li‘lllllllllllfllll

93 00835 8412

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Milling!

3

This is to certify that the

thesis entitled

The Nature of DNA—Protein
Interactions Studied by Polyacrylamide Gel
Electrophoresis

presented by

John Anthony Ceglarek

has been accepted towards fulfillment
of the requirements for

MS degree in Biochemistry

GIL/“(Kg Gab/axm

Major professor

Date May 15, 1987

O~7639 MSUis rm AWN-"film ‘ ‘ "1 '“y, .. ', Institution

-AA.‘

 

 

PViESIcJ RETURNING MATERIