Volume3 no.10 October 1976 Nucleic Acids Research

Studies on the interaction of H] histone with superhelical DNA: Characterization
of the recognition and binding regions of H1 histone!

 

Dinah S.Singer and Maxine F . Singer

 

Laboratory of Biochemistry, National Cancer Institute, National Institute of
Health, Bethesda, MD 20014, USA

 

Received 14 July 1976

ABSTRACT
e very lysine rich histone, Hl, isolated from a variety of sources

interactspreferentially with superhelical DNA compared to relaxed DNA
duplexes”. The nature of this specific interaction has been investi-
gated by studying the ability of various purified fragments of H1 histone
from calf thymus to recognize and bind superhelical DNA. The data suggest
that the globular region of the H] histone molecule (amino acid residues
72-106) is involved in the recognition of superhelical DNA. Thus, the HI
histone carboxy-terminal fragment, 72-212, resembles native HI histone both
quantitatively and qualitatively in its ability to discriminate between and
bind to superhelical and relaxed DNA while the H1 histone carboxy-terminal
fragment, residues 106-212, has lost this specificity, binding superhelical
and relaxed DNA equally well. Furthermore, under conditions in which the
globular region of the intact H1] histone has been unfolded, the molecule
oses its ability to discriminate between superhelical and relaxed DNA, and
binds both forms of DNA equally.

INTRODUCTION

Of the five major histone fractions present in the chromatin of most
eukaryotic cells, four appear to be associated in a core about which the
DNA is wrapped@*. This unit of chromatin, commonly called a nucleosome
or nu body, consists of two molecules each of the arginine-rich and moder-
ately lysine-rich histones (H3, H4, H2a, and H2b) and is associated with
140-200 base pairs of DNA. The structural relationship of the very
lysine-rich histone H1 to this fundamental unit of chromatin is poorly
understood and its function completely unknown. Whereas histones H2a, H2b,
H3 and H4 have remained remarkably constant in structure throughout eukary-
ote evolution, the primary sequence of H] histone has undergone extensive
interspecies and interorgan variation as a result of both conservative and
non-conservative amino acid replacements®, Some species are known to

have as many as five H1] histone subfractions within a single organ; the
other four classes of histones have no more than two or three®, Further-

more, H] histone is modified through phosphorylation’. Distinct sites for

 

2531
© Information Retrieval Limited 1 Falconberg Court London W1V 5FG England
Nucleic Acids Research

 

phosphorylation of HI are reported to be associated with various stages in
the cell cycle and hormonal stimulation’ ~!°,

Calf thymus H1 histone, a protein of 212 amino acid residues with a
molecular weight of 21,000, is highly asymmetric with respect to both the
distribution of basic amino acid residues and its secondary structure! !»!2
The molecule has a short highly basic tail at the amino terminal end
and a globular central core region containing a relatively high proportion
of the hydrophobic amino acid residues; the carboxy terminal half of the
molecule consists largely of alanine, proline, and lysine residues !@

Recent reports from this laboratory indicated that H] histone from a
variety of sources interacts with superhelical DNA in preference to
relaxed DNA duplexes, and that the efficiency of the binding of DNA to HI
increases with the negative or positive superhelical density of the DNA
molecules! 37!5, The other four major classes of histones do not dis-
tinguish between the relaxed and superhelical forms of ona’, Bottger
and co-workers have confirmed these observations !©,

In this report we present experiments indicating that the globular
region of the calf thymus H] histone molecule is involved in the recognition
of superhelical DNA. Various polypeptide fragments of the H1 molecule
have been purified and their ability to interact with superhelical and
relaxed DNA has been studied.

MATERIALS

All ONA preparations were simian virus 40 labeled with  "C-thymidine.
Closed circular duplex superhelical DNA (ONA I) was isolated from purified
virions and then purified through two cycles of ethidium bromide-CsCl
banding!’ The DNA I sample was divided into aliquots, precipitated
with ethanol and stored at -20°C. Prior to use, a single aliquot of
precipitated ONA was centrifuged for 1 hour at 10,000 rpm, resuspended in
0.01 X SSC (1.5 mM sodium chloride, 0.15 mM sodium citrate) containing 1 mM
EDTA, and dialyzed against the same solution. By handling the DNA I in
this way there is less than 10% contamination with nicked, relaxed DNA.

