative location of the positioned nucleosomes
remains the same at every ionic strength. Numbering
the nucleosome between the operators 183 bp apart
as 0 (Fig. 2A), we used MNase footprinting to map
the positions of nucleosomes 0, ) 1 and + 1 by using
different primers to look at one nucleosome at a time

(Fig. 4). Translationally positioned nucleosomes
showed a clear 145 bp protection in the profile com-
parisons of the chromatin lanes with and without R3.
The + 1 nucleosome was found with its 5¢ edge
located 10 bp downstream of the lac operator L2
(Fig. 4D), whereas the 3¢ edge of nucleosome ) 1 was
found 10 bp upstream of the lac operator L1 (Fig. 4B)
at every ionic strength tested. Mapping of nucleosome
0 between the lac operators L1 and L2 showed its
location to be 25 bp downstream of L1 on the 5¢ side,
and 12 bp upstream of L2 on the 3¢ side. Similarly, the
location of the + 2 and ) 2 nucleosomes did not
change with changing ionic strength. This analysis
shows that the number of positioned nucleosomes
AB C
D
Fig. 4. Structural analysis of the pU6lac3 chromatin. High-resolution MNase footprinting was used to map the positioned nucleosomes over
plasmid pU6lac3 without or with bound R3. A seven-fold molar excess of R3 over operators was added at the start of the assembly. Aliqu-
ots of the same assembly were subjected to three levels of MNase digestion. Ellipses mark the nucleosomal size protections, and solid
boxes represent the R3 footprint and the lac operators. The positions of nucleosomes ) 1, 0, + 1 and + 2 are marked. (A) Mapping the posi-
tion of the nucleosomes in the presence of R3 by the primer extension footprinting of the chromatin assembled at 70 m
M salt. The primer
was located 196 bp upstream of the lac operator L1, hybridizing to the bottom strand, as depicted in the cartoon in the left-hand bottom cor-
ner. Lanes 1–8 show extension products of naked DNA digestions, and lanes 9–16 show chromatin samples. R3 was added to the samples
in lanes 5–8 and 13–16. GATC shows the sequencing ladder generated with the same primer. (B) Primer extension footprinting of the chro-
matin reconstituted at 90 m
M salt. A comparison of the profiles of the digested chromatin without and with R3 is shown. The primer was
the same as in (A). (C) Higher-resolution MNase footprinting of the chromatin with R3 bound or unbound shows that the lac operator L3 gets
included in the positioned nucleosome + 1. Lanes 1–8 show the naked DNA pattern with (lanes 5–8) and without (lanes 1–4) R3, and
lanes 9–16 show the digestion pattern of the chromatin assembled at 70 m

M salt (R3 was added to lanes 13–16). The primer was located
64 bp upstream of lac operator L2, hybridizing to the bottom strand as shown schematically in the right-hand bottom corner. GATC shows
the sequencing ladder generated by using the same primer. (D) A comparison of the profiles of lanes 9 and 13 in (C) is shown. Protection of
145 bp due to a positioned nucleosome + 1 overlaps with part of L3 occupied by R3.
Mechanisms of nucleosome positioning R H. Pusarla et al.
2400 FEBS Journal 274 (2007) 2396–2410 ª 2007 The Authors Journal compilation ª 2007 FEBS
changes with ionic strength, but the spacings of these
nucleosomes do not change with ionic strength. It is
important to note that the reported increase in the
measured NRL with increasing ionic strengths [18]
represents the change in the average NRL of the bulk
chromatin, and we did not see any disruption of the
regular spacing in the presence of R3 (Fig. 3B). There-
fore, the small changes in the average nucleosome spa-
cing at lower ionic strengths ([18], data not shown) can
easily adjust with unaltered spacing of a few posi-
tioned nucleosomes (Fig. 4) in a relatively shorter
region of the chromatin. At higher salt concentrations
(> 110 mm), the extreme nucleosomes ) 2 and + 2,
which are positioned due to alignment rather than
restricted mobility, moved out of register and behaved
like bulk nucleosomes, resulting in loss of positioning.
As the change in ionic strength did not disturb the
positioned nucleosomes at lower ionic strengths, we
conclude that nucleosome positioning dominates the
nucleosome-spacing effects on the bulk chromatin.
The distance between protein-binding sites
decides nucleosome positioning
Positioning of five nucleosomes in a single register as a
result of the binding of R3 to two lac operators in plas-

mid pU6LNS [26] suggested that the bound R3 mole-
cules serve as bookmarks for alignment of nucleosomes
on both sides. However, as compared to pU6LNS with
two lac operators, the presence of a third operator L3
did not increase the number of positioned nucleosomes in
pU6lac3. Whereas the 183 bp distance between the two
operators L1 and L2 in pU6LNS and pU6lac3 can easily
accommodate one nucleosome, the subnucleosomal dis-
tance of 140 bp between L2 and L3 in pU6lac3 did not
disrupt positioning of the nucleosomes downstream of
L2 (Fig. 2). Footprinting to map the two downstream
nucleosomes, + 1 and + 2 in Fig. 4C, showed protec-
tion at L3 contiguous with the remaining 130 bp
between L2 and L3 (nucleosome + 1 is positioned 10 bp
downstream of L2), suggesting that L3 becomes inclu-
ded in the core particle + 1 at its 3¢ edge (Fig. 4C,D).
Although there is a possibility that R3 at L3 was
removed by nucleosome + 1, a total protection size of
more than 145 bp DNA downstream of L2 and span-
ning over L3 suggests that it was generated by the co-
occupancy of R3 and nucleosome at L3. Indeed, the R3
footprint at both L3 and L2 at higher digestion levels
(Fig. 4C, lane 13 versus lane 9) was clearly visible. This
means that R3 is included in the positioned nucleosome
rather than working as a protein block at L3, as dis-
cussed later. The lac repressor, a major groove-binding
small protein, is reported to form tripartite complexes
with a nucleosome [31,32], and once formed, the repres-
sor–DNA complex is stable [33]. The properties of bind-
ing of R3 to DNA are similar to those of the wild-type

protein [34]. Therefore, it appears that the observed pro-
tection is due to a positioned nucleosome encroaching
on half of the R3-bound L3. These results explain why
the presence of the third operator L3 did not increase
the range of nucleosome positioning in pU6lac3.
At every ionic strength, nucleosome 0 was positioned
25 bp from L1 and 12 bp from L2, whereas the flanking
nucleosomes ) 1 and + 1 were positioned 10 bp from
L1 and L2, respectively. This observation suggests that
a minimum distance of  10 bp is maintained between
the operators and the histone octamer position, giving a
minimum requirement of 165 bp between the two pro-
teins if they are to work as flag-ends for a positioned nu-
cleosome, protecting 145 bp of core DNA length
between them. The results shown in Fig. 4 suggest that
the distance between the two binding sites should be
more than 140 bp if they are both to work as demarca-
tion boundaries. In order to find this minimum, optimal
distance, we derived a set of plasmids from pU6lac3 by
reducing the 183 bp distance between L1 and L2 in 5 bp
increments (Fig. 1B). Reducing this distance by 10 or
20 bp (as well as 5 and 15 bp, not shown) did not result
in any loss of positioning (Fig. 5A). A distance of
163 bp in plasmid d20, more than the core DNA length,
shows the presence of five positioned nucleosomes in the
presence of R3. In contrast, the deletion by 25, 30 or
35 bp resulted in the loss of the array of positioned nu-
cleosomes, and only one positioned nucleosome show-
ing less pronounced protection between L1 and L2
could be seen (lanes 3, 6 and 8) in the IEL analysis in

Fig. 5B. However, a comparison of the profiles in
Fig. 5C revealed the presence of three nucleosomal size
protections in d30, whereas no protection could be seen
in the profile comparisons for d35 (Fig. 5C). A strong
nuclease sensitivity (marked by asterisks) next to the R3
bound to all three operators, suggesting an alteration in
the chromatin structure, could also be seen.
The higher-resolution MNase footprinting used to
ascertain protection confirmed the presence of the
145 bp protection (nucleosome 0) between L1 and L2
(Fig. 5D,E) in plasmids d25–d35. A 10 bp distance
was maintained from L2 on its 3¢ edge, but the dis-
tance from L1 was reduced to less than 9 bp on the 5¢
side in all three plasmids (Fig. 5D), suggesting that L2
works as the central reference point for nucleosome
deposition, leaving a minimum gap of  9 bp from L2
on both of its sides. A distance of 158 bp between L1
and L2 in d25 allows L1 to remain excluded from
145 bp protection (middle profile, Fig. 5E), whereas
the distance of 153 bp in d30 makes the 5¢ boundary
R H. Pusarla et al. Mechanisms of nucleosome positioning
FEBS Journal 274 (2007) 2396–2410 ª 2007 The Authors Journal compilation ª 2007 FEBS 2401
A
DE
F
BC
Fig. 5. Position of the nucleosome between two protein-binding sites changes with the distance between them. Chromatin was assem-
bled using an S-190 extract of Drosophila embryos in the presence or absence of the R3 protein. Ellipses indicate the nucleosomes, and
solid boxes represent the R3 footprint over the lac operator. Arrowheads and arrows mark the positions of the lac operators. The names
of the plasmid DNAs are given for each panel. (A) IEL analysis of the chromatin assembled over plasmids d10 (lanes 1–7) and d20 (lanes 8

and 9). M, molecular size marker, and positioned nucleosomes are numbered. The probe was the same as that in Fig. 2. (B) IEL analysis
of the chromatin assembled over plasmids d25, d30 and d35. M, molecular size marker. Lanes 1–3 represent d25, lanes 4–6 represent
d30, and lanes 7–9 represent chromatin and naked DNA samples of d35. The probe was the same as that in Fig. 2. The black dot marks
the spot in lane 9, which is an artefact, and asterisks denote the positions of the hypersensitivity generated due to R3 binding in lanes 3,
6 and 8. (C) Comparisons of the chromatin from (B), showing profiles of lanes 1 and 3 for d30 (upper panel) and lanes 7 and 8 for d35
(lower panel). The asterisk denotes the position of the hypersensitivity generated due to R3 binding. (D) Higher-resolution MNase footprint-
ing of the chromatin assembled over d30 and d35. The 5¢ boundary of the nucleosome between L1 and L2 is close to the 3¢ edge of L1
in d30 and encroaches on it in d35. The primer hybridizes to the top strand 53 bp downstream of L2. GATC shows the sequencing ladder
generated by using the same primer. (E) Comparison of digestion profiles of the chromatin without and with R3 for d10, d25 and d35 from
high-resolution footprinting gels identifies the 145 bp protection due to the positioned nucleosome 0 between L1 and L2. The comparison
for d35 [lanes 4 and 6 from (D)] in the lower panel shows that the protection between L1 and L2 is less than 145 bp. The black dot near
L1 in all the profiles marks the small peak in the presence of R3 (d10 and d25), which is absent in the lower panel (d35). (F) Higher-resolu-
tion MNase footprinting of the chromatin assembled over d35 shows the increased nuclease sensitivity of the top strand of the DNA in
the presence of R3. The primer hybridizes to the bottom strand 93 bp upstream of L1. GATC shows the sequencing ladder generated by
using the same primer.
Mechanisms of nucleosome positioning R H. Pusarla et al.
2402 FEBS Journal 274 (2007) 2396–2410 ª 2007 The Authors Journal compilation ª 2007 FEBS
of nucleosome 0 touch operator L1 (Fig. 5D). A fur-
ther 5 bp deletion in d35 included L1 in the core parti-
cle (Fig. 5D,E). In the presence of R3, the large peak
marked with an asterisk (Fig. 5E) marks the 5¢ bound-
ary of nucleosome 0 in d10, but nucleosomal protec-
tion was seen over it in d25 and d35. Similarly, the
small peak in the presence of R3, near L1 in the pro-
files (Fig. 5E, black dot), coincided with the 3¢ demar-
cation boundary of the protection due to R3 at L1. As
compared to d10, the absence of this peak at the same
position in d35 (Fig. 5E, lowest panel) showed a
145 bp protection overlapping L1, suggesting that L1
was included in the nucleosome at the 5¢ boundary. In

contrast to this, the MNase footprinting on the com-
plementary strand to probe nucleosome 0 in d35 from
the 5¢ side did not show a nucleosomal size footprint,
but R3 protection at both L1 and L2 was seen, along
with slightly higher nuclease sensitivity of the DNA
between L1 and L2 (Fig. 5F). As the operator is
accommodated by the nucleosome such that the
repressor-binding surface is always facing away from
the octamer [31,32], and binding of repressor to the
operator bends it away from the repressor [35], it is
possible that the phase difference of 5 bp from d30 dis-
turbs the protection of the ends of the core DNA in
d25 and d35. Thus, despite the clear R3 footprint on
L1 and L2 in d35 (Fig. 5D,F), and the probable pres-
ence of the positioned nucleosome, the higher MNase
sensitivity of the only top strand between L1 and L2
and deformation of DNA at the core ends may result
in the less clear protection (Fig. 5B,C, lane 8, lower
panel) in the IEL analysis, which revealed the double-
strand cuts by MNase. Therefore, as R3 was not
removed in d35 (Fig. 5F), we conclude that R3
and histone octamer co-occupy the DNA between L1
and L2.
The IEL profile of d30 showed three positioned
nucleosomes, whereas only one positioned nucleosome
was confirmed by the footprint profiles on the bottom
strand in d25 and d35 (Fig. 5D,E). Thus, R3 can work
A
CE
BD

Fig. 6. The number of positioned nucleosomes decreases with the number of R3-binding sites. Chromatin was assembled over plasmids
with two or one R3-binding sites, in the presence or absence of the R3 protein. Ellipses mark the nucleosomes, and the solid box repre-
sents the R3 footprint over the lac operator. Arrowheads and arrows mark the positions of the lac operators. The asterisk denotes the posi-
tion of the hypersensitivity generated due to R3 binding. (A) IEL analysis of plasmid pU6lac2 (with two lac operator sites) as chromatin and
naked DNA, with and without R3. Chromatin was assembled using the S-190 extract of Drosophila embryos. The probe position was the
same as that in Fig. 2. M, molecular size marker. (B) IEL analysis of plasmid pBKS+ assembled into chromatin by using the S-190 extract of
Drosophila embryos in the absence and the presence of R3. The probe position was the same as that in Fig. 2. M, molecular size marker.
(C) Positioned nucleosome upstream of L2 in the presence of R3 on the pU6lac2 chromatin assembled by using the S-190 extract of Dro-
sophila embryos. A comparison of the profiles of MNase-digested chromatin without and with R3 from a higher-resolution footprinting gel is
shown. A radiolabeled primer used for the extension reaction hybridized 53 bp downstream of L2. (D) A low-resolution structure analysis of
the chromatin over plasmid pU6lac2 reconstituted by the salt gradient dilution method of nucleosome assembly. The probe position was the
same as that in Fig. 2. M, molecular size marker. (E) A profile comparison of the low-resolution IEL analysis of the plasmid pBKS+ chromatin
reconstituted by the salt dilution method, with and without R3, shows the single nucleosome.
R H. Pusarla et al. Mechanisms of nucleosome positioning
FEBS Journal 274 (2007) 2396–2410 ª 2007 The Authors Journal compilation ª 2007 FEBS 2403
as a true boundary when the distance between two R3-
binding sites is 165 bp at a minimum, as in d20. The
nucleosomes flanking the bound protein should have
flexibility of leaving minimum 8–10 bp from the pro-
tein-binding site when they occupy their positions.
Positioning is adjusted according to the distance avail-
able between two blocks, but a constraint on the flexi-
bility due to a shorter distance than this minimum
leads to loss of positioning.
Number of protein-binding sites decides the
number of positioned nucleosomes
The presence of two suboptimally placed pairs of
R3-binding sites in d35, L1 at 148 bp and L3 at 140 bp
from L2, resulted in disruption of the alignment of
nucleosomes on both sides of L2. These results suggest

that in the absence of L1, the remaining two operators,
L2 and L3, separated by less than the nucleosome core
length, should behave like a single operator. Therefore,
we removed the lac operator L1 from pU6lac3 to
obtain plasmid pU6lac2 (Fig. 1A). Chromatin was
assembled on the plasmids with R3 bound at a single
position (pBKS+) or at two sites 140 bp apart
(pU6lac2), and IEL was used to find the presence of
positioned nucleosomes on both plasmids (Fig. 6).
Four positioned nucleosomes on pU6lac2 (Fig. 6A,
lane 4) only in the presence of R3, two downstream of
L2, probably with inclusion of L3 in the nucleosome,
and two upstream of L2, confirmed by the MNase
footprinting analysis (Fig. 6C, gel profile comparisons),
were seen. In comparison, in pBKS+, only two
weakly positioned nucleosomes could be seen flanking
L3 (see Fig. 6B, lanes 3 and 4). Interestingly, strong
nuclease sensitivity was seen downstream of L3 in both
plasmids (marked by an asterisk), which may be due
to the exclusion of more nucleosomes downstream of
L3, as discussed later. Therefore, the results confirm
that the number of positioned nucleosomes reduces
with the number of protein blocks.
Binding of R3 leads to positioned nucleosomes
only on one side
All chromatin assemblies in the experiments mentioned
above were assisted by the various S-190 extract com-
ponents. Nevertheless, the presence of positioned
nucleosomes in the absence of uniform spacing and the
ability of R3 to bind to its site even at high ionic

strengths (Figs 2C and 3A) suggested that nucleosomes
can be positioned in the presence of R3 even by the
salt gradient method of chromatin assembly. Chroma-
tin assemblies formed by the salt dilution method on
the three plasmids with different numbers of operators
showed different numbers of translationally positioned
nucleosomes in the presence of R3. Whereas no posi-
tioned nucleosomes were seen downstream of L3, only
one positioned nucleosome was found on the 5¢ sides
of L2 in pU6lac2 (Fig. 6D) and L3 in pBKS+
(Fig. 6E) in the presence of R3. Although R3 bound
to both the operators in pU6lac2, the presence of pro-
tection between L2 and L3 could not be confirmed,
either by the IEL profile comparisons or MNase foot-
printing. Moreover, a clear 5¢ boundary of the nucleo-
some upstream of L2 was not visible in Fig. 6D,
probably due to its unrestricted mobility on the 5¢ side.
Nevertheless, the double-stranded nature of the MNase
cuts of DNA in IEL maps, and the size of this protec-
tion ( 220 bp of DNA), suggest that the protection is
due to weak translational positioning or unique rota-
tional phasing of multiple nucleosome positions there.
Thus, these results show that at least one nucleosome
gets positioned on one side of a DNA-bound R3 in
the salt dilution method of chromatin assembly. The 3¢
side of the DNA occupied by R3 in the chromatin
shows hypersensitivity to MNase, which gives an
important structural insight into the nucleosome posi-
tioning seen over these plasmids, as discussed in the
next section.

