LIIZKJ‘LRY

Michigan Sum
University

 

This is to certify that the

thesis entitled

THE NON-SPECIFIC DNA BINDING
ACTIVITY OF CATABOLITE
ACTIVATING PROTEIN OF E. COLI

presented by

Stephen Alan Saxe

has been accepted towards fulfillment
of the requirements for

Master OLScjence— degree in Biochemistry l

mfimcw. EDAW

Major professor

Date—MW

0-7639

 

THE NON-SPECIFIC DNA BINDING ACTIVITY OF CATABOLITE

ACTIVATING PROTEIN OF E. COLI

BY
Stephen Alan Saxe

A THESIS

Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of

MASTER OF SCIENCE

Department of Biochemistry

1978

ABSTRACT

THE NON-SPECIFIC DNA BINDING ACTIVITY OF CATABOLITE
ACTIVATING PROTEIN OF E. COLI

BY
Stephen Alan Saxe

The non-specific binding of catabolite activating
protein (CAP) of E. coli to double-stranded DNA has been
studied by sedimentation velocity and circular dichroism
techniques. It was found that C00perative binding, i.e.,
the binding of protein molecules in clusters along the DNA,
occurs in the absence of CAMP whereas the binding is non«
cooperative when CAMP is present. Circular dichroism
measurements were used to determine that about 13 base
pairs of DNA are covered by a molecule of bound CAP for
both the cooperative and noncooperative binding. Both
types of binding are very ionic strength dependent. Values
for the intrinsic association constant of the protein to
DNA and for a cooperativity parameter which measures the
extent of protein-protein interactions have been determined
for the cooperative binding of CAP to calf thymus DNA over
the range 50-80 mM Na+. The results imply that in viva CAP

is bound to the chromosome whether CAMP is present or absent.

DEDICATION

To my parents

ii

ACKNOWLEDGEMENTS

I would like to express my deepest appreciation to
Dr. Arnold Revzin for his patience, his guidance, and his
continual encouragement which caused me to keep trying.
I would also like to thank Gail McDole and Mildred Martin

for growing the bacteria used for these studies.

iii

TABLE OF CONTENTS

Page
LIST OF TABLES. . . . . . . . . . . . . . . . . . . vi
LIST OF FIGURES . . . . . . . . . . . . . . . . . . vii
LIST OF ABBREVIATIONS . . . . . . . . . . . . . . . ix
INTRODUCTION. . . . . . . . . . . . . . . . . . . . 1
LITERATURE REVIEW . . . . . . . . . . . . . . . . . 4
Discovery of CAP . . . . 4
Binding of Nucleotides by CAP. . . . . 5
Binding of CAP to DNA. . . . . . . . . . . . 7
Biological Activity of CAP . . . . . . . 10
C00perative Binding of Proteins to DNA . . . 12
MATERIALS AND METHODS . . . . . . . . . . . . . . . 17
Materials. . . . . . . . . . . . . . . .‘. . 17
Methods. . . 17
PurifiCation of Catabolite Activating
Protein (CAP) . . . . . . . . . . . 17
Protein Measurement . . . 20
Measurement of CAMP Binding to CAP. . 20
Gel ElectrOphoresis . . . . . . . 21
Determination of Extinction
Coefficients. . . . 21
Measurement of CAP Binding to DNA . . 23
Calibration of Model B Absorbance
Readings. . . . 23
Determination of Binding Site Size. . 24
DNA Thermal Melting Experiments . . . 24
Paper Chromatography of CAMP. . . . . 24
RESULTS . . . . . . . . . . . . . . . . . . . . . . 26
Purification of CAP. . . . . 26
Determination of Extinction Coefficients
for CAP. . . . . . . . . . 28
DNA Thermal Melting Experiments. . . . . . . 29
Determination of Binding Site Size . . . . . 30
Binding of CAP to DNA. . . . . . . . . . . . 39
IDISC USSION . . . . . . . . . . . . . . . . . . . . 59

iv

Page

Binding of CAP to DNA. . . . . . . . . . . . 60
In vivo Implications . . . . . . . . . . . . 64
LIST OF REFERENCES. . . . . . . . . . . . . . . . . 67

Table

II

III
IV

VI

LIST OF TABLES

Physical Properties of CAP

Extinction Coefficients (1/cm-mole) of
Tryptophan, Tyrosine, and Cysteine in 6 M
Guanidine Hydrochloride, pH 6.5 (from
Bdelhoch, 1967). . . . .

, Purification of CAP.

Extinction Coefficients for Native CAP

Relationship Between [CAP] and Sedimenta-
tion Coefficient . . . . . . . . .

Effect of Ionic Strength on Binding
Parameters . . . . . . . . .

vi

Page

21
27
28

43

58

Figure

10

11

LIST OF FIGURES

The genetic map of the lac genes

Thermal melting curves for poly dCA- T) plus
CAP in 1.0 mM NaZHPO4, 0.1 mM NazEDTA,
pH 7. 7 . .

CD spectra of CAP and CTDNA in 5.0 mM
NazHPO4, 15.0nmINaCl, 0.1 mM NazEDTA,
0.05 mM DTT, pH 7.7. . . . .

Difference CD spectrum obtained by titrat-
ing CAP with CTDNA in the presence of CAMP

Difference CD spectrum obtained by titrat-
ing CAP with CTDNA in the absence of CAMP.

Difference CD spectrum obtained by titrat-
ing CTDNA with CAP

Model E ultracentrifuge scan of sedimenta-
tion of 1.0 x 10- 6 M CAP plus 3.4 x 10-5 M
CTDNA in 10.0 mM Tris- HCl, pH 7. 9, 0.1 mM
NazEDTA, 0.05 mM CAMP, 100.0 mM NaCl

Model B ultracentrifuge scan of sedimenta-
tion of 1. 8 x 10- 6 M CAP plus 2. 8 x 10-5 M
CTDNA in 10.0 mM Tris HCl, pH 7. 9, 0.1 mM
NazEDTA, 69. 8 mM NaCl. . . . . . . .

Scatchard plot for the cooperative binding
of CAP to CTDNA at 22°C in 10 mM Tris-HCl,
0.1 mM NazEDTA, 49.8 mM NaCl, pH 7.9

Scatchard plot for the C00perative binding
of CAP to CTDNA at 22°C in 10 mM Tris-HCl,
0.1 mM NazEDTA, 59.8 mM NaCl, pH 7.9

Scatchard plot for the cooperative binding

of CAP to CTDNA at 22°C in 10 mM Tris-HCl,
0.1 mM NazEDTA, 69.8 mM NaCl, pH 7.9

vii

Page
13

32

34

36

38

41

45

48

51

S3

55

Figure

12

13

Scatchard plot for the cooperative binding
of CAP to CTDNA at 22°C in 10 mM Tris-HCl,
0.1 mM NazEDTA, 79.8 mM NaCl, pH 7.9 .

Binding constant as a function of [Na+]

for CAP binding to CTDNA at 22°C in 10 mM
Tris-HCl, 0.1 mM NazEDTA, plus NaCl.

viii

Page

57

63

ADP

AMP, S'-AMP

3'-AMP

ATP

CAMP

2',3'-CAMP

CAP

CD

CGMP

CTDNA

CTP

CTuMP

D209W

2'-dA

DEAE-Cellulose
(DE-52)

DNase

DTNB

DTT

E. coli

EDTA

gal

GMP

GTP

1,5-I-AENS

K

Kd
L
Zac

n
Nsz
pl

PMSF
poly d(A-T)

poly d(I-C)
jpoly(rC)
IPPO

RNase

ESDS
'Tris-HCl

LIST OF ABBREVIATIONS

adenosine 5'-diphosphate
adenosine 5'-monophosphate

.adenosine 3'-monophosphate

adenosine 5'-triphosphate
adenosine 3',5'-monophosphate
adenosine 2',3'-monophosphate
catabolite activating protein
Circular dichroism

guanosine 3',5'-mon0phosphate
calf thymus DNA

cytidine 5'-triphosphate
tubericidin 3',5'-monophosphate
diffusion coefficient
2'-deoxyadenosine
diethylaminoethyl cellulose

deoxyribonuclease
5,5'-dithiobis(2-nitrobenzoic acid)

dithiothreitol

Escherichia coli

ethylenediaminetetraacetiC acid

the galactose operon

guanosine 5'-monophosphate

guanosine S'-triphosphate

N-(iodoacetylaminoethyl)~1-naphthyl-
amine-S—sulfonate

intrinsic association constant

dissociation constant

free ligand (CAP) concentration

the lactose Operon

binding site size
5,5'-dithiobis(Z-nitrobenzoic acid)
isoelectric point

phenylmethylsulfonylfluoride

3'+5' copolymer of alternating deoxy-
adenosine-deoxythymidine

3’+5' copolymer of alternating deoxy-
inosine-deoxycytidine

3'+5l polymer of ribocytidine
2,5-diphenyloxazole

ribonuclease

sodium dodecylsulfate
tris(hydroxymethyl)aminomethane
hydrochloride

ix

tryptophan

tyrosine

uridine S'-triphosphate

ultraviolet

extinction coefficient
1,N6-ethenoadenosine 3',5'-monophosphate
bound CAP/total DNA

cooperativity parameter

INTRODUCTION

Much recent work has been directed towards elucidation
of the molecular basis for control of cellular processes.
Probably the most studied and best understood control
system involves the regulation of transcription at the
lactose Operon of E. 001i. The Zac repressor, RNA polymer-
ase, catabolite activator protein (CAP), and CAMP are all
involved in this process. Transcription from the lac
Operon is under both negative and positive control.
Negative control is mediated by lac repressor which binds
to the Operator region of DNA and prevents transcription
from occurring. (For a review see Beckwith and Zipser,
1970.) The binding of repressor to operator can be reduced
in the presence of small molecule inducers which bind to
the protein and diminish its affinity for the operator.
Removal of repressor permits transcription to proceed.

Positive control of Zac mRNA production is involved
in the phenomenon of catabolite repression. This is
ciescribed as the decreased rate of production of specific
<3nzymes ("catabolite-sensitive" enzymes) which occurs

then glucose (or similar compounds, e.g., glucose-6-
Ffliosphate or gluconic acid) is present in the growth
nRadium. The presence of glucose decreases the cellular
lsavel of CAMP whereas addition of CAMP to the growth medium

1

2
can overcome this repression (Perlman and Pastan, 1968).
The CAMP effect is mediated by a protein (CAP) which can
bind CAMP. The CAP-CAMP complex binds to the promoter
region and enhances transcription at catabolite-sensitive
operons (de Crombrugghe et a1., 1971), but the mechanism of
this action is as yet obscure.

Our overall goal is to understand the molecular mechanisms
involved in Zac operon control, especially the manner in which
CAP may stimulate RNA polymerase activity. As one approach
to elucidate the details of CAP function we have performed a
quantitative study of the interaction of CAP with DNA. Our
emphasis is on the situation where the CAMP level is low or
zero. It is shown that in the absence of CAMP, CAP will
bind cooperatively to DNA which does not contain a catabolite
sensitive operon. The association constant for this "non-
specific" CAP-DNA interaction has been determined over a
range of ionic strengths and the protein-protein cooperativity
has been Characterized. It is obviously desirable to measure
the binding constant of CAP to the lac operon itself, but
this has proved difficult due to the high affinity of CAP
for non-specific DNA (Majors, 1975).

Study of non-specific binding may yield information
aibout the specific CAP-DNA interaction. For example, the non-
SIDeCifiC association constant shows a dependence on CAMP
(PJissley et a1., 1972). Furthermore, one may be able to draw
irrferences about the specific binding through study of, for

exxlmple, the effects of Mg++ on non-specific binding or of

3

variations in CAP affinity for single-stranded and double-
helical DNAs of different base compositions and sequence.

Finally, work done with Zac repressor has shown that
most of the repressor in viva is not free in solution
but rather is bound to non-specific DNA (Kao-Huang et a1.,
1977). The results to be reported here indicate that
much of the CAP may also be bound to non-specific DNA
irrespective of the CAMP level in viva. Thus, non-specific
binding greatly affects the concentrations of regulatory
molecules free in the cytOplasm and must be considered as

an element of Zac operon control.

LITERATURE REVIEW

Discovery of CAP

The phenomenon of catabolite repression was observed
as long ago as 1900 (Dennert, 1900). Epps and Gale (1942)
described this effect as the suppression of the formation
of certain enzymes by the presence of glucose in the growth
medium. As an example, the presence of lactose or galactose
in the growth medium will normally induce synthesis of
enzymes necessary for their catabolism. These enzymes are
coded for by the Zac and gal operons, respectively. How-
ever, when glucose is also present it prevents induction
of the enzymes coded for by these two catabolite-sensitive
operons. Real progress in understanding the mechanism of
this effect did not begin until it was shown that the
cellular concentration of CAMP rapidly decreased in the
presence of glucose (Makman and Sutherland, 1965). From
there Perlman and Pastan (1968) and Ullman and Monod
(1968) showed that addition of CAMP to the medium in which
t11e bacteria were growing could overcome the repression due
t<3 glucose. Another major step occurred when Zubay et al.
(11970) and Emmer et a1. (1970) isolated bacterial mutants
111 ‘which CAMP does not relieve catabolite repression.
Tiiirs increased speculation that a protein might be neces-

sarfiy to mediate the action of CAMP. Zubay et a1. (1970)

4

5

partially purified such a protein, which has been named
catabolite activating protein, or CAP. Their assay for
this protein was a cell-free system, derived from their
mutant bacterial strain, for synthesizing the catabolite-
sensitive enzyme, B-galactosidase. Addition of CAP to
the cell extract increases the synthesis of this enzyme.
Emmer et a1. (1970) also partially purified CAP. Their

3H-CAMP to

assay involved measurement of the binding of
their protein fractions. CAP can now be purified to
apparent homogeneity (Anderson et a1., 1971). Some of

the physical Characteristics of CAP are given in Table I.

Table I. Physical PrOperties of CAP

 

Molecular weight of CAP 44,600 daltons
Molecular weight of each subunit 22,300 daltons

pI 9.12 _7 2
D20,w 7.7 x 10 cm /seC
d-helix 31%

~SH groups 4

Partial Specific volume, V 0.752 ml/g
Frictional coefficient, f/fo 1.17

Sedimentation Coefficient, 820’W 3.53

 

Table is from Anderson et a1., 1971.

Binding of Nucleotides by CAP

 

Purified CAP has a CAMP binding activity. Emmer et
ail. (1970) determined a dissociation constant, Kd, for the
CIXPHCAMP complex of 1 x 10'6 M. This was measured by incu-

bating CAP with 3

H-CAMP and then precipitating the
CMXP’-CAMP complex with (NH4)ZSO4. Because the (NH4)ZSO4

may'- affect the equilibrium, other workers have used

6
equilibrium dialysis to quantitate the CAP-CAMP binding.
Zubay et a1. (1970), using their partially purified CAP,
measured a Kd = 1.7 x 10.5 M in a 10 mM Tris-acetate,
pH 8.2, 10 mM Mg-acetate, 60 mM K-acetate, 1.4 mM DTT
buffer. Anderson et a1. (1971) used the same buffer and

6

found K = 9.1 x 10- M.

d
Several other nucleotides have been tested for their
ability to bind to CAP. No binding was observed for
3'-AMP, 5'-AMP, ADP, ATP, GTP, or 2'-dA (Emmer et a1.,
1970). Anderson et a1. (1972) measured the ability of
several CAMP analogues to compete with CAMP for binding
to CAP. Cyclic TuMP competed very well while other
analogues competed less strongly or not at all. Cyclic
3',S'-GMP is a competitive inhibitor of CAMP binding to

S M (Emmer et a1., 1970).

CAP and has a Kd = 1-2 x 10'
CAP undergoes a conformational Change upon binding

CAMP. This is seen from the result that CAP to which

CAMP is bound becomes more susceptible to proteases
(Krakow and Pastan, 1973). This attack of the CAP-CAMP
complex by proteases produces a resistant protein fragment
called the a-core (Eilen and Krakow, 1977) having a
Inolecular weight of 12,500 as compared to 22,300 for the
analtered subunit. It is capable of binding CAMP and con-
tziins two sulfhydryl groups. These are buried but can be
eacposed by denaturation. The a-core is rapidly denatured
i;E added to a 3 M urea solution in the absence of CAMP.

Thcis is measured by monitoring the rate of reaction of the

sulgfhydryl groups with Nbs2 following addition of the

d-Core to the urea solution. In.the presence of CAMP the
rate of reaction is much slower, indicating a resistance
to denaturation. Cyclic TuMP and CGMP affect the a-core
in a manner similar to CAMP. Eilen and Krakow (1977)
proposed that the CAMP "tightens” the fragment, thus
increasing its resistance to denaturation. Although CGMP
also caused this ”tightening” effect, it does not make
CAP susceptible to proteases.

Wu and Wu (1974) also observed a conformational
Change in CAP by attaching a fluorescent probe to it and
observing a relaxation process upon addition of CAMP.
The presence of the probe had little effect on both the
ability of CAP to bind CAMP and on its ability to stimulate
in vitra gal transcription. Wu et a1. (1974) alSo observed
a fluorescence enhancement and a blue shift upon binding
CAMP to the labeled CAP. These Changes were observed only

in the presence of Mg++. Cyclic TuMP and N6,O2

t

-dibutyryl-
CAMP, both of which are active in viva, also enhance and
shift the fluorescence spectrum. Cyclic GMP and e-CAMP

cause a quenching of the fluorescence.

Binding of CAP to DNA

 

A second "activity" of CAP is its ability to bind to
DNA. Much of the work done to date studying the binding
of CAP to DNA has utilized non-specific DNA and has been
qualitative or semiquantitative. Riggs et a1. (1971), using
a nitrocellulose filter assay, reported that binding of CAP

to DNA occurs both in the presence and in the absence of

8
CAMP. Under their conditions the presence of CAMP
increased the amount of binding, whereas CGMP prevented
binding. The binding was unaffected by AMP and GMP.
They suggested that binding which occurs in the absence
of CAMP may be cooperative since they observed sigmoidal
binding curves. In a "Note Added in Proof" they retracted
this statement and instead Claimed that no binding of CAP
to DNA occurs unless CAMP is present. In any case, they
observed no specificity of binding, CAP being bound
equally well to poly d(A-T) and to DNA from salmon sperm,
Clastridium perfringens, and Micracaccus Zuteus, and to
DNA from bacteriophage Ah 80 and Ah 80d lac which contains
a catabolite-sensitive promoter. Binding of CAP to E. aali
rRNA and tRNA is much less tight than to DNA.

Nissley et a1. (1972) looked at the binding of CAP to
various DNAs using both nitrocellulose filter assays and
band sedimentation in sucrose density gradients. There
was binding to all of the DNAs tested, including the sepa-
rated strands of Ap gal and Ap lac. Binding was CAMP
dependent under their ionic conditions (20 mM Tris-HCl,
pH 7.8, 10 mM MgC12, 100 mM KCl for their filter assay,

40 mM Tris, pH 7.9, 10 mM MgCl 40 mM KCl for the sedi-

2)
mentation studies). To Check for specificity of binding,
they performed competition experiments. This involved

32p-1abe1ed Ah 80d lac pS DNA (which

incubating CAP and
contains a promoter that presumably binds CAP tightly)
with varying concentrations of unlabeled non-lac containing

DNAs and passing the mix over a nitrocellulose filter. If

9
binding to the lac DNA is much stronger than to the other
DNA, the presence of the unlabeled DNA should have little
effect on the amount of lac-CAP complex formed. However,
all of the DNAs tested caused a decrease in counts bound
to the filter as the concentration of unlabeled DNA was
increased. This indicates competition between the DNAs
for binding of CAP; the data implied that the non-lac and
lac DNAs bound CAP with about the same affinity.
Krakow and Pastan (1973) used the nitrocellulose

filter assay to show that half maximal binding of CAP to

7 M CAMP and half maximal

5

poly d(I—C) occurs at 7 x 10-
binding of CAP to poly d(A-T) occurs at l x 10' M CAMP.
The optimum pH for binding of CAP to DNA is 8.0. No
binding occurs at pH 10.0 and there is binding at pH 6.0
.in both the presence and absence of CAMP. The a-core
does not show any CAMP dependent binding but does retain
CAMP independent binding at pH 6.0 under their ionic con-
ditions. Eilen and Krakow (1977) showed that reacting
the two available sulfhydryl groups of CAP with DTNB
eliminates CAMP dependent DNA binding. However, Wu and
Wu (1974) stated that reacting 1,5-I-AENS to the sulf-
hydryl groups of CAP had no effect on its CAMP binding
activity nor its ability to promote gal transcription.

Wu et a1. (1974) reported that CAP labeled with a
fluorescent probe undergoes a fluorescence Change when it
binds to Ah 80d lac DNA in the presence of CAMP. This

Change is not seen upon binding to DNA without the lac

promoter. They stated that this is indicative of a unique

10
conformational Change in CAP upon binding to the lac
promoter, but their data are questionable. They reported

9 M Ah 80d lac DNA with 5.3 x 10‘7 M

that mixing 6 x 10-
CAP causes a 15% reduction in fluorescence. This seems
unlikely since only about 1% of the CAP molecules could

be bound to promoter sites in their solution.

Majors (1975) has apparently shown specific binding
of CAP to promoter DNA. Using a filter assay, he found
that CAP bound more strongly to short DNA fragments con-
taining the lac promoter than it did to fragments without
this promoter. While his data seem convincing, his experi-
ment apparently requires certain special reaction condi-
tions. This is consistent with the other data reviewed

here that demonstration of the specific CAP-DNA interaction

remains a difficult problem.

Biological Activityof CAP

 

In vitra studies show that CAP plus CAMP promotes
transcription at catabolite—sensitive operons. Zubay et
a1. (1970) used a DNA-directed cell-free system to monitor
this stimulation of B-galactosidase. In the presence of
CAMP, there was a linear increase in the synthesis of B-
galactosidase with a corresponding increase in the amount
of CAP added to the assay system. Without CAMP the addi-
tion of CAP had no effect.

Nissley et a1. (1971), using an RNA-DNA hybridization
assay, measured the type of RNA formed in a transcription

system containing Ap gal 8 DNA and RNA polymerase of E.

ll
cali. Addition of CAP and CAMP to this system caused a
specific 15-fold increase in gal mRNA synthesis. No
increase in mRNA synthesis was seen if non-gal DNA was
used. The addition of 2',3'-CAMP or S'-AMP to the tran-
scription system had no effect on the rate, whereas
3',S'-CGMP inhibited transcription.

In an effort to determine the stage of transcription
(e.g., initiation, elongation) at which CAP acts, Nissley
et a1. (1971) performed transcription assays by preincubat-
ing CAMP, CAP, DNA, and RNA polymerase and then adding
ATP, GTP, CTP, UTP, and rifampicin. Stimulation of gal
transcription, relative to the control assay without CAMP
or CAP, was observed under these conditions. If either
CAMP or CAP was added with the rifampicin and not earlier,
no stimulation of transcription was seen. Since rifampicin
prevents initiation but not elongation, it was prOposed
that during the preincubation a rifampicin-resistant
complex of CAP, CAMP, DNA, and RNA polymerase was formed
and a single transcript was then made. These results
indicate that CAP acts at initiation.

In Ap lac, the lac mRNA is complementary to the
Ap lacL strand of DNA. Eron et a1. (1971) reported that
in vitra in the absence of CAP and CAMP there is much
transcription from the incorrect Ap lacH strand. Addition
of CAP and CAMP to the system increases transcription from
the proper strand and decreases other transcription. This
stimulation depends on the presence of sigma factor in the

RNA polymerase.

12

The effect of Mg++ on transcription at lac, gal, and
A promoters was studied by Nakanishi et a1. (1975). Mg++
strongly interferes with formation of ”open” complexes
(i.e., essentially irreversible promoter-RNA polymerase
complexes) at the lac and gal promoters but has less of
an effect at the A promoters. Local denaturation of the
DNA may be required to form these open complexes. Mg++
stabilizes the DNA making it resistant to denaturation,
and it appears that CAP plus CAMP can overcome this inhibi-
tory effect at the lac and gal promoters. It has been
shown genetically that the order of the lac genes is i,
p, o, z, y, a (Magasanik, 1970). Thus, the interaction

site of CAP at promoter is directly adjacent to the RNA

polymerase interaction site (Figure 1).

.C00perative Binding_of Proteins to DNA

 

CAP binds cooperatively to DNA under the conditions
described in this thesis. A discussion of the cooperative
binding of proteins to DNA will be helpful in understanding
the results to be presented.

Several proteins have been isolated which bind coopera-
tively to DNA. Included among these are E. cali unwinding
LDrOtein (Molineux et a1., 1974), gene 5 protein of bacterio-

thage fd (Alberts et a1., 1972), and gene 32 protein of
‘bacteriophage T4 (Alberts and Frey, 1970). These all bind
to single-stranded DNA, but they do show some differences.
than gene 5 protein binds to closed Circular single-stranded

UNIX it causes the DNA to collapse into a rod (Pratt et a1.,

as» new movou ocou w ash

.hommanaoh ee~ we mgmosucxm

.mocou CCN 029 no gas ufiuocou 0

ea .H osswam

 

 

 

 

 

T m.>.~ o a _
u 22mm:
zch<mmmc Gyros—Om:
35:5 35 £5 33
4M4»: :e_uom..ou:: gaze—:3:— :o_uum..ou:. 020w;
macaw Commocaum mambo—:35. <2: :_3o..m.m<u

 

14

1974). In contrast, gene 32 protein acts as a DNA unwind-
ing protein (Alberts and Frey, 1970). It is required for
DNA replication (Epstein et a1., 1963), repair (Wu and Yeh,
1973) and genetic recombination (Tomizawa et a1., 1966).
The ability of gene 32 protein to perform these functions
probably depends on its c00perative binding to the single-
stranded DNA (Alberts and Frey, 1970; Delius et a1., 1972;
Alberts et a1., 1968). At high concentrations, the gene

32 protein self-aggregates and it is probably these protein-
protein interactions which cause the cooperative binding
(Carroll et a1., 1975). Since cooperative binding seems

to have a physiological role for gene 32 protein, it is
likely that the cooperative binding of other proteins will
also serve a purpose.

Quantitation of the cooperativity has been performed
for a few DNA-protein systems. A definition, on a quanti-
tative level, of cooperativity is given in an excellent
paper by McGhee and von Hippel (1974), in which they
describe the theoretical aspects of DNA-protein interaction.
They begin by showing that the Scatchard (1949) plot
obtained when a protein binds noncooperatively to DNA is
curved rather than linear. For the theoretical case of a
protein binding equally well anywhere on the DNA, this
curvature is due solely to the fact that the protein covers
more than one base pair. As the binding density increases,
gaps develop along the DNA which are too small to allow
other proteins to bind without a rearrangement of the

previously bound proteins. This causes a sort of negative

15
COOperativity since free energy must be expended to cause
this rearrangement. If a protein covered only a single
base pair on DNA, this negative cooperativity would not
appear and a linear Scatchard plot would result. The
equation which they develop to describe this noncooperative

binding of a protein to DNA is of the form
v/L = f(K,n,v) (3)

where v = [bound protein]/[tota1 DNA], L = [free protein],
K = the intrinsic association constant, i.e., the associa-
tion constant for the protein binding to an isolated region
of DNA, and n = the number of base pairs (or bases for
single-stranded DNA) covered by a single protein molecule.
When cooperativity is present, another parameter, w,
must be introduced. This measure of the cooperativity is
defined as:
the probability of the protein binding next
to another protein already on DNA

the pr5bability of tHE proteinbinding to
an isolated region of DNA

Ruyechan and Wetmur (1975) measured a value oft» =

S for E. cali DNA unwinding protein binding to

2.7 x 10
single-stranded DNA in 0.15 M Na+. This value decreased
with.an increase in the ionic strength. Their study was
performed using electron micrOSCOpy and a statistical

inechanical model to analyze the data. A problem with the

method is that the preparation of the protein-DNA complex

for electron micrOSCOpy may alter the binding.

16

Alberts and Frey (1970) determined that gene 32—
protein binding to single-stranded DNA has a w of at least
80. This was determined by running a mixture of gene 32-
protein and fd single-stranded DNA on a sucrose gradient
and noting that strong cooperative binding occurred even
when the number of isolated sites was 80 times as great
as the number of contiguous sites. Jensen et a1. (1976)
and Kelly et a1. (1976) also studied the cooperative
binding of gene 32-protein to single-stranded DNA. They

reported a value of w = 103

and it appears to be relatively
independent of ionic strength. Jensen et a1. (1976) used
thermal melting data to determine this value, whereas Kelly
et a1. (1976) used a fluorescence quenching technique to
measure binding of the protein to polynucleotides and to
short oligonucleotides capable of binding only one protein
molecule. The ratio of these two values is w.

Finally, Draper and von Hippel (1978) measured a value
of w = 31 for poly(rC) binding to site II of E. cali ribo-
somal protein 81. They used fluorescence titrations and
compared the relative binding of site II to short oligo-
nucleotides and to polynucleotides.

Research on these systems of protein binding coopera-
tively to DNA is continuing and questions concerning the
mechanisms and function of C00perativity are being asked,
as seen by a recent paper by Williams and Konigsberg (1978),
who determined which part of gene 32-protein of bacteriophage

T4 is involved in protein-protein interactions and which

part is involved in protein-DNA interactions.

MATERIALS AND METHODS

Materials

 

Catabolite activating protein (CAP) was purified from
E. cali strain CR63, which was a gift from Dr. Loren
Snyder. DNase, RNase, lysozyme, bovine serum albumin,
casein, trypsin, calf thymus DNA, poly d(A-T), poly d(I-C),
3',S'-cyCliC AMP, 5'-AMP, 3',S'-CGMP, phenylmethylsulfonyl-
fluoride, deoxycholic acid, guanidine-HCl, bromphenol blue,
and Coomassie brilliant blue R were obtained from Sigma
Chemical Company. 3H-CAMP (26 Curie/mmole) was from
Amersham/Searle Corporation. DEAE-Cellulose (DE-52),
cellulose phosphate (coarse fibrous P1), and cellulose
powder CF11 were purchased from Whatman. Sephadex G-75
was from Pharmacia. Acrylamide and bis-acrylamide were
purchased from Bio-Rad Laboratories and 2,5-diphenyloxazole
(PPO) was from Research Products International Corporation.
All other chemicals were the highest grade available and

were purchased from the usual commercial sources.
Methods

Purification of Catabolite Activating_Protein (CAP)
CAP was purified from E. cali by a modification of
the procedure of Anderson et a1. (1971). A typical puri-

fication was as follows:
17

18

Preparation of crude extract: Two hundred fifty
grams of frozen E. cali were placed in 500 m1 of 4°C
buffer (10 mM Tris-acetate, pH 8.2, 10 mM Mg-acetate,
60 mM K-acetate, 0.05 mM DTT, 0.1 mM EDTA, 130 pg/ml
lysozyme, 23 ug/ml PMSF). This was blended slowly and
allowed to sit for 20 minutes. Then 10 ml of 4% Na+
deoxycholate were slowly blended in and the solution was
again left for 20 minutes. The solution became quite
viscous at this point. Approximately 2 mg each of RNase
and DNase were added with stirring and the solution was
left overnight at 4°C. (One preparation performed without
adding RNase yielded results similar to those in which
RNase was used.) This crude extract was centrifuged at
16,000 x g for 2 hours. The supernatant was saved and
dialyzed for 3 days vs. six 10 1 Changes of buffer A
(10.0 mM K phosphate, pH 7.7, 0.05 mM DTT, 0.1 mM EDTA).

Batch separation with DEAE-Cellulose: The dialyzed
crude extract was batch treated with DE-SZ. Five hundred
milliliters of wet-packed DE-SZ equilibrated with buffer A
were mixed with the dialyzed crude extract. This was
slowly filtered through Whatman 41 filter paper on a
Bfichner funnel. ‘The DE-SZ was washed with three 100 m1
rinses of buffer A. These rinses were added to the rest
of the flowthrough.

Phosphocellulose column: The DE-SZ flowthrough was
adjusted to pH 7.0 with 0.5 N acetic acid and loaded onto
a phosphocellulose column (2.5 x 35.0 cm) equilibrated

with buffer B (10 mM K phosphate, pH 7.0, 0.05 mM DTT,

19
0.1 mM EDTA). The column was washed with buffer B, which

contained 0.3 M KCl, until the effluent had an absorbance
reading less than 0.1 at 280 nm. CAP was then eluted
with a 1.0 liter linear gradient of buffer B + 0.3 M KCl
to buffer B + 1.0 M KCl. Fractions were collected and
monitored for A280’ conductivity, and CAMP binding activity.
Active fractions were pooled and the protein was precipi-
tated by adding 390 mg/ml (NH4)ZSO4 and centrifuging at
16,000 x g for 30 minutes. The precipitate was resuspended
in 6.0 ml buffer A + 0.1 M KCl and dialyzed overnight at
4°C vs. 1.0 liter of buffer A + 0.1 M KCl.

DNA-cellulose column: The above dialysis was stOpped
and any aggregated material was removed by a two minute
low speed spin in a Clinical centrifuge. The supernatant
was diluted 3-fold with buffer A and loaded onto a DNA-
cellulose column (0.9 x 13.5 cm) equilibrated with buffer
A. The DNA-cellulose was prepared according to the method
of Alberts and Herrick (1971) using native calf thymus DNA.
The column was washed with buffer A until the absorbance
was less than 0.05 at 280 nm. Most of the CAP passed
directly through the column. The fractions showing CAMP
binding activity were pooled. An (NH4)ZSO4 precipitation
was performed as before. The pellet was resuspended in
2.5 m1 of buffer A + 0.5 ml KCl and dialyzed overnight at
4°C vs. 1.0 liter of buffer A + 0.5 M KCl.

Sephadex G-75 column: The above dialysis was stopped
and the sample was loaded onto a Sephadex G-75 column

(2.5 x 90.0 cm) equilibrated with buffer A + 0.5 M KCl.

20
Elution was performed with the same buffer. Fractions
were assayed for A280 and CAMP binding activity. Active
fractions were pooled and precipitated as before. The
pellet was resuspended in 5.0 ml buffer C (10 mM TrisoHCl,
pH 7.9, 0.1 mM EDTA, 0.05 mM DTT, 0.1 M NaCl) and dialyzed
overnight at 4°C vs. 1.0 liter of buffer C. The dialyzed

sample was stored at -20°C in 500 pl aliquots.

Protein Measurement

 

Protein was measured by the method of Lowry et a1.
(1951) with crystalline bovine serum albumin used as a

reference standard.

Measurement of CAMP Binding to CAP

 

Binding of CAMP to CAP was measured by a modification
of the procedure of Anderson et a1. (1971). Each assay
consisted of 100 pl containing 10 mM 5'-AMP, 200 pg casein,

10 mM K phosphate, pH 7.7, 3

H-CAMP (200,000 cpm), and CAP.
This was incubated in an ice-water bath for 5 minutes.
Four hundred microliters of cold saturated (NH4)ZSO4 were
added and the sample was centrifuged at 8,000 x g for 10
minutes. The supernatant was removed and discarded. The
pellet was resuspended in 10.5 ml of scintillation fluid
(0.5% PPO in ethanolztoluene [1:3]) and counted. Back-
ground counts due to non-specific binding of CAMP to

other proteins were determined by repeating the assay with

20 mM unlabeled CAMP added to the reaction mix.

21

Gel Electrgphoresis

 

SDS polyacrylamide gel electrophoresis was performed
according to the procedure of Laemmli (1970). Twelve
percent gels were run. Crystallized trypsin was used as

a molecular weight standard.

Determination of Extinction Coefficients

 

The method for determining extinction coefficients was
based on a procedure by Edelhoch (1967). Tryptophan,
tyrosine, and cysteine are the only amino acid residues
which absorb ultraviolet light above 275 nm when the protein
is denatured in 6 M guanidine hydrochloride. Edelhoch
determined approximate extinction coefficients for each
of these amino acid residues under the above condition.
These are shown in Table II. If the number of each of

these residues is known for a protein, then the extinction

Table II. Extinction Coefficients (l/cm-mole) of Tryptophan,
Tyrosine, and Cysteine in 6 M Guanidine Hydro-
chloride, pH 6.5 (from Edelhoch, 1967)

 

A 280 A 288
Tryptophan 5690 4815
Tyrosine 1280 385

Cysteine 120 73

 

22
coefficients for that protein in 6.0 M guanidine hydro-

chloride, 20.0 mM phosphate, pH 6.5 at A 280 and A 288 are:

8280 = 5690 (# of trp) + 1280 (# of tyr) + 120 (# of cystine)

4815 (# of trp) + 385 (# of tyr) + 73 (# of cystine)
(1)

Anderson et a1. (1971) reported that CAP contains 4 trypto-

8288

phan, 10 tyrosine, and 4 cysteine (2 cystine) residues.
It follows from equation (1) that for CAP in 6.0 M guani-

dine hydrochloride:

2280 = 5690 (4) + 1280 (10) + 120 (2) = 35,800

5288 4815 (4) + 385 (10) + 73 (2) = 23,256 (2)

The concentration of CAP in 6.0 M guanidine hydro-
chloride can be determined by reading the absorbance of
the solution at A 280 or 288 nm. If this solution had
been made by adding concentrated guanidine hydrochloride
to a solution of CAP, the concentration of CAP in the
original solution could be determined by multiplying by
the dilution factor.