Relaxed closed circular duplex DNA (DNA 1") was prepared from DNA I
by treatment with mammalian DNA-relaxing enzyme!8»19 as follows. " DNA-
relaxing enzyme was partially purified from mouse LA-9 cells according to
the procedure of Germond et at!9, by Dr. M. Bina-Stein. One py) of
enzyme extract was used to relax 3 yg of DNA I. A reaction mixture of 160
ug of DNA I in 0.05 M Tris-HCl, pH 7.8, 1 mM EDTA, 0.1 M NaCl, and 1 mg/ml
bovine serum albumin (BSA) in a total volume of 0.6 ml was incubated without

14

 

2532
Nucleic Acids Research

 

DNA-relaxing enzyme for 30 minutes at 23°C. The appropriate volume of
enzyme was then added and the reaction mixture incubated for an additional
hour at 23°C. The reaction was stopped by adding sodium dodecylsulfate to
a final concentration of 1%. The reaction mixture was extracted with an equal
volume of buffer saturated phenol. The phenol phase was re-extracted with
buffer and the two supernatant fluids combined. The DNA 1” was precipitated
with ethanol and stored at -20°C. The reaction product was characterized
by agarose gel electrophoresis“?. Under these reaction conditions all the
DNA I was converted to the relaxed closed circular form, DNA I".

The specific activity of the two forms of DNA was 11,309 cpm/yg. A
single preparation of DNA was used throughout these experiments.

Calf thymus H1 histone was obtained from Worthington Biochemical Corp.
and was further purified by gel filtration through a 1.9 cm x 40 cm column
of Biogel P-60. Column elution was with 0.05 M NaCl, 0.02 N HCI, 0.02%
sodium azide, according to Bohm et ai.2!, The flow rate of the column
was 6 ml/hr; 0.6 ml fractions were collected. Peak fractions from the
column were pooled and precipitated with 50 volumes of 0.05% HCl-acetone at
-20°C. The precipitate was washed twice with acetone, dried in a desiccator,
and resuspended in water. The concentration of H] histone was determined
from the absorbance at 230 nm (absorbance of 1 is equivalent to 4.25
mg/ml )*<, Analysis of this material by gel electrophoresis gave only the
expected doublet protein band characteristic of calf thymus H1 histone. 3
labeled Hi histone was prepared as previously described!®,
METHODS

N-Bromosuccinimide Cleavage of Calf Thymus Hl Histone. Histone H1
contains a single tyrosine at amino acid residue 7204208 which can be
cleaved by N-bromosuccinimide (nas)?! A modification of the procedure
of Bustin and cole!! was used for the present work. Calf thymus H]1
histone (0.1 mole), purified as described above, was suspended in 0.4 ml of
50% acetic acid. To this, 6 umoles of NBS (Sigma Chem. Co.) freshly dis-
solved in 0.4 ml of 50% acetic acid (2.7 mg/ml NBS) were added. The re-
action mixture was incubated for 4 hours at 22°C.

The reaction products were separated by gel filtration on Biogel P-60
as described above. The elution of polypeptides was monitored at 230 nm
and 260 nm. NBS cleavage of tyrosine results in the formation of a dienone
spirolactone which absorbs at 260 nm. Therefore, the amino-terminal frag-
ment (residues 1-72) can be distinguished from the carboxy-terminal fragment
(residues 73-106) by absorbance of the amino-terminal fragment at 260 nm.

H-

 

2533
Nucleic Acids Research

 

The Biogel P-60 column elution profile of the NBS-reacted H1 is shown in
Fig. 1. Each of the two peaks corresponding to the fragments was pooled
and repurified on another Biogel P-60 colum. Gel electrophoresis of the
separated fragments following the second column chromatography did not
reveal any cross contamination (Fig. 2). Amino acid analyses (data not
shown) of the two fragments were consistent with the analyses reported by
Bustin and cote!!,

Chymotrypsin Cleavage of Calf Thymus Hi Histone. Calf thymus H]
histone has a unique initial cleavage site for a-chymotrypsin on the carboxy-
terminal side of the single phenylalanine residue in the middle of the HI]
molecule?o*2? , The cleavage procedure used in these studies is a modifi-
cation of that described by Bradbury et al.°’. To a 10 mg/ml solution
of calf thymus H] histone (purified as described above) in 0.05 M Tris-HCl,
pH 8.0, a-chymotrypsin (Sigma) was added to a final concentration of 0.02
mg/ml. The reaction mixture was incubated for 10 min. at 26°C. Under
these conditions most of the H] was cleaved into two half-molecules. After
longer times of incubation, further digestion of the amino terminal frag-
ment was observed.

 

   

 

 

 

qT T q q | qT 1 t T qT q qT | T
1.0-F
0.8 F
8 0.6 F
gt
0.4-
40.3 !
0.2 do2 !
0.2 3
as 4o1 *
a. ia he al i
55

0 5 aC ee
FRACTION (Number)

Fig.1 Separation of NBS digestion products of H1 histone on Biogel P-60. Hl
histone was cleaved with NBS and the products resolved on a Biogel P-60 column as
described in the section on methods. The first peak contains the carboxy-terminal
fragment, residues 73-212; the second peak contains the amino-terminal fragment,
residues 1-72.

 

2534
Nucleic Acids Research

 

Fig.2 Gel electrophoresis of purified
NBS fragments. Following an initial
separation on Biogel P-60 column, each
of the two fragments was rechromato-
graphed on the same column. Pooled
fractions were subjected to gel electro-
phoresis as described in the section on
methods. The gels were intentionally
overloaded (50 pg/gel) to reveal any
minor cross contamination. a) Intact
H1; b) carboxy-terminal fragment 73-212;
c) amino terminal fragment 1-72.

 

Following exposure to a-chymotrypsin, the reaction mixture was ad-
justed to pH 2 by the addition of HC]. The digest was then layered onto a
Biogel P-60 column. The fragments were eluted as described above. The
elution profile of the digest from the column is shown in Fig. 3. The
material in the first peak is unreacted histone Hl. The material in the
second peak is the carboxy-terminal fragment (residues 107-212). The
material in the third peak is the amino-terminal fragment (residues 1-106).
Each of the fragments was further purified on Biogel P-60 columns. Gel
electrophoresis of the separated fragments did not reveal any cross con-

 

2535
Nucleic Acids Research

 

 

T
i

1.8
| 1.6
1.4
1.25
1.07
0.8
0.6
0.4
0.2

T
L

T

Aiso

T

 

 

 

pm’ o . én rN | i 1 ] 1 {

0 5 10 15 20 25 30 35 40 45 50 55 60
FRACTION (Number)

 

Fig.3 Elution profile of chymotrypsin digestion products of H1 on Biogel P-60. HI
histone was cleaved with chymotrypsin and the products resolved on a Biogel P-60
column as described in the section on methods. The first peak is unreacted H1; the
second peak contains the carboxy-terminal fragment 107-212; the third peak contains
the amino-terminal fragment 1-106.

tamination (Fig. 4). The carboxy-terminal fragment (residues 107-212)

ran as a doublet (data not shown), as previously observed by Bradbury

et 1,27 this presumably reflects the microheterogeneity known to exist in
the calf thymus H1 histone fraction@’. Amino acid analyses of the fragments
(data not shown) were consistent with those reported by Bradbury et al.?’
and verified the identification of the fragments.

Binding Assay. Interaction between DNA and H] histone or H1 fragments
was measured by determining the protein-dependent accumulation of V4.
labeled DNA on nitrocellulose filters, as described previously!?, Standard
reaction mixtures (0.1 ml) contained 50 mM Tris-HCl pH 7.8, 1 mM EDTA,

1 mg/ml BSA, 0.1 M NaCl, histone (or fragments) and '4c-1abeled DNA as

indicated in the text. The mixture was incubated at 23°C for 15 min, and
was then diluted with 2 ml of the buffer mixture used for the reaction (50
mM Tris-HCl pH 7.8, 1 mM EDTA, 0.1 M NaCl), filtered, and the filter washed
four times with the same solution prior to determining the radioactivity
bound to the filter.

The results are expressed as the percentage of input counts retained on
the filter as a function of the R value. The R value is defined as the
molar ratio of protein to DNA (DNA is expressed as molecules of DNA, not

 

2536
Nucleic Acids Research

 

 

Fig.4 Gel electrophoresis of purified
chymotrypsin fragments. Procedure was
as described for Fig.2. The gels were
intentionally overloaded (50 p1g/gel) to
reveal any minor cross contamination.

oO
ee

nucleotides). Molecular weights of 21,000 for Hl histone, 6,000 for amino-
terminal fragment (residues 1-72), 15,000 for carboxy-terminal fragment
(residues 73-212), 10,500 for both amino-terminal fragment (residues 1-106)
and carboxy-terminal fragment (residues 107-212), and 3.6 x 10° for SV40
pnae® were used to calculate the R value. In all cases, the concentration
of DNA used was approximately 1 ug/ml; the Hl histone concentration ranged
from 0.01 to 0.5 yg/ml.

In the range of R values studied here, Hl forms largely soluble com-
plexes with DNA I. However, at high R values (around 100), H1 histone forms
large aggregates with pnalé, (D. Singer, unpublished observations).