Multiple translationally positioned nucleosomes
due to R3 binding in the salt dilution method of
chromatin assembly
With the two methods of assembly used in Fig. 6, R3
binding did not give the same number of positioned
nucleosomes on pU6lac2 and pBKS+. However, on
pU6lac3, similar to S-190-assembled chromatin, five
nucleosomes were positioned in the presence of R3
(Fig. 7A; compare lanes 7–9 with lanes 5–6) in the salt
gradient method of chromatin assembly. Positioning of
two nucleosomes flanking both sides of L1 as well as
L2 was confirmed by MNase footprinting analyses,
which showed well-positioned nucleosomes on both
sides of the middle operator, L2, whereas L3 was
included at the 3¢ edge of the + 1 nucleosome
(Fig. 7B,C). In the absence of R3, the sequence
upstream of L2 in both pU6lac3 and pU6lac2 was
found to give rotational but not translational position-
ing of a nucleosome, whereas the sequence upstream
of L3 in pBKS+ did not show even rotational posi-
tioning (data not shown). However, the translationally
positioned nucleosomes in Figs 6 and 7 are sequence-
independent, as they could not be seen in the absence
of R3 and did not show any rotational positioning in
DNaseI footprinting.
Mechanisms of nucleosome positioning R H. Pusarla et al.
2404 FEBS Journal 274 (2007) 2396–2410 ª 2007 The Authors Journal compilation ª 2007 FEBS
Deposition of five positioned nucleosomes on
pU6lac3 in the presence of R3 can serve as a potential
method to obtain the array of translationally posi-

tioned nucleosomes by the salt gradient dilution
method of chromatin assembly. However, the presence
of R3 in the array may interfere with the use of this
chromatin for further analyses and restrict the utility
of the method. Therefore, we used isopropyl thio-b-d-
galactoside (IPTG) to dislodge R3 from the assembled
chromatin to obtain a chromatin with positioned and
spaced nucleosomes by this method. IPTG is reported
to quickly shatter the DNA–repressor complex. The
30 min incubation of chromatin with IPTG gives the
nucleosomes ample scope for readjustment. However,
the IEL analysis in Fig. 7D showed that addition of
IPTG to the chromatin assembled in the presence of
R3 did not disrupt the array (compare lanes 9 and 10
with lanes 7 and 8). Addition of IPTG did not inter-
fere with MNase activity (Fig. 7D; compare lanes 3–6
and 11 with lanes 1–2). Further analysis of the struc-
ture by MNase footprinting (Fig. 7E) showed that R3
was removed by IPTG but the nucleosomal protection
between L1 and L2 was not lost (compare lane 9 with
lanes 7–8). Comparison of profiles of digestion of
chromatin with R3 to that with R3 and IPTG
(Fig. 7F) showed that the profiles matched and com-
pletely overlapped each other, except at the operators,
where the footprint of R3 was lost in the presence of
IPTG. The analysis confirmed that the nucleosomes
positioned due to R3 binding at the operators stayed
in place even after removal of R3.
We have previously used R3 to obtain templates
with translationally positioned nucleosomes in S-190-

assisted chromatin assembly. The presence of R3 on
the template does not interfere with chromatin remode-
ling or transcription [36]. Unlike other published proto-
cols, which depend on the presence of positioning
sequences of substantial length, the proposed method
would require the introduction of only 20–30 bp of
DNA containing a lac operator at the appropriate posi-
tion. Thus, R3 can be used to generate nucleosomal
arrays by the salt dilution method on any DNA.
A
CF
BD E
Fig. 7. Nucleosomes are positioned in the presence of R3 on the chromatin assembled by the salt gradient dilution method. Chromatin
assembled as described in Experimental procedures was subjected to MNase digestion. Nucleosomes are marked by ellipses, and the posi-
tions of lac operators are indicated. (A) Low-resolution IEL analysis of the pU6lac3 chromatin assembled by the salt gradient method. The
probe was the same as that in Fig. 2. (B) Mapping of the positioned nucleosomes in the presence of R3 on the pU6lac3 chromatin. High-
resolution gel analysis of the extension products of the MNase-digested chromatin without and with R3, using a primer that hybridized
53 bp downstream of L2, is shown. Gray boxes denote the R3 footprint, and ellipses mark the nucleosomal protection. GATC shows the
sequencing ladders generated with the same primer. (C) Comparison of the profiles of the MNase-digested chromatin without and with R3
from a high-resolution footprinting analysis of the extension products of a primer that hybridized 64 bp upstream of L2 shows that L3 gets
included in a positioned nucleosome. (D) Effect of IPTG on the chromatin assembled in the presence of R3. Excess IPTG was added at the
end of the chromatin assembly by the salt gradient dilution method, and incubation was continued for 30 min at 30 °C. IEL analysis of the
chromatin structure is shown. The probe was the same as that in Fig. 2. M, molecular size markers. (E) Addition of IPTG dislodges R3 from
its sites. High-resolution MNase footprinting shows loss of protection at the operators while nucleosomal protection is maintained in the
presence of IPTG (lanes 7–9). The primer was same as that in (B). GATC shows sequencing ladders generated with the same primer. (F)
Comparison of the profiles of the MNase-digested chromatin with R3 from the gel in (E) shows that profiles in the presence or absence of
IPTG perfectly match everywhere except at the lac operators L1 and L2.
R H. Pusarla et al. Mechanisms of nucleosome positioning
FEBS Journal 274 (2007) 2396–2410 ª 2007 The Authors Journal compilation ª 2007 FEBS 2405
Taken together, the results of this study have shown