A solution of CAP was prepared and the absorbance
spectrum was read. This solution was then diluted with
concentrated guanidine hydrochloride to make the final
solution 6.0 M in guanidine hydrochloride. The absorbance
of this solution was read at A 280 and A 288. The concen-
tration of the diluted CAP was determined from each of
these readings. Multiplication by the dilution factor

yielded the concentration of CAP in the initial solution.

23
This taken together with the measured absorbance spectrum
allowed calculation of the extinction coefficients for the
native CAP.
All absorbance measurements were made on a Gilford
Model 250 spectrOphotometer with the slit width held

constant at 0.40 mm.

Measurement of CAP Binding to DNA

The binding of CAP to DNA was measured using the
sedimentation velocity method of Jensen and von Hippel
(1976). The method was adapted for use with the Beckman
Model B analytical ultracentrifuge equipped with UV optics
and a photoelectric scanner. In this way the contents of
a single cell could be examined several times during a
run. Solutions of CAP and DNA were dialyzed overnight
vs. the apprOpriate buffer and were mixed approximately
2 hours prior to each run. The sedimentation velocity
runs were performed at 22°C at speeds from 10,000 to

40,000 rpm.

Calibration of Model B Absorbance Readings

 

Absorbance readings in the Model B analytical ultra—
centrifuge were made with the slit fully open (2.0 mm).
Because of the wide spectral band pass, absorbance read-
ings in the Model E were different than those measured
on the Gilford. Correction factors to account for this
difference were determined by measuring the absorbance

of CAP and the absorbance of CTDNA first in the Gilford

24
spectrophotometer and then in the Model E. The ratios
of the absorbances (corrected for different path lengths)

were used as the correction factors.

Determination of Binding Site Size

 

The binding site size of CAP on DNA, i.e., the number
of base pairs covered by each molecule of bound CAP, was
determined by Circular dichroism (CD) measurements (Butler
et a1., 1977). Binding of CAP to DNA yields a CD spectrum
in the range 230-300 nm, which is not equal to the sum of
the spectra for CAP and DNA alone. Titrations of CAP with
concentrated CTDNA and of CTDNA with concentrated CAP were
performed at low ionic strength to insure complete binding.
CD spectra were measured on a Jasco spectrOpolarimeter

modified for CD by Sproul Scientific.

DNA Thermal Melting Experiments

Melting experiments were performed to investigate the
effect of CAP on the stability of double-stranded DNA.
Samples were prepared by mixing stock solutions of CAP and
DNA. Melts were performed on a Gilford Model 250 spectro-
photometer equipped with a thermal programmer. The tempera-
ture was increased linearly at a rate of 1°C/minute.

Absorbance readings were made at 260 nm.

Paper Chromatography_of CAMP

 

The purity of the CAMP was Checked using two different
solvent systems and descending paper Chromatography. The

procedure is that of Smith et a1. (1960). To Check for

25
CGMP contamination, 5.0 ul of 0.1 M CAMP were spotted on
Whatman No. 40 acid washed filter paper. Five microliters
of 0.1 M CGMP were also run as a control. The solvent
system was isobutyric acid-l M ammonium hydroxide-0.1 M
disodium EDTA (100:60:1.6). To Check the CAMP for degra-
dation to AMP, the solvent was Changed to isopropyl
alcohol-concentrated ammonia-water (7:1:2). Again 5.0 ul
of 0.1 M CAMP was spotted and 5.0 ul of 0.1 M 5'-AMP was
run as a control. The nucleotides were located by observing
fluorescence under a UV lamp. In both solvent systems only
a single spot which corresponded to CAMP was seen for the
CAMP track. No spots corresponding to CGMP or AMP were

seen in the CAMP sample.

RESULTS

Purification of CAP

 

The results of a typical purification are shown in
Table III. The final product appeared to be greater than
95% pure as judged from SDS polyacrylamide gel electro-
phoresis. The purification procedure was changed slightly
from that of Anderson et a1. (1971) because in my hands
their method yielded CAP which was only about 50% pure.
Their procedure included DEAE-Cellulose Chromatography,
phosphocellulose Chromatography, (NH4)ZSO4 precipitation,
and Sephadex G-100 chromatography. The addition of a
DNA-cellulose Chromatography step in my procedure improved
the final purity. As run, the CAP passed directly through
this column at a fairly low ionic strength while other
(presumably DNA-binding) proteins were retained. This
removal of other proteins which bind native DNA tightly
was obviously desirable since the CAP was used for DNA
binding experiments. Use of a high-salt Sephadex G-75
step following the DNA-cellulose column should separate
any DNA, which may have leached off of that column, from
the CAP. This sizing step also separated CAP from other
protein contaminants and resulted in a further 3- to 4-

fold increase in the purity.

26

27

.nflmmav .Hm pa xpzoq mo eczema ecu xn CCCHEHCHCC mm: :wouoemm

 

 

”we p.4N Rea x Ha.H m.o~ m.mN ms-o xppsnaom

NmH 9.0m AOH x ow.~ 0.0m o.wm . pmofisfifiao-<za

m.N4 5.5m Ros x se.e o.msm meN pmofisfifipaoeamoga

om.N HOH Boa x Hw.a oam.w ems Nm-ma

--- cos Rod x mm.s mmN.ON coo paehaxp pessu

CofimHhsm xua>ooom w Emu Hmuoe mnwev :Houopm maev oESHo> . :ofluowhm
CHom Hmpoe ngoe

 

m<u mo cofiumufiwwpsm .HHH oHan

28
Determination quExtinction Coefficients for CAP
Our quantitative experiments require that we know pre-
cisely the concentration of CAP in various solutions. This
concentration can be conveniently determined by absorbance.
Hence it is important that accurate extinction coefficients
for CAP are known. The measured extinction coefficients for

CAP are shown in Table IV.

Table IV. Extinction Coefficients for Native CAP

_———¢—_—_.—- -__,-—__.__.-__—______ —.-.—————--_—.—__-.——_._—.——u__——__——_-—._—--——

 

8
A (l/Cm-mole)
230 N 2.48 x 105
265 3.03 x 104
280 3.98 x 104

 

The accuracy of the procedure used here depends upon how
Closely the extinction coefficients of tryptophan, tyrosine,
and cysteine in the denatured CAP compare to those of the
model compounds (Table II). Edelhoch (1967) compared the
number of tryptophan and tyrosine residues calculated for
several proteins using this procedure with the number of
those amino acid residues actually present and the correla-
tion is quite good.

The procedure as I have used it also depends upon the
accuracy of the amino acid composition of CAP as determined
by Anderson et a1. (1971). It seems certain from their data
that there are 2 tryptophan and 2 cysteine residues per 22,500

molecular weight subunits. However, the analysis of tyrosine

29
gave a calculated result of 5.41 i 0.27 residues per sub-
unit. They Chose S as the correct number of residues per
subunit. If there are actually 6 tyrosine residues per
subunit, my final values for the extinction coefficients
would be lowered by only 2%.

The average value of 2.98 x 104 1/cm-mole which was
determined for the extinction coefficient at A 280 nm com-
pares well to a value of 3.5 x 104 l/Cm-mole which can be
calculated from the absorbance spectrum reported by Anderson
et a1. (1971) for a CAP solution of specified concentration.
The protein concentration had been determined by the Lowry

(1951) method. The value of 3.98 x 104

is 11% greater than
the 8280 for the denatured CAP. This is within the 0-20%
hyperchromic effect usually seen for native compared to
denatured proteins (Beaven and Holiday, 1952).

Extinction coefficients for native CAP can be evaluated
by Edelhoch's procedure using data either at A = 280 nm or
A = 288 nm. It is noteworthy that there was only a 5% dif-
ference between extinction coefficients based on the A280 vs.

A288 readings of the denatured CAP sample. The results in

Table IV are an average of the two sets of data.

DNA Thermal Melting Experiments
To test the hypothesis that CAP destabilizes the DNA
double-helix, melting experiments were performed. A low
ionic strength buffer (1.0 mM NaZHPO4, 0.1 mM NaZEDTA, pH
7.7) and p01y d(A-T) were used so that the DNA would melt

before the CAP became heat denatured. If CAP does

30
destabilize native DNA, then the poly d(A-T) should show a
lower melting temperature in the presence of CAP., Conversely,
a higher melting temperature would indicate that the protein
stabilizes the double helix. The poly d(A-T) alone melted
over a narrow temperature range centered at 26.8°C (Figure
2). When CAP was added to a ratio of 75 base pairs/CAP mole-
cule, there was a sharp partial melt at 27.5°C followed by a
gradual melt, and when the CAP was increased to a ratio of
13 base pairs/CAP there was a gradual melt centered at 47.5°C.
These results indicate that under the conditions used for
this experiment, the CAP stabilizes rather than destabilizes

the poly d(A-T).

Determination of Binding Site Size

 

The binding site size of CAP on DNA, i.e., the number of
base pairs which are physically covered when a molecule of
CAP binds to DNA, was determined by circular dichroism measure-
ments. Figure 3 shows that the sum of the spectra due to
CAP and DNA separately does not add up to the spectrum observed
when these are present together. Figures 4 and 5 show the
ellipticity readings obtained at A = 250 nm by titrating CAP
with native calf thymus DNA in the presence and absence of
CAMP, respectively. The early part of the titration in the
presence of CAMP has some scatter due to aggregation. This
disappears late in the titration when the CTDNA is in excess.
The break point in both of these figures occurs at about 12-13
base pairs CTDNA/CAP molecule. As a Check on this result, a

titration was done in the reverse order, i.e., DNA was

 

31

 

 

.msasa than Ae-<ve smog z m-c~ x sv.~ .a<u 2
o-o~ x om.~ no Nashua when Ae-<vp sfioa : m-o~ x mm.~ .a<u z s-eH x o.4 am
“manna ohms Ae-<vp saoa 2 m-c~ x ee.~ A< .s.s 2a .<enm~mz :5 H.o . oaz~mz
ze o.H cw d<u mafia AH-<VC zdom How mo>uso m:«u~vs awakens .N opsmwm

 

 

fi..mm:h<mmisuh
66 CV ON
_ _ q q .. .. J. .. .4.« l
u
3 m .
F I 4
4
<4 o
a <, o
I ‘ O O O
.444 l.
““““ I'll-A09
.1. 41.469

.l. as

me.