 

2537
Nucleic Acids Research

 

Gel Electrophoresis. 15% polyacrylamide gels containing 2.5 M urea,
0.9 N acetic acid were used throughout@°. 10 cm gels were run for 4 hr
at 150 volts and stained with amido black?7,
RESULTS

Interaction of H]l Histone With Relaxed Closed Circular DNA. Previous
studies from this laboratory! 2"! demonstrated that H] histone has a
specific ability to interact with closed circular duplex superhelical DNA
(DNA I), compared to either nicked circular duplex DNA (DNA II), or relaxed
closed circular duplexes (DONA 1) prepared with polynucleotide ligase. The
interaction of H] histone with DNA I", generated by DNA-relaxing enzyme, is
shown in Fig. 5. The results are expressed as the percentage of the input
DNA which is bound to filters as a function of the value, R (defined as the
molar ratio of Hl histone to DNA added to the reaction). Whereas all the
DNA I is retained on the filter at an R value of 50, DNA I" is not completely
retained until an R value of 300 is attained. This observation confirms
previous results!?, indicates a more striking difference between the
two DNA molecules than in our earlier work. This difference is presumably
due to the difference in the preparation of the DNA I substrates. In the
present study DNA I was extracted from purified SV40 virions, whereas in
earlier studies DNA I was extracted from SV40 infected cells. DNA I from

 

 

 

 

 

T T T T T T
°
100 + 7
90 - e ? a oe ~~ 4
/
sor 7 7
/

ee 7
5 a

60r 7 7
2 yx
ee
a 40+ / 4

i
30F / 7
/
204 / 7
ok /
it i L i L ! !
50 100 150 200 250 300 350
R

Fig.5 Hl histone dependent accumulation of 14Cc-DNA I and 14C-DNA I* on
nitrocellulose filters. Ris the molar ratio of HI/DNA. The concentration range
of HI used was 0.05-1.0 pg/ml; 14C-DNA was approximately 1 pg/ml. DNAI.— ;
DNA It----.

 

2538
Nucleic Acids Research

 

virions has a higher average superhelical density and a smaller distribution
of superhelical turns than does DNA I from infected cells??. Therefore,
there is a larger difference between the superhelical densities of the DNA
I and DNA I” preparations used in this study than in previous ones.
Interaction of Amino-Terminal Fragments of Hl Histone With DNA I
and DNA I". The two purified amino-terminal fragments resulting from the
NBS and chymotrypsin cleavages were examined for their ability to interact
with DNA I and DNA I” by the nitrocellulose filter technique. The extent of
binding of DNA by both amino-terminal fragments 1-72 and 1-106 is markedly
reduced compared to the binding by native Hl at comparable R values (Table
I). The amino-terminal fragment 1-72 does not complex DNA 1” at the highest
concentration tested (R = 568) and only a small amount of DNA I is com-
plexed at an R of 323. The amino-terminal fragment 1-106 complexes DNA I
significantly but only at high R values, and has only a very limited
ability to complex DNA 1",
Interaction of Carboxy Terminal Fragments of Hl Histone With DNA I and
DNA I". Unlike the amino-terminal fragments, both carboxy terminal frag-
ments complexed DNA efficiently (Fig. 6). The carboxy-terminal fragment

TABLE I
The Binding of Amino Terminal Fragments to DNA I and DNA I”

 

 

 

DNA I oNA I
Fragment
Conc. (ug/ml R Binding (%) R Binding (%)
Amino-terminal
Fragment (1-72)
0 0 0.9 0 0.4
0.05 32.5 1.9 28.6 -
0.10 64.6 1.5 56.9 0.3
0.20 129.2 2.5 ‘113.7 0.3
0.50 322.9 7.2 284.3 0.5
1.0 - - 568.5 1.0
Amino-terminal
Fragment (1-106)
0 0 0.3 0 0
0.04 14.6 2.8 10.6 0.5
0.08 29.2 6.8 21.1 2.8
0.10 36.5 10.7 26.4 0.2
0.20 73.1 21.6 52.8 4.6
0.50 182.7 24.4 131.9 4.7

 

tk is the molar ratio of fragment/DNA.

 

2539
Nucleic Acids Research

 

 

 

 

 

 

 

 

100
b c 4S
80 r- — 7
- A ff
& 60 — / — i
/
3 L f+ /
Q !
=z 40 i / Le f
o oe t
. / L |e
/ I
20 / ff
/ !
/ +} 9
06—<0==01 1 4 wt 1 j J i 1
0 50 100 0 50 100 0 50 100

Fig.6 Binding of H1 and Hl carboxy terminal fragments to DNAI and I*, The
procedure was as described in the section on methods. DNA I: —; DNA I’: ---.

a) Carboxy-terminal fragment 73-212; b)intact H1; c) carboxy-terminal fragment 107-
212.