that the positioning of nucleosomes in the vicinity
of a sequence-specific DNA-bound trans-acting protein
does not require a context of uniformly spaced
nucleosomes. With increasing ionic strength, fewer
nucleosomes get aligned in the flanking regions. A
DNA-bound protein does not always work as a
demarcating flag for the positioned nucleosomes. The
number of positioned nucleosomes in a single register
depends on the ionic strength, the number of binding
sites, and the distance between the binding sites.
Discussion
Nucleosome spacing and positioning
with changing ionic strength
Nucleosomes are nucleoprotein complexes formed due
to the neutralization of the acid ⁄ base nature of their
components and held together by attractive forces of
opposite charges. Like those of other such complexes,
the structure and stability of nucleosomes are greatly
influenced by the ionic strength of the medium in vitro
[16,17]. Folding of the chromatin into higher-order
structures involves interparticle interactions of the
nucleosome cores. Several charge-neutralizing mecha-
nisms operating on histones or the DNA backbone
may modulate these interactions similarly in vivo.
Some of these modifying activities are covalent modifi-
cations of the charged histone tails, DNA methylation,
incorporation of histone variants, and DNA-binding
basic proteins. Thus, the ionic strength of the medium
in the in vitro experiments represents different charged
states of genomic DNA in vivo. The previously repor-

ted change in the repeat length of regularly spaced
nucleosomes and different superhelical densities
(Fig. 3C) in response to the changing ionic strength
thus explain how the packing densities of nucleosomal
arrays may change in response to any of the above-
mentioned activities.
It is known that the ionic strength of the medium
affects the sliding of the nucleosomes. Persistence of
nucleosome 0 between L1 and L2 in the presence of
R3 at higher ionic strengths (Fig. 2) suggests that it is
probably more stable than the flanking nucleosomes.
The sequence between L1 and L2 in pU6lac3 was
found to give rotationally positioned nucleosomes in
the absence of R3 only in the salt gradient assembly.
However, chromatin assembly in the S-190 extract is
ATP-dependent and sequence-independent, which
ensures uniform occupancy of the full DNA length by
the nucleosomes with equal spacings. Therefore, the
apparent stability of nucleosome 0 at higher ionic
strengths is conferred on it by the boundary effect of
the R3 pair bound to L1 and L2.
Turning a gene into an active state from a state of
inactivation often involves binding of a single activator
molecule to its site in a repressive chromatin structure
[37]. Thus, the binding of proteins at singular sites in
an array of densely packed and equally spaced nucleo-
somes can cause substantial decondensation and rear-
rangements in condensed chromatin regions, working
as trigger for the opening of regions. Active genes with
nucleosome-free regions flanked by densely packed

nucleosomal regions or widely spaced but positioned
nucleosomes on gene regions have been observed in vivo
[38–40]. In agreement with this, the results in Figs 2
and 7 show that nucleosomes can be positioned in the
absence of an array of equally spaced nucleosomes.
Mechanism of nucleosome positioning by R3
Depending upon their location, nuclease-hypersensitive
regions and DNA distortions are reported to be
important factors for nucleosome positioning [41,42].
Binding of the lac repressor to DNA has been studied
in great detail. Earlier studies based on gel mobility
shift assays have reported that the wild-type operator
does not show a detectable bend on wild-type repres-
sor binding [43]. However, the crystal structure of the
repressor–operator complex has shown that the bind-
ing of the repressor to the 21 bp symmetric operator
distorts the DNA conformation, bending it away from
the repressor [35]. All three operators used in this
study constitute the high-affinity, wild-type O1 oper-
ator sequence of the Escherichia coli lac operon. An
unidentified conformational change was reported to
accompany the binding of the lac repressor to nucleo-
somal DNA [32]. Similar to this, we have observed an
asymmetric DNA deformation when R3 binds to the
lac operator only on chromatin. The location of the
operator site in nucleosomal DNA was also reported
to affect the binding affinity of the lac repressor [32],
which follows an asymmetric mode of logging onto the
operator [35]. The MNase hypersensitivity in the pres-
ence of R3, only on one side (3¢ side) of L2 and L3 in

the chromatin (Figs 5 and 6), suggests a distortion of
the DNA there, probably because of the asymmetric
nature of the operators [35]. R3 is a mutant of the lac
repressor that binds to the operator only as a dimer.
Therefore, its binding to the asymmetric operator may
result in only asymmetric deformation of the operator
without subsequent looping of the intervening DNA
that is seen with the wild-type protein. This asymmetry
in the location of the DNA deformation is found to
exclude the positioned nucleosomes on the 3¢ side of
Mechanisms of nucleosome positioning R H. Pusarla et al.
2406 FEBS Journal 274 (2007) 2396–2410 ª 2007 The Authors Journal compilation ª 2007 FEBS
both L2 and L3 (Fig. 6). However, chromatin assem-
bly and spacing of nucleosomes in S-190 extract are
energy-driven processes, which even out the effects of
the DNA deformations by including them in nucleo-
somes rather than excluding them. This may be the
reason why different numbers of nucleosomes were
positioned due to R3 in both methods of assembly.
L1 in pU6lac3 is present in an opposite orientation
from that of the other two operators, suggesting that a
similar deformation can be expected on its 5¢ side. In
contrast to the L2 and L3 pair, the distortions on the
opposing and flanking sides of the L1 and L2 pair may
result in facilitated nucleosome deposition between
them. Owing to the 183 bp distance between L1 and L2,
the R3s bound to these sites can work as boundaries,
restricting the translational mobility of nucleosomes
between them. Fixing the position of nucleosome 0 in
this way can therefore lead to alignment of more posi-