OH.

mu.

 

 

aewvuo aoNvauosav

 

33

.ocon <za mafia

anode a<u mo ssh ass ma ---- .s.s za .een :5 mo.o .<enm~sz :5 H.o .Husz

:5 o.mH .eoazmmz :5 o.m ch <zaeu was a<u Co asaupam cu

.m ohamflm

 

 

 

7,‘

Q

C

O

‘3

<

z

a
Q)
C
.3
N]
of”
<
o

 

34

 

Figure 3

DNA/CAP
mixture'

 
 

296

 

 

(sigun Mengqie)

AllOlldl'l'lE

270

WAVELENGTH, n m

250

 

3S

m<zceu mafia a<o 2 o-o~ x H.H A<

.pcofis <zoeu mm
.s.s 2a .a2<a m-o~ x H .eea 2e mo.c .<enmmmz

2e H.o .fiumz 22 o.mH .eoa2st 2a o.m ma shaman pee .a2<a Co oucomoha one an
<zaeu 2a“: a<u meaumsafia 22 epehsano ezsaaaam no pucphpmmfin .4 shaman

cm

.33 2c:—
mbma mama <20 no.3:

me. am

 

36

 

Figure 4

 

_ _

In
I

o
‘I‘

diltasAm—v
.11.. A5

1
m
'1‘

 

wu osz=\< ‘(seeJfiapuum AllOlldl'l'la

 

37

 

 

.pcoHe <zceo Am m<zneu mafia a<u 2 c-o~ x ~.H A<
.m opsmwm how :30:m page we 65mm one mg Mommas och .a2<u we oocomnm any :«
<zaeu new: m<o mamumguwu >3 cocwmuno sapwooam no oocohommHo .m ouswwm

53 «.3:
3.22. 0mm: <29 mo_oE

 

om mv on ma
4 a _ _
4 o
. .\
.1 .\
arm. . \\

 

 

<I|< Amy
. II. A<v

m
I

O
‘1'
we oszr-v ‘(seeafiepuum Ailolidnu

 

 

39
titrated by addition of CAP in the absence of CAMP (Figure
6). The ratio of base pairs of CTDNA/CAP molecule is again
about 13 at the break point, consistent with the other two
titrations.

The fact that the binding of CAP to CTDNA results in a
non zero Circular dichroism difference spectrum is indicative
of a conformational change in either or both the CAP and the
CTDNA. Since this difference is apparent in the region from
A = 250 to A = 300 nm in which CAP does not have a signifi-
cant CD spectrum, it is tempting to state that the CTDNA
undergoes a conformational Change when it binds CAP. Also,
no Change in the CD spectrum occurs over the range A = 210-230
when CAP is titrated with CTDNA both in the presence and
absence of CAMP (data not shown). CAP alone has a large CD
reading at low wavelengths, whereas CTDNA has a relatively
small reading. These results taken together, i.e., a differ-
ence occurring in the region in which CTDNA but not CAP has a
large CD spectrum but no difference seen where CAP has a large
CD spectrum while CTDNA does not, suggest that the CTDNA
undergoes a conformational Change upon binding CAP. Interpre-
tation of CD spectra in terms of A-form or B-form DNA struc-
tures, etC., is very difficult. An unambiguous statement
about the specific conformational Changes which occur cannot

be made without a good deal of additional effort.

Binding of CAP to DNA

 

As a preliminary to studying the binding of CAP to

DNA, control experiments using CAP alone were performed

4O

 

 

.cofiumpufiu one mcwusc mazuoo sown: sawusawc can on can ma ammopoaw

asmaflm age .5: omN u 2 as sphaaaaafifip <zaeu escapaxm Am .=owaappm gasp
pouwm acme mcwon :mom no a .Cocvc who: m<u ecumeacoocou we mcofiuwcum a: om
Co>om .COHHSHOm <zoeo 2 m-oH x HH.~ a mo H: ooNN nu“: :swon mm: CCMHMMuMu
age .<zceu mafia a<u A< .m panama pom Hana he 828m 828 he shaman pee
.m<u\mpflmm omen can» hogan» swam ommn\m<u ohm mafia: mmmwomnm any umcu ouoz

.m<u new: <znbu wcfiumuufiu 2n Cocflmuno Espuomam no monopowwfia .c ohsmfim

 

mtg. omen <20 East. .
. L<o 3:9:

 

41 Cl

 

Figure 6

 

um oszzt ‘(seelfiapmsm Annual-I13

 

 

42
to see if there is any protein-protein interaction in
the absence of DNA. Table V shows that there is only
a slight Change in the sedimentation coefficient over
the concentration range of CAP tested both in the presence
and absence of CAMP. This implies that there is essen-
tially no interaction occurring between CAP molecules
under these conditions.

The binding of CAP to DNA was studied using a sedi-
mentation velocity technique. The CAP and DNA were mixed
in an appropriate buffer to give a homogeneous solution
before centrifugation was begun. During the centrifuga-
tion run a boundary develops as the molecules sediment.
The Z-phase boundary seen in Figure 7A indicates the
presence of two molecular species having different sedi-
mentation coefficients. Both of these species are present
in the bottom part of the cell where the absorbance is
equal to the sum of the absorbances for each species
taken separately. However, the more rapidly sedimenting
material is no longer present in the top portion of the
cell. Here the absorbance is due to only the more slowly
sedimenting molecules. The exact composition of the
material at any point in the cell was determined by
measuring the absorbance at both 230 and 265 nm and

solving equations 4 and 5 simultaneously;

A230 = sgég (b)[CAP] + ngé (b)[DNA] (4)
A265 = 632? (b)[CAP] + egg? (b)[DNA] (5)

where b is the path length of the cell.

43

 

 

 

 

Table V. Relationship Between [CAP] and Sedimentation
Coefficient
No CAMP 1 x 10‘4 M CAMP
(CAP) 5 [CAPTI s
-7 -7
3.5 x 10 2.4 5.1 x 10 3.1
4.5 x 10‘7 2.9 9.0 x 10‘7 3.3
1.3 x 10‘6 3.4 1.6 x 10‘6 3.4
3.05 x 10'6 3.7 3.2 x 10'6 3.4

 

CAP was present in a 5.0 mM NazHPO4, 15.0 mM NaCl, 0.1 mM
NazEDTA, and 0.05 mM DTT, pH 7.7 buffer. Sedimentation
was performed at 40,000 rpm and 22°C.

5

It was found that in the presence of 5 x 10- M

CAMP, CAP will bind to native CTDNA. The binding is

ionic strength dependent, being very tight at ionic

strengths less than 75 mM Na+ and virtually disappearing

at 125 mM

Na+. At very low ionic strengths there is no

free protein (as long as the DNA is in excess) and only

a single boundary is seen in the sedimentation run. The

only species present is CAP-DNA complexes as determined

by absorbance measurements and the sedimentation coef-

ficient.

The absence of free CAP implies that every CAP

molecule is active in binding to DNA. At very high ionic

strengths no binding occurs; a two-phase boundary is seen

consisting of a relatively rapid section due to free DNA

and a slow section due to free CAP. At ionic strengths

between 75 and 125 mM Na+ the amount of binding is inter-

mediate and a two-phase boundary is seen during the

44

 

 

.Coomm mcwcomoh poumm mou::we

cm :oxmu ope: mcmum .UoNN u 9 .59h occ.o~ mm; coomm pouom .5: mom u 4

as seem an me: cm~ u 2 as seam A< .2082 2a o.oo~ .a2<6 2s mo.o .<eom~mz

:5 H.o .0.“ In .Hu:.mwph :5 o.o~ :w <znhu z m-oH x e.m mafia m<u 2 cued
x o.~ mo newumucosfiwom mo :mom amzmfipaccochufis m Hope: .5 opnmwm

45

Figure 7

 

 

 

n22<o +5.: moNH/m Am.

 

 

 

 

m5<o .T .E: OMNHK 33

46

centrifugation (Figure 7). The rapidly sedimenting
material is a CAP-CTDNA complex, and the slowly sedi-
menting material is free CAP determined from the absorb-
ance readings and sedimentation coefficients. There is
no section of the boundary which corre5ponds to free DNA
molecules because the "free" DNA in this system consists
of the gaps along the DNA molecules where no protein is
bound. Because no DNA totally free from CAP is seen,
these sedimentation results imply that all of the DNA
molecules have some CAP bound to them.