73-212 complexed DNA I completely at an R value of 52, whereas DNA I" is
not completely complexed until an R value of 260 (Fig. 6a). The carboxy-
terminal fragment 73-212 therefore resembles: native H] histone (Fig. 6b) both
qualitatively and quantitatively in its ability to discriminate between and
bind to superhelical and relaxed DNA. On the other hand, the carboxy-
terminal fragment 107-212 binds both forms of DNA equally well; virtually
all of the DNA I or DNA I" is complexed at an R value of about 75 (Fig. 6c).
The carboxy-terminal fragment 107-212 apparently has lost the specificity
for superhelical DNA molecules and binds DNA 1" with increased efficiency
relative to native H1 histone.

Binding of NBS Fragments to Nitrocellulose Filters. It has been
demonstrated previously '* that nitrocellulose filters quantitatively
retain 3H-H1 histone in the concentration range used in these studies.
The possibility that the differential ability of the various fragments of
Hl to complex DNA results from a differential retention by the nitro-
cellulose filters was investigated. SHH histone was cleaved with NBS
and the two tritium labeled fragments isolated in the same way as before.
Various concentrations of the fragments were incubated in a mock binding
experiment (standard binding buffer without DNA), and filtered as usual.
The results (Table II) demonstrate that over a concentration range of 0.1

 

2540
Nucleic Acids Research

 

TABLE IT
The Binding of N-Bromosuccinimide Fragments of SHH Histone to

Nitrocellulose Filters

 

Fragment Concentration 34-Fragment 73-212 Bound 34-Fragment 1-72 Bound

(ug/m1) cpm & cpm &
0.1 313 80.9 352 98.4
0.5 332 80.6 338 94.0
1.0 293 75.0 336 93.4
2.0 287 73.2 310 85.3
0.5 yg unfiltered 378 100 357 100

SHH (specific activity 800 cpm/ug) was cleaved with NBS and the
fragments purified as described in Methods. The fragment concentration
was calculated from the specific radioactivity, assuming that the specific
radioactivity of the fragments was the same as that of the starting H1.
0.5 ug of each of the fragments was diluted in the appropriate volume of
binding buffer to give the concentrations listed above. The samples were
incubated, filtered, washed and counted as in a standard binding assay.
0.5 ug (in 5 ul of binding buffer) was applied directly to a filter,
without washing and served as the control for 100% binding.

 

 

 

to 2.0 g/ml both the carboxy-terminal fragment 73-212 and the amino-
terminal fragment 1-72 are quantitatively retained on nitrocellulose filters.
Effect of Salt Concentration on the Interaction of DNA With HI and
H] Carboxy-Terminal Fragments. Previous work in this laboratory demon-
strated that H] interacts with superhelical DNA optimally at 0.1 M nact!®,
This is confirmed by the data in Fig. 7a. Fig. 7a also shows that
the extent of interaction of Hl histone with DNA I” (at a constant R value)
decreases with increasing salt, as would be expected for a purely electro-
static interaction. The effect of salt concentration on complex formation
between the carboxy-terminal fragments and both DNA I and DNA 1” was also
examined (Fig. 7). The reaction between the carboxy-terminal fragment 73-
212 and DNA I was optimal between 0 and 0.1 M NaCl; complex formation with
DNA I” decreased with increasing salt concentration (Fig. 7b). Thus the
salt dependencies of the reaction with Hl histone and with carboxy-terminal
fragment 73-212 are similar. The results obtained with carboxy-terminal
fragment 107-212 are markedly different (Fig. 7c). In this case, binding
of both DNA I and DNA I" by the carboxy-terminal fragment 107-212 was
highest in 0.2 M NaCl. Consistent with the observed loss of specificity
for superhelical DNA at 0.1 M NaCl, the carboxy-terminal fragment 107-212
does not distinguish between superhelical and relaxed DNA at any salt
concentration tested. It is surprising that this fragment, which is approxi-
mately one third lysine, should bind DNA optimally at 0.2 M NaCl. However,

 

2541
Nucleic Acids Research

 

 

8
as

>
”

BINDING (%)
886 £838 8 8

 

_
oe

 

 

 

 

 

NaCl (M)

Fig.7 Binding of HI and H1 carboxy terminal fragments to DNA as a function of

salt concentration. H1 and fragments, at constant R values, were incubated in 10 mM
Tris, pH 7.8, 1 mg/ml BSA, 1 mM EDTA with either DNA I or DNA I at various