tioned nucleosomes on both of its sides. With L3 at a
distance of 140 bp on the 3¢ side of L2, the nucleosome
formation in this region would be possible only by inclu-
ding and accommodating L3 as well as R3 bound to it
in the 3¢ edge of a positioned nucleosome.
The presence of lac operators in low-energy regions
such as the ends of the nucleosomes as compared to
the high-energy dyad axis region for conformational
changes was reported to facilitate the formation of the
ternary complex between the nucleosome and the lac
repressor [31]. The lac operator always faces the his-
tone octamer in a manner that allows repressor bind-
ing on the outward surface [31]. We have previously
shown that binding of a transcription factor next to
the R3-positioned nucleosome leads to its upward shift
and chromatin remodeling, resulting in inclusion of R3
bound to its site in a positioned nucleosome [36],
rather than R3 being dislodged. Thus, it is most likely
that R3 is not removed from L1 in d35 and L3 in
pU6lac3 or pU6lac2, which get included in the nucleo-
somal protection at the 3¢ edge in both the salt gradi-
ent and S-190 extract-assisted chromatin assemblies.
R3 as a tool for positioning nucleosomes in vitro
Assembly extracts, capable of depositing nucleosomes
over the DNA with uniform spacings, represent an
important breakthrough in the investigation of the
assembly and disassembly of chromatin in the context
of gene expression. We have previously reported the
use of R3 to prepare templates with positioned nucle-
sosomes using the S-190 extract method of assembly

[36]. Significantly, the protein binding at specific sites
led to positioned nucleosomes even in the salt dilution
method of chromatin assembly, presenting a poten-
tially new method of preparing templates with posi-
tioned nucleosomes in vitro. However, less efficient
positioning on pU6lac2 as well as pBKS+ as com-
pared to pU6lac3 suggests that the presence of two
operators at more than the core nucleosome distance
in pU6lac3 may be the reason for the difference. When
R3 is bound to both of the sites separated by more
than 145 bp, the mobility of the nucleosome between
them is restricted, generating strong translational posi-
tioning (e.g. in pU6lac3), which facilitates the align-
ment of flanking nucleosomes also.
Operator spacing also shows some differences
between the methods of assembly. In d35, where both
L1 and L3 come closer to L2, arrays are lost in the
S-190-assisted assembly, whereas in pU6lac2, where L2
and L3 are at a subnucleosomal distance, arrays are
lost in the salt gradient assembly. Although we have
changed spacing only with reference to core DNA
length, by extrapolation of the results in Fig. 5, any
distance less than 130–140 bp may exclude the nucleo-
some from the intervening DNA, as it may be difficult
to accommodate a protein within the core as compared
to at the ends. Similarly, two adjacent sites would
work as single block, and nucleosomes can be posi-
tioned on both of its sides. When R3 is bound to a sin-
gle operator, one nucleosome can align on one or both
of its sides, according to the method of assembly.

Nucleosome mobility is limited only in one direction
and its position is less restricted; because of this, the
alignment of flanking nucleosomes shows greater
flexibility, and the properties of the underlying DNA
sequence can influence the positioning. These results
suggest that, depending on the requirement, variation
in number and distance of R3-binding sites can be
used in vitro to obtain chromatin with nucleosome-free
regions or with positioned nucleosomes.
The presence of transcription factor-binding sites is
reported to regulate the presence of nucleosomes on a
genomic region [4]. We have shown in this study that
trans-acting proteins are the dominant determinants
for nucleosome positioning and spacing. Several exam-
ples of nucleosome positioning due to the binding of a
transcription factor to the target site and subsequent
remodeling are known. Positioned nucleosomes bring
two distant sites closer and facilitate the interaction
between the factors bound to them [8,9,44,45]. Posi-
tioned nuclesomes are often found at the borders of
hypersensitive sites in the regulatory regions of genes
[6], which may have multiple factor-binding sites. On
the basis of our results, we propose that any protein
showing DNA-interaction properties, such as R3,
can show similar effects on nucleosome positioning.
It is not the size of a factor that makes it work as a
R H. Pusarla et al. Mechanisms of nucleosome positioning
FEBS Journal 274 (2007) 2396–2410 ª 2007 The Authors Journal compilation ª 2007 FEBS 2407
boundary marker for disruption or alignment of nucleo-
somes next to the protein. It is, rather, the distance

between the binding sites of two factors that will decide
either the positioning or exclusion of nucleosomes
between them. Thus, the results can be used to predict
whether a positioned nucleosome would be involved in
juxtaposing two known distant sites for the productive
interaction of the protein factors binding to them.
Experimental procedures
DNA
Three plasmid DNA constructs with different numbers of
lac operator sequence copies were used in this study
(Fig. 1A). Plasmid pU6lac3 containing three copies of high-
affinity lac operator sequences of type O1 was similar to
plasmid pU6LNS [29] with two lac operators at a distance
of 183 bp, reported previously. Plasmid pU6lac2 was
derived from pU6lac3 such that L1 was deleted to give two
lac operator sites, and the vector DNA Bluescript, pBKS+
(Stratagene, La Jolla, CA, USA) was used in place of the
plasmid with a single lac operator site. Figure 1A shows
the relative positions of the three sites in these plasmids.
The distance between L1 and L2 is 183 bp, and L3 is at a
distance of 140 bp from L2. The plasmids had unique sites
for the restriction enzymes: AlwN1 at 709 bp from the third
lac operator, L3; and Xmn1 1117 bp upstream from the
first lac operator, L1. Plasmids d5, d10, d15, d20, d25, d30
and d35 were derivatives of pU6lac3 wherein the distance
between L1 and L2 was reduced by deleting the intervening
DNA in 5 bp increments such that the distance between L1
and L2 was reduced to 148 bp in d35 (Fig. 1B), 153 bp in
d30, 158 bp in d25, and so on.
Preparation of R3 protein