A fairly extensive study of the binding of CAP to
DNA in the absence of CAMP was performed. As with CAMP
the binding was very ionic strength dependent. However,
there are two major differences when CAMP is not present.
One is that the overall binding is weaker. No binding
was observed when the [Na+] was above 80 mM, and binding
was too tight to observe any free CAP only at ionic
strengths below 40 mM Na+. This range is comparable to
the 75-125 mM Na+ range when CAMP is present. The second
major difference is that a 3-phase rather than a Z-phase
boundary is observed when 50-80 mM Na+ is present (Figure
8). It was found that this boundary corresponds to
rapidly sedimenting CAP-DNA complexes, free DNA molecules,
and slowly sedimenting free CAP. The fact that there are
DNA molecules which are completely devoid of protein is
indicative of cooperative binding of the CAP to the CTDNA.
The cooperativity causes CAP to bind in Clusters so that

some of the DNA molecules contain many CAP molecules and

47

.Coon magnomou

nouwm mouzcwe we :oxmu one: mcmum .UoNN u e .592 coc.o~ mm: coomm Heuom
.2: moN u 2 as show 2m we: cmN u 2 68 seem 2< .2082 22 m.sc .<eam~mz

2e 2.o .m.s :2 .2u=.m2se 2e o.o2 :2 <zceu 2 m-c2 x m.~ h=2a a<u 2 e-O2
x w.H mo :ofiumucoswcom mo Cmum omzmwpucoom»u~: m Hove: .m ouzmfim

48

 

 

 

   

.2526 2. 6:. menu/m 8..

R

 

 

 

 

 

n22<o o: .E: OMNNK A<v

49
the rest of the DNA is essentially free of CAP. The ratio

of DNA to CAP for the rapidly sedimenting material varied
depending on the total amounts of DNA and CAP in the solu-
tion but approached a limiting value of about 13 base
pairs per molecule of CAP at low DNA/protein input ratios.
Scatchard plots for the data obtained for the experiments
in 50-80 mM Na+ in the absence of CAMP are shown in
Figures 9 through 12. The theoretical equation derived
by McGhee and von Hippel (1974) for this cooperative

binding is given by

3 = . _ . (2w-l)(l-nv)+v-R n-l. 1-(n+l)v+R 2
L K (1 nu) ( 2(m-I)(l-nv) ) ( 2(l-nv) ) (6)
The parameters v, L, K, n, and m were defined previously

(see literature review). R is given by

R = {[l-(n+l)\)]2 + 401\)(l-n\))}1/2 (7)

The data shown in Figures 9 through 12 were fitted to this
equation by setting n = 13 and using a nonlinear least
squares program to determine best values of K and 6.

These are given in Table VI. The values of v and v/L shown
in Figures 9 through 12 were determined from the total
concentrations of CAP and CTDNA throughout the cell before
centrifugation was begun (determined by the two-wavelength
analysis previously described) and from the concentration
of free CAP. This concentration of free ligand was
determined as follows. Because CAP sedimented much more

slowly than both the CAP-CTDNA complexes and also free

50

Figure 9. Scatchard plot for the cooperative bind-
ing of CAP to CTDNA at 22°C in 10 mM Tris-HCl, 0.1 mM
NazEDTA, 49.8 mM NaCl, pH 7.9.

 

 

51

Figure 9

 

20——

37/1. x 10”4

pa
0

 

r,

 

 

 

52

Figure 10. Scatchard plot for the cooperative bind-
ing of CAP to CTDNA at 22°C in 10 mM Tris-HCl, 0.1 mM
NazEDTA, 59.8 mM NaCl, pH 7.9.

 

 

V/L' x 10"4

pa
0

(II

53

Figure 10

 

 

 

 

 

54

Figure 11. Scatchard plot for the cooperative bind-
ing of CAP to CTDNA at 22°C in 10 mM Tris-HCl, 0.1 mM
NazEDTA, 69.8 mM NaCl, pH 7.9.

V/L x 10’4

 

 

'l

Figure 11

SS

 

.02

.04

<l

.06

 

56

Figure 12. Scatchard plot for the cooperative bind-
ing of CAP to CTDNA at 22°C in 10 mM Tris.HC1, 0.1 mM
NazEDTA, 79.8 mM NaCl, pH 7.9.

V/L x 10'4

pa

 

 

57

Figure 12

I

 

.02

 

<l

 

58

Table VI. Effect of Ionic Strength on Binding Parameters

 

 

[Na+] (mM) K (l/Cm-mole) w K-w (l/Cm-mole)
50 5.96 x 104 70 4.17 x 10°
60 1.98 x 104 110 2.18 x 10°
70 6.51 x 103 175 1.14 x 10°
80 7.29 x 103 60 4.37 x 105

 

CTDNA, it was a simple matter to completely sediment
everything except for the free CAP, measure its absorbance,
and determine its concentration. To evaluate v, the
amount of bound CAP was taken as the difference between
total and free CAP.

In an attempt to extend the results for cooperative
binding, some additional experiments were performed. To
Check whether there is any base specificity for the bind-
ing, CTDNA was replaced by poly d(A-T) and by poly d(I-C).
Enough data to make Scatchard plots were not obtained,
but CAP appears to bind equally well to poly d(A-T),
poly d(I-C), and CTDNA. The final experiments to be
reported involved the effects of Mg++ on the binding.

4.

At low levels of Mg+ (3-10 mM) cooperative binding

was still present whereas the presence of 25 mM Mg++
plus 15 mM Na+ was enough to prevent all CAMP independent
binding from occurring. Thus the effects of Mg++ seem

to arise primarily from its contribution to the ionic

strength; no specific effects are apparent.

DISCUSSION

The binding site size is an important parameter in
Characterizing the CAP-DNA interaction. It was interest-
ing to see whether the site size was affected by the
presence of CAMP. The results indicate that CAP covers
approximately 13 base pairs whether binding is coopera-
tive or nonCOOperative. Changes in CAP certainly occur
when it binds CAMP. This was shown by Wu et a1. (1974)
using their fluorescent probe and by Krakow and Pastan
(1973), who noticed differences when CAP was incubated
with proteases in the presence and absence of CAMP, and
it is also obvious as seen by the Change from cooperative
to noncooperative binding of CAP to DNA when CAMP is
added. However, since the binding site size is unchanged
by the addition of CAMP, the change in the CAP conforma-
tion is probably a subtle one rather than any gross
difference. It also implies that the orientation of
CAP on the DNA may be similar for both types of binding.

The binding site size of 13 base pairs is reasonable
considering the size of CAP. Ovalbumin, which has the
same molecular weight and frictional coefficient as
does CAP, has a molecular radius of 27.6 A as determined
by its diffusion coefficient (Tanford, 1961). If CAP is

O
assumed to have a diameter of about 55 A by comparison,

59

60
the fact that it covers only 44 A (13 base pairs) on

DNA can be explained by assuming slight overlap or from

the fact that CAP is not perfectly spherical.

Binding of CAP to DNA

 

Two important results concerning the binding of CAP

to DNA have been obtained. One is that, contrary to most
reports, there is CAMP independent binding of CAP to
DNA at a physiological pH (7.9). Second, this binding is
highly cooperative as opposed to the nonCOOperative or
only slightly cooperative binding which occurs in the
presence of CAMP. The only previous report of CAMP-
independent binding of CAP to DNA was that of Riggs et
a1. (1971). This was later retracted, the results being
attributed to impure preparations of CAP (Nissley et a1.,
1972). I do not attribute the results presented here to
impure preparations of CAP. Several preparations of CAP
gave similar results and SDS-polyacrylamide 9915 showed
that these preparations were greater than 95% pure.
It seems likely that other workers have overlooked this
CAMP-independent binding due to their experimental con-
ditions. As I have shown, the CAMP-independent binding
is very ionic strength dependent, being undetectable by
my technique when the Na+ concentration was above 80 mM.
Nissley et a1. (1972) used ionic strengths greater than
this for both their sedimentation and filter assays.

Riggs et a1. (1971) used low ionic strength buffers in

their report of CAMP-independent binding, and it is not

61
Clear why they later observed no binding. It is the
report of Nissley et al. (1972) and the retraction by
Riggs et al. (1971) that are commonly cited for the obser-
vation that only CAMP-dependent binding occurs.

The highly cooperative binding is an interesting
effect. As mentioned previously, a measure of the coopera-
tivity parameter, m, has been determined for only a few
protein-DNA systems. The binding parameters K and w for
the CAP—CTDNA system are shown in Table VI. From the
data at 50—70 mM Na+ it appears that the intrinsic binding
constant K decreases about 3-fold for each 10 mM increase
in the Na+ concentration. Also, there is a corresponding
50-60% increase in w. This pattern does not hold, however,
when the data for 80 mM Na+ are included. It must be
noted that the nonlinear least squares program used to
fit the data is somewhat insensitive to Changes in K and
w. A visual fit indicates that any value of w from 50
to 200 is reasonable for all four of the ionic conditions.
This is due to the fact that the regions most sensitive
to Changes in K and w are at very low and very high values
of v where it was difficult to obtain accurate data.
However, for any fit at a single ionic strength the Changes
in K and w are almost inversely proportional. Therefore,
at a given ionic strength the overall binding constant
K-w remains relatively constant no matter what values of
K and w are Chosen to fit the data. A plot of log(K-w)

vs. log[Na+] is shown in Figure 13.

62

Figure 13. Binding constant as a function of
[Na+] for CAP binding to CTDNA at 22°C in 10 mM Tris-HCl,
0.1 mM NazEDTA, plus NaCl.

 

LOG K-LU

42

 

 

1.3

Figure 13

 

l ' l
1.2 1.1

-LOG [Nafl

1.0

 

64

In viva Implications

 

It is interesting to consider what fraction of CAP
may be bound to DNA in viva. A linear extrapolation of

the results in Figure 13 shows that the overall binding

4

constant K-w is approximately 5 x 10 l/m at an ionic

strength of 130 mM Na+ (considered to approximate ”physio-

logical” conditions). The in viva concentration of CAP

is about 2 x 10'6

3

M (Anderson et a1., 1971) and DNA is
about 6.6 x 10- M in base pairs. If it is assumed that
only half of the DNA in the cell is free and that K = 500
and w = 100, then a calculation shows that about 64% of

the CAP will be bound to the DNA even in the absence of
cAMP. This indicates that some fraction of the CAP is
probably bound in viva, although the actual amount is
uncertain.

The presence of CAMP in the cell will cause noticeable
Changes. Because the nonspecific binding is tighter in
the presence of.CAMP, more CAP will become bound to the
DNA. Also, the previous Clustering of CAP on the DNA due
to the cooperativity will be mostly eliminated resulting
in a more even distribution.

The above has been limited to the binding of CAP to
nonSpecifiC DNA sites and has not included any possible
effects due to Mg++ nor binding of CAP to catabolite
sensitive promoters. However, the in viva concentration
of Mg++ is probably on the order of a few millimolar.

This concentration of Mg++ showed no specific effects upon

binding and it is negligible relative to the 130 mM Na+

65
concentration assumed to approximate in viva ionic condi-
tions. It would be desirable to study this specific
binding of CAP to promoter DNA, but this has proved
difficult as other workers have shown. Both Riggs et al.
(1971) and Nissley et a1. (1972) were unable to show any
specific binding of CAP to lac promoter. Majors (1975)
has reported tighter binding of CAP to lac but his results
depended on very exacting conditions. Apparently there
is specific binding of CAP to catabolite sensitive operons
but it is only slightly stronger than is the nonspecific
binding. The results obtained here deal with nonspecific
CAP-DNA interactions and thus do not directly provide
information about the CAP- or CAP-RNA polymerase-DNA
interactions. However, the results are suggestive as to
what these interactions may or may not be. A generally
pr0posed model is that CAP binds at the promoter and by
interacting with DNA and/or with RNA polymerase facilitates
initiation (If transcription. For example, CAP could
somehow help RNA polymerase to ”melt" into the promoter
region. The CD results show that binding of CAP to DNA
Changes the conformation of the DNA but does not give
a detectable Change in the protein conformation. However,
this conformational Change appears to be the same both
in the presence and absence of CAMP. Since transcription
at catabolite sensitive operons is stimulated only when
CAMP is present, this conformational Change may be unin-
volved in promoting this transcription. The results of

the thermal melting experiments do not indicate that CAP

66

is a DNA melting protein. These results are certainly
not conclusive but provide evidence that the hypothesis
that CAP destabilizes the DNA allowing RNA polymerase to
begin transcription may be incorrect. Because these two
sets of results do not indicate any specific Changes in
the DNA which may occur to promote transcription, the
specificity may reside in the proteins. Cooperative
binding induced by DNA has been observed, and it is
tempting to believe that it has an in viva role. Although
this cooperativity seems to disappear in the presence of
CAMP, possibly the specific binding of CAP to promoter DNA
will revive this effect. This might favor binding addi-
tional CAP molecules to the promoter in the presence of
CAMP and a direct interaction of a small Cluster of CAP
molecules with the nearby RNA polymerase might result.
Evidence of the true mechanism will probably have to
await experimentation with the promoter region itself.

In summary, the results which have been presented
add to the picture of the interaction of CAP with DNA.
In particular, it has been shown that nonspecific c00pera-
tive binding of CAP to DNA probably occurs in viva in the
absence of CAMP and also that the presence of CAMP both
increases the amount of binding and changes it from a
cooperative to a nonCOOperative system. These results
lead to some interesting speculation about the in viva

mechanism by which CAP stimulates transcription.

LIST OF REFERENCES

LIST OF REFERENCES

Alberts, B., Amodio, P., Jenkins, M., Gutmann, E., and
Ferris, F. (1968) Cold Spring Harbor Symp.
Quant. Biol. 33, 289.

Alberts, B. and Frey, L. (1970) Nature 227, 1313.

Alberts, B., Frey, L., and Delius, H. (1972) J. Mal.
Biol. 68, 139.

Alberts, B. and Herrick, G. (1971) In Methods in
Enzymology, eds. Grossman, L. and Moldave, K.
(Academic Press, New York), Vol. XXI, p. 198.

Anderson, W. B., Perlman, R. L., and Pastan, I. (1972)
J. Biol. Chem. 247, 2717.

Anderson, W. B., Schneider, A. B., Emmer, M., Perlman, R. L.,
and Pastan, I. (1971) J. Biol. Chem. 246, 5929.

Beaven, G. H., and Holiday, E. R. (1952) Adv. Protein
Chem. 7, 319.

Beckwith, J. R., and Zipser, D. (1970) "The Lactose
Operon", Cold Spring Harbor Laboratory.

Butler, A. P., Revzin, A., and von Hippel, P. H. (1977)
Biochemistry 16, 4757.

Carroll, R., Neet, K., and Goldthwait, D. (1975) J.
Mal. Biol. 91, 275.

deCrombrugghe, B., Chen, B., Anderson, W., Nissley, P.,
Gottesman, M., Pastan, I., and Perlman, R. (1971)
Nature New Biology 231, 139.

Delius, H., Mantell, N. J., and Alberts, B. (1972) J.
Mol. Biol. 67, 341.

Dennert, F. (1900) Ann. Inst. Pasteur 14, 139.

Draper, D. E., and von Hippel, P. H. (1978) submitted
for publication.

Edelhoch, H. (1967) Biochemistry 6, 1948.

67

 

68
Eilen, E., and Krakow, J. S. (1977) J. Mol. Biol. 114, 47.

Emmer, M., deCrombrugghe, B., Pastan, I., and Perlman, R.
(1970). Proc. Natl. Acad. Sci. 0.5.4. 66, 480.

Epps, H. M. R., and Gale, E. F. (1942) Biochem. J. 36,
619.

Epstein, R., Bolle, A., Steinberg, C., Kellenberger, E.,
Boy de la Tour, E., Chevalley, R., Edgar, R.,
Susman, M., Denhardt, G., and Lielausis, A. (1963)
Cold Spring Harbor Symp. Quant. Biol. 28, 375.

Eron, L., Arditti, R., Zubay, G., Connaway, S., and
Beckwith, J. R. (1971) Proc. Natl. Acad. Sci.
U.S.A. 68, 215.

Jensen, D. E., Kelly, R. C., and von Hippel, P. H. (1976)
J. Biol. Chem. 251, 7215.

Jensen, D. E., and von Hippel, P. H. (1976) Chemistry
251, 7198.

Kao-Huang, Y., Revzin, A., Butler, A. P., O'Conner, P.,
Noble, D. W., and von Hippel, P. H. (1977) Proc.
Natl. Acad. Sci. U.S.A. 74, 4228.

Kelly, R. C., Jensen, D. E., and von Hippel, P. H. (1976)
J. Biol. Chem. 251, 7240.

Krakow, J. S., and Pastan, I. (1973) Proc. Natl. Acad.
Sci. U.S.A. 70, 2529.

Laemmli, U. K. (1970) Nature 227, 680.

Lowry, 0., Rosebrough, N., Farr, A., and Randall, R.
(1951) J. Biol. Chem. 193, 265.

Magasanik, B. (1970) In The Lactose Operon, eds.
Beckwith, J. R., and Zipser, D. (Cold Spring
Harbor Laboratory), p. 189.

Majors, J. (1975) Nature 256, 672.

Makman, R. S., and Sutherland, E. W. (1965) J. Biol.
Chem. 240, 1309.

McGhee, J. D., and von Hippel, P. H. (1974) J. Mol.
Biol. 86, 469.

Molineux, I. J., Friedman, S., and Gefter, M. L. (1974)
J. Biol. Chem. 249, 6090.

Nakanishi, S., Adhya, S., Gottesman, M., and Pastan, I.
(1975) J. Biol. Chem. 250, 8202.

69

Nissley, P., Anderson, W. B., Gallo, M., Pastan, I., and
Perlman, R. L. (1972) J. Biol. Chem. 247, 4264.

Nissley, S. P., Anderson, W. B., Gottesman, M. E., Perlman,
R. L., and Pastan, I. (1971) J. Biol. Chem. 246,
4671.

Perlman, R. L., and Pastan, I. (1968) J. Biol. Chem. 243,
5420.

Pratt, D., Laws, P., and Griffith, J. (1974) J. Mol.
Biol. 82, 425.

Riggs, A. D., Reiness, G., and Zubay, G. (1971) Proc.
Natl. Acad. Sci. U.S.A. 68, 1222.

Ruyechan, W. T., and Wetmur, J. G. (1975) Biochemistry
14, 5529.

Scatchard, G. (1949) Ann. N.Y. Acad. Sci. 51, 660.

Smith, M., Drummond, G. I., and Khorana, H. G. (1960)
J. Amer. Chem. Soc. 83, 698.

Tanford, C. (1961) Physical Chemistry of Macromolecules
(John Wiley 8 Sons, Inc., New York), pp. 358-359.

Tomizawa, J., Anraku, N., and Iwama, Y. (1966) J. Mol.
Biol. 21, 247.

Ullman, A., and Monod, J. (1968) FEBS Letters 2, 57.

Williams, K. R., and Konigsberg, W. (1978) J. Biol. Chem.
253, 2463.

Wu, F. Y., Nath, K., and Wu, C. (1974) Biochemistry 13,
2567.

Wu, C., and Wu, F. Y. (1974) Biochemistry 13, 2573.
Wu, J., and Yeh, Y. (1973) J. Virol. 12, 758.

Zubay, G., Schwartz, D., and Beckwith, J. (1970) Proc.
Natl. Acad. Sci. U.S.A. 66, 104.

 

lllllllllllljlllllllllllljlflllllllllllllllllll

9 3678