NaCl concentrations for 15' at 23°. R values for DNA I were: fragment 73-212,

14,8; Hl, 15.2; fragment 107-212, 30.4. R values for DNA I* were: fragment 73-212.
92.4; Hl, 95.2; fragment 107-212, 38.0. DNAI:—;DNAF:---. a) intact H1;
b)carboxy-terminal fragment 73-212; c) carboxy-terminal fragment 107-212.

it is known from NMR studies that in the absence of salt, 15% of the lysine
residues of this fragment are not bound to DNA, whereas at 0.15 M NaCl, all
lysines are bound<’, Bradbury et al. proposed@/ that at 0.15 M NaCl,
intermolecular cross-linking of linear DNA molecules by the carboxy-terminal
fragment 107-212 can occur, mediated by the lysine residues which remained
unbound in the absence of salt. Alternatively, it is possible that in-
creasing the salt concentration induces a conformational change in the
fragment which enables more effective binding of the DNA. The superhelicity
of the DNA molecule should not be significantly affected over this range of
concentration. It should also be noted that the carboxy-terminus of the
intact molecule may also bind DNA optimally at 0.2 M NaCl but that this
effect would go largely undetected, due to the marked optimum at 0.1 M NaCl.
Interaction of Hi Histone With DNA in the Presence of Urea. The
observation that the carboxy-terminal fragment containing amino acid
residues 107-212 has lost the ability to discriminate between superhelical
and relaxed DNA, while the carboxy-terminal fragment 73-212 retains this
ability, suggested that the region of the native H] molecule between re-
sidues 73 and 106 is involved in the recognition of superhelical molecules.

 

2542
Nucleic Acids Research

 

 

 

 

 

 

 

160
b c
90} a - -
80+ - r
70+ - -
® eo} r r ?
/
2 50+ - “ r /
5 7 /
2 7 /
= 40+ - ’ r /
° a ’ /
30 ae - Yr 5 /
“ / /
20 Tr Ped - i - /
’ / /
y ?
10 / E / - i
te 1 1 sy Lo {ob
4 8 12 16 20 4 8 12 16 20 4 8 12 16 20
R

Fig. 8 Binding of Hl to DNA in the presence of urea. Filter binding assays were
done as described, except that urea was added to the incubation mixture before the
addition of either Hl or DNA. DNA I:—; DNA T:---. a)nourea; b) 4 M urea;
c) 6M urea.

It is known from both NMR and CD studies that this region of the molecule
is globular and contains a-helical structure !?, If the observed loss of
specificity for superhelices results from a loss of the structure in this
hydrophobic region, a similar loss might be expected when the binding assay
was performed under conditions that denature apolar regions of Hl. Bradbury
et al, '? have demonstrated that between 4 and 6 M urea, in 0.15 M NaCl,
nearly all of the globular structure of the Hl molecule is disrupted.
Therefore, native H] was tested for its ability to complex DNA I and DNA 1"
in the presence of 4 M (Fig. 8b) and 6M (Fig. 8c) urea. The results clearly
demonstrate that at both urea concentrations tested, H!] is able to complex
DNA. However, as the urea concentration is increased, the H] molecule
loses its ability to discriminate between superhelical and relaxed DNA. At
low R values, the binding of HT to DNA I decreases with increasing urea
concentration, whereas at higher R values, the binding of H1 to DNA 1"
increases. The net effect is that in 6 M urea, the binding of H1 to DNA 1°
is nearly the same as to DNA I. This is the same result as that obtained
with the carboxy-terminal fragment 107-212 in the absence of urea (Fig. 6c).
Since urea is known to denature DNA (unwind duplexes), it is possible
that some of the effect seen in Fig. 6c results from changes in the super-
helical density of DNA I and DNA I", Therefore, we studied the relative
sedimentation rates of DNA I and DNA I compared to nicked relaxed circular

 

2543
Nucleic Acids Research

 

duplex SV40 DNA in sucrose gradients in the presence of 6 M urea. The
results (not shown) indicated that there were no gross changes in the
relative superhelical densities of either DNA I or DNA 1",
DISCUSSION