A mutant lac repressor protein, R3, which can dimerize but
not tetramerize [35], was prepared from an overexpression
clone (gift from K. Matthews, Rice University, Houston,
TX, USA) according to the method described therein. The
activity of the protein was estimated by a gel shift assay,
and saturating amounts of protein (typically, 12.2 pmol of
R3 dimers to 550 fmol of plasmid DNA with three lac
operator sites) were used for both the IEL and footprinting
experiments.
Chromatin assembly
Chromatin with regularly spaced nucleosomes in the absence
of histone H1 was assembled at an NaCl concentration of
50 mm, using an S-190 extract of Drosophila embryos and
exogenously added rat liver core histones as described
previously [8,26]. Rat liver core histones were salt-extracted
octamers, prepared by using hydroxylapatite [17]. Taking
the ionic strength of the S-190 extract as equivalent to
100 mm NaCl, the ionic strength of the assembly mix was
increased by addition of different amounts of NaCl to give
the total NaCl concentration as stated. Saturating amounts
of R3 protein (7–8-fold molar excess of dimers over the
operators) or its storage buffer were incubated with DNA in
the presence of appropriate buffer for 30 min before the
start of assembly. Aliquots of the same assembly mix were
used for structural analyses using the DNA-cleavage enzymes
MNase or DNaseI as described previously [8,26].
Topological analysis of chromatin assembly
Chromatin assembled with or without R3 by using the
S-190 extract of Drosophila embryos was deproteinized at
the end of assembly, and 250 ng of DNA per lane was loa-

ded in a 1% agarose gel containing 15 lm chloroquin. The
gel was run at 60 V for 12 h in 1 · TBE buffer containing
15 lm (10 lgÆmL
)1
) chloroquin. Chloroquin was removed
before staining the gel with ethidium bromide by shaking
the gel for 10 min in deionized water with five changes.
Salt gradient dilution method of chromatin
assembly
For chromatin assembly, 2 lg of DNA and rat liver core
histones (1 : 1.6 molar ratio of DNA to histones) were
incubated with 1 lg of BSA in a 10 lL volume at a final
NaCl concentration of 2 m at 37 °C for 10 min. Serial dilu-
tions were made to bring down the salt concentration
slowly to 1.5, 1.0, 0.8, 0.7, 0.6, 0.5, 0.4, 0.25 and 0.1 m
by adding a buffer containing 20 mm Tris ⁄ HCl (pH 8),
10 mm b-mercaptoethanol, 1 mm EDTA, 0.1 mm phenyl-
methanesulfonyl fluoride and 0.33% NP-40 with a 15 min
incubation at 30 °C for each dilution, followed by a 15 min
incubation at 37 °C for the last dilution. R3 protein
(12.2 pmol) was added at the 0.8 m salt step, and 120 nmol
of IPTG was added at the end of the assembly. The assem-
bly was incubated for 30 min with IPTG to dislodge R3.
Reconstituted chromatin samples were subsequently diges-
ted with MNase or DNaseI to carry out further analysis.
Chromatin structure analysis
The structure of the assembled chromatin was analyzed by
the low-resolution IEL method or footprinting analysis by
the primer extension method as described previously [8,26].
Briefly, 125 ng of both chromatin and naked DNA samples

partially digested with MNase or DNaseI were deproteinized
before secondary digestion with restriction enzymes (IEL) or
primer extension with Vent Exo- DNA polymerase (foot-
printing). Profiles of DNA digestions resolved in Southern
blots (IEL) or by electrophoresis on 6% polyacrylamide ⁄ 8 m
Mechanisms of nucleosome positioning R H. Pusarla et al.
2408 FEBS Journal 274 (2007) 2396–2410 ª 2007 The Authors Journal compilation ª 2007 FEBS
urea gels (footprinting) were generated from the Phosphori-
mager using the image gauge program of Fuji (Tokyo,
Japan). All protections were ascertained by matching the
profiles of the lanes with similarly digested DNA. A protec-
tion of 145 bp DNA seen in footprinting gels marked the
location of the positioned nucleosome, whereas a larger size
protection for the same could be seen in IEL analyses.
Acknowledgements
We thank P. Vanathi for pure rat liver core histones.
Financial support from CSIR, Government of India
is acknowledged. Rama-Haritha and Vinesh are recipi-
ents of Senior and Junior Research fellowship, respect-
ively, from CSIR.
References
1 Grunstein M (1990) Histone function in transcription.
Annu Rev Cell Biol 6 , 643–678.
2 Paranjape SM, Kamakaka RT & Kadonaga JT (1994)
Role of chromatin structure in the regulation of tran-
scription by RNA polymerase II. Annu Rev Biochem 63,
265–297.
3 Simpson RT (1986) Nucleosome positioning in vivo and
in vitro. Bioessays 4, 172–176.
4 Bernstein BE, Liu CL, Humphrey EL, Perlstein EO &

Schreiber SL (2004) Global nucleosome occupancy in
yeast. Genome Biol 5, R62.1–R62.11.
5 Simpson RT, Roth SY, Morse RH, Patterton H-G,
Cooper JP, Murphy M, Kladde MP & Shimizu M
(1993) Nucleosome positioning and transcription. Cold
Spring Harb Symp Quant Biol 58, 237–245.
6 Simpson RT (1991) Nucleosome positioning: occur-
rence, mechanisms and functional consequences. Prog
Nucleic Acid Res Mol Biol 40, 143–184.
7 Komura J-i & Ono T (2005) Disappearance of nucleo-
some positioning in mitotic chromatin in vivo. J Biol
Chem 280, 14530–14535.
8 Shivaswamy S, Kassavetis GA & Bhargava P (2004)
High-level activation of transcription of the yeast U6
snRNA gene in chromatin by the basal RNA polymer-
ase III transcription factor TFIIIC. Mol Cell Biol 24,
3596–3606.
9 Wolffe AP (1994) Nucleosome positioning and modifi-
cation: chromatin structures that potentiate transcrip-
tion. Trends Biochem Sci 19, 240–244.
10 Eisfeld K, Candau R, Truss M & Beato M (1997) Bind-
ing of NF1 to the MMTV promoter in nucleosomes:
influence of rotational phasing, translational positioning
and histone H1. Nucleic Acids Res 25, 3733–3742.
11 van Holde KE (1988) Higher-order structure. In Chroma-
tin (Rich A, ed.), pp. 289–317. Springer-Verlag, New
York, NY.
12 Fletcher TM & Hansen JC (1996) The nucleosomal
array: structure ⁄ function relationships. Crit Rev
Eukaryot Gene Expr 6, 149–188.

13 Thoma F & Zatchej M (1988) Chromatin folding modu-
lates nucleosome positioning in yeast minichromosomes.
Cell 55, 945–953.
14 De Ambrosis A, Ferrari N, Bonassi S & Vidali G
(1987) Nucleosomal repeat length in active and inactive
genes. FEBS Lett 225, 120–122.
15 Berkowitz EM & Riggs EA (1981) Characterization of
rat liver oligonucleosomes enriched in transcriptionally
active genes: evidence for altered base composition and a
shortened nucleosome repeat. Biochemistry 20, 7284–7290.
16 Dieterich AE, Axel R & Cantor CR (1979) Salt-induced
changes of nucleosome core particles. J Mol Biol 129,
587–602.
17 Bhargava P (1993) Dynamics of interaction of RNA
polymerase II with nucleosomes. I. Effects of salts. Prot
Sci 2, 2233–2245.
18 Blank TA & Becker PB (1995) Electrostatic mechanism
of nucleosome spacing. J Mol Biol 252, 305–313.
19 Blank TA & Becker PB (1996) The effect of nucleosome
phasing sequences and DNA topology on nucleosome
spacing. J Mol Biol 260, 1–8.
20 Clark DJ & Kimura M (1990) Electrostatic mechanism
of chromatin folding. J Mol Biol 211, 883–896.
21 Hansen JC, Ausio J, Stanik VK & van Holde KE
(1989) Homogeneous reconstituted oligonucleosomes,
evidence of salt-dependent folding in the absence of his-
tone H1. Biochemistry 28, 9129–9136.
22 Laskey RA & Earnshaw WC (1980) Nucleosome assem-
bly. Nature 286, 763–767.
23 Lusser A & Kadonaga JT (2004) Strategies for the

reconstitution of chromatin. Nature Methods 1, 19–26.
24 Banerjee S & Cantor CR (1990) Nucleosome assembly
of Simian virus 40 DNA in a mammalian cell extract.
Mol Cell Biol 10, 2863–2873.
25 Pusarla R-H & Bhargava P (2005) Histones in func-
tional diversification: core histone variants. FEBS J 272,
5149–5168.
26 Pazin MJ, Bhargava P, Geiduschek EP & Kadonaga JT
(1997) Nucleosome mobility and the maintenance of
nucleosome positioning. Science 276, 809–812.
27 Becker PB (2002) Nucleosome sliding: facts and fiction.
EMBO J 21, 4749–4753.
28 Kornberg RD (1981) The location of nucleosome in
chromatin is specific or statistical. Nature 292, 579–580.
29 Alberti S, Oehler S, von Wilcken-Bergmann B &
Muller-Hill B (1993) Genetic analysis of leucine heptad
repeats of Lac repressor: evidence for a 4-helical bundle.
EMBO J 12, 3227–3236.
30 Depew RE & Wang JC (1975) Conformational fluctua-
tions of DNA helix. Proc Natl Acad Sci USA 72,
4275–4279.
R H. Pusarla et al. Mechanisms of nucleosome positioning
FEBS Journal 274 (2007) 2396–2410 ª 2007 The Authors Journal compilation ª 2007 FEBS 2409
31 Chao MV, Gralla S & Martinson HG (1980) The lac
operator nucleosomes. 1. Repressor binds specifically to
the operator within the nucleosome core. Biochemistry
19, 3254–3260.
32 Chao MV, Gralla S & Martinson HG (1980) The lac
operator nucleosomes. 2. lac nucleosomes change con-
formation to strengthen binding by lac repressor. Bio-

chemistry 19, 3260–3269.
33 Riggs AD, Suzuki H & Bourgeois S (1970) Lac repres-
sor–operator interaction. I Equilibrium studies. J Mol
Biol 48, 67–83.
34 Chen J, Alberti S & Mathews KS (1994) Wild-type
operator binding and altered cooperativity for inducer
binding of lac repressor dimer mutant R3. J Biol Chem
269, 12482–12487.
35 Lewis M, Chang G, Horton NC, Kercher MA, Pace
HC, Schumacher MA, Brennan RG & Lu P (1996)
Crystal structure of the lactose operon repressor and its
complexes with DNA and inducer. Science 271, 1247–
1254.
36 Shivaswamy S & Bhargava P (2006) Positioned nucleo-
somes due to sequential remodeling of the yeast U6
snRNA chromatin are essential for its transcriptional
activation. J Biol Chem 281, 10461–10472.
37 Gross DS & Garrard WT (1987) Poising chromatin for
transcription. Trends Biochem Sci 12 , 293–296.
38 Almer A & Horz W (1986) Nuclease hypersensitive
regions with adjacent positioned nucleosomes mark the
gene boundaries of the PHO5 ⁄ PHO3 locus in yeast.
EMBO J 5, 2681–2687.
39 Perez-Ortin JE, Estruch F, Matallana E & France L
(1987) Fine analysis of the chromatin structure of the
yeast SUC2 gene and of its changes upon derepres-
sion. Comparison between the chromosomal and
plasmid-inserted genes. Nucleic Acids Res 15,
6937–6956.
40 Yuan G-C, Liu Y-J, Dion MF, Slack MD, Wu LF,

Altschuler SJ & Rando OJ (2005) Genome-scale identifi-
cation of nucleosome positions in S. cerevisiae. Science
309, 626–630.
41 Fitzgerald DJ & Anderson JN (1999) DNA distortion
factor as a factor in nucleosome positioning. J Mol Biol
293, 477–491.
42 Thoma F (1986) Protein–DNA interactions and nucle-
ase-sensitive regions determine nucleosome positions on
yeast plasmid chromatin. J Mol Biol 190, 177–190.
43 Wu H-M & Crothers DM (1984) The locus of sequence-
directed and protein-induced DNA bending. Nature
308, 509–513.
44 Thomas GH & Elgin S (1988) Protein ⁄ DNA architec-
ture of the DNase I hypersensitive region of the Dro-
sophila hsp26 promoter. EMBO J 7, 2191–2201.
45 Zhao X, Pendergrast PS & Hernandez S (2001) A posi-
tioned nucleosome on the human U6 promoter allows
recruitment of SNAP
c
by the Oct-1 POU domain. Mol
Cell 7, 539–549.
Mechanisms of nucleosome positioning R H. Pusarla et al.
2410 FEBS Journal 274 (2007) 2396–2410 ª 2007 The Authors Journal compilation ª 2007 FEBS