H1 histone extracted either by salt or acid from a variety of tissues,
including calf thymus, rat hepatoma cells, cultured monkey cells (BSC-1),
chicken erythrocytes and HeLa cells, interacts with superhelical DNA in
preference to relaxed DNA duplexes /4+!5 (M. Bina-Stein et al, manuscript
in preparation; unpublished observ.). Here we have presented experiments
suggesting that the globular region of the molecule is involved in deter-
mining this specificity. The carboxy-terminal fragment of the molecule,
containing residues 73-212, retains both the ability to recognize super-
helical DNA and to bind DNA. It is known from other studies that this
fragment contains 1) a globular region similar to that in the intact HI,

and 2) a non-globular, highly lysine rich region, extending from approxi-

mately residue 123 to the carboxy-terminus at residue 21290, On the

other hand, the carboxy-terminal fragment, containing residues 107-212,
which has lost most of the globular region of the intact H] molecule, binds
DNA but no longer distinguishes between superhelical and relaxed DNA. It
binds both forms of DNA equally and with an efficiency somewhat lower than
that with which the intact H] molecule binds superhelical DNA. Therefore,
while the globular region of the H] molecule extending from amino acid
residues 73 to 106 confers on the H] molecule the ability to recognize
superhelical DNA, it also appears to limit the ability of the molecule to
interact with relaxed DNA duplexes. Since the carboxy-terminal fragment
107-212 binds DNA, the binding of the DNA molecule appears to be pre-
dominantly mediated by the lysine rich portion of the molecule beyond
residue 107.

Results consistent with this model were obtained by studying the
interaction of H] and DNA in 6M urea, where the globular structure of the
protein is denatured. In this case, the Hl] molecule lost the ability to
recognize superhelical DNA (at low R values, see Fig. 8) and gained enhanced
ability to complex relaxed DNA (at higher R values, see Fig. 8).

The amino-terminal fragment containing residues 1-106 does bind to
DNA, albeit with a low efficiency, and, moreover, displays specificity for
superhelical DNA. It is known that the amino terminal end of the molecule
contains a small lysine rich stretch from residues 1 to 407°, It is
possible that this cationic region can interact with DNA and stabilize the

 

2544
Nucleic Acids Research

 

recognition of superhelical DNA by the globular region. In the absence of
any globular region (amino-terminal fragment 1-72), this lysine rich
portion does not interact efficiently with either DNA I or DNA 1", under
the conditions used here.

The salt dependency curves for both intact H] and the fragment con-
taining residues 73 to 212 suggest that interaction with DNA 1" is purely
electrostatic, whereas interaction with DNA I is more complex. The re-
quirement for 0.1 M NaCl for optimal binding of H1 and fragment 73-212 to
DNA I may reflect a salt-induced change in the conformation of the H]
molecule related to its ability to recognize superhelical DNA.

Recent experiments (D. Singer, unpublished) indicate that the complexes
formed when H1] interacts with DNA I differ from those formed when H1 inter-
acts with DNA I", Velocity sedimentation gradient analyses of the H1-DNA
complexes demonstrate that at low H1/DNA ratios (w/w), H1] primarily forms a
soluble complex with DNA I, but forms only large aggregates with DNA 1",

These results suggest that the interaction of H1] with superhelical DNA
has two components: 1) recognition of superhelicity by the globular region
of the molecule extending from amino acid residues 73-106, and 2) inter-
action between the DNA and the lysine rich portion of the molecule beyond
amino acid residue 106. The overall reaction may also involve a conforma-
tional change in the H] protein such as to stabilize the superhelical turns
in the DNA. It should be noted that while this discussion does not imply
any particular structural description of superhelical turns, it does suggest
that a physical conformation characteristic of superhelicity is recognized
by Hl. Partial single-strandedness, known to exist in superhelices®! #9,
is not likely to be involved since Ht binds much more poorly to single-
stranded DNA than to double-stranded DNA 13,35

It is interesting to note that whereas there is considerable variation
in the amino acid sequence of different H1 histones in the regions of amino
acid residues 1-40 and 110-212, the amino acid sequence of the globular
region of the molecule is nearly invariant (R. Cole, personal communication).
Therefore, it may be that recognition of superhelical configurations of
DNA by H1 is an important component of the physiological role of H] histone.
This speculation is of particular interest in view of recent proposals for
chromatin structure, all of which involve some sort of folding or super-
coiling of the DNA duplex about the nucleosome. The nature of the inter-
action of H1 histone with nucleosomes is currently being investigated.

 

2545
Nucleic Acids Research

 

ACKNOWLEDGEMENTS
We gratefully acknowledge helpful discussions with Drs. M. Bina-Stein
and C. Klee. We thank Mr. J. Williams for performing the amino acid analyses.

REFERENCES

1. A preliminary report of this work was presented at the meeting of

the American Society of Biological Chemists, June 6-10, 1976.

(Fed. Proceedings, 35, 1722).

Kornberg, R.D. (1974) Science 184, 868-871.

Olins, A. and Olins, D. (1974) Science 183, 330-332.

Van Holde, K., Sahasrabuddhe, C., and Shaw, B. (1974) Nucl. Acids

Res. (1974) 1, 1579-1585.

Hewish, D.R. and Burgoyne, L A. (1973) Biochem. Biophys. Res. Commun.

52, 504-510.

nn R. J. and Smith, E. L. (1971) Ann. Rev. Biochem. 40,

9-314.

Langan, T. A. and Hohmann, P. (1975) Chromosomal Proteins and

their Role in the Regulation of Gene Expression. Academic Press,

N.Y. pp. 113-125.

8. Balhorn, R., Chalkley, R. and Granner, D. (1972) Biochemistry
11, 1094-1097.

9. Lake, R. S. and Salzman, N. P. (1972) Biochemistry 11, 4817-4826.

10. Bradbury, E. M., Inglis, R. J., and Matthews, H.R. (1974) Nature
247, 257-261.

11. Bustin, M. and Cole, D. (1969) J. Biol. Chem. 244, 5291-5294.

12. Bradbury, E. M., Cary, P. D., Chapman, G. E., Crane-Robinson, C.,
Danby, S.G.E., Rattle, H., Boublik, M. and Palau, J. (1975)
Eur. J. Bioch. 52, 605-613.

13. Vogel, T. and Singer, M. (1975) J. Biol. Chem. 250, 796-798.

14. Vogel, T. and Singer, M. (1972) Proc. Natl. Acad. Sci (USA)
72, 2597-2600.

15. Vogel, T. and Singer, M. (1976) J. Biol. Chem. 251, 2334-2338.

16. Bottger, M., Scherneck, S., and Fenske, H. (1976) Nucl. Acids Res.
3, 419-429.

17. Sebring, F. D., Kelly, T.J. Jr., Thoren, M. M. and Salzman, N. P.
(1971) J. Virol. 8, 478-490.

18. Champoux, J. J. and Dulbecco, R. (1972) Proc. Natl. Acad. Sci. (USA)
69, 143-146.

19. Germond, J., Hirt, B., Oudet, B., Gross-Bellard, M. and Chambon, P.
(1975) Proc. Natl. Acad. Sci. (USA) 72, 1843-1847.

20. Keller, W. (1975) Proc. Natl. Acad. Sci. USA) 72, 2550-2554.

21. Bohm, E. L., Strickland, W. N., Strickland, M., Thwaits, B. H.,
Van der Westhuizen, D. R. and Von Holt, C.'(1973) FEBS Letters
34, 217-221.

22. Ohlenbusch, H., Olivera, B. M., Tuan, D., and Davidson, N.
(1967) J. Mot. Biol. 25, 299-315.

23. Panyin, S. and Chalkley, R. (1969) Arch. Biochem. Biophys. 130,
337-364.

24. Kinkade, J. M., Jr. and Cole, R. D. (1966) J. Biol. Chem.
241, 5790-5797.

25. Rall, S. C. and Cole, D. (1971) J. Biol. Chem. 246, 7175-7190.

26. The only Hi histone fraction that has been sequenced to position
106 is rabbit thymus fraction RTL-3, which has a phenylalanine at
position 106 (33). Peptide analysis of calf thymus histone indicates

~ a go fm WP
. s . e # e@

 

2546
27.
28.

29.
30.

31.
32.
33.

34,
35.

Nucleic Acids Research

a similar placement of phenylalanine in this histone (34) but the
precise position is not known. In this paper, the phenylalanine
position is referred to as position 106 by analogy with the RTL-3
fraction. The position of the single tyrosine residue in the calf
thymus H1] is at residue 72 (11).

Bradbury, E. M., Chapman, G., Danby, S., Hartman, P. and Riches, P
(1975) Eur. J. Biochem. 57, 521-528.

Tai, H. T., Smith, C. A. and Sharp, P. A. and Vinograd, J. (1972)
J. Virol. 9, 317-325.

Eason, R. and Vinograd, J. (1971) J. Virol. 7, 1-7.

chopeans G., Hartman,P., and Bradbury, £. (1976) Eur. J. Biochem.
6 9-75.

Woodworth-Gutai, M. and Lebowitz, J. (1976) J. Virol. 18, 195-204.
Beard, P., Morrow, J. and Berg, P. (1973) J. Virol. 6, 1303-1313.
Jones, G.M.T., Rall, S.C., and Cole, R. D. (1974) J. Biol. Chem.
249, 2548-2553.

Langan, T.A. (1971) Fed. Proc. 30, 1089.

Akinrimisi, E.D., Bonner, J., and Ts'o, P.D.P. (1965) J. Mol. Biol.
11, 128-136.

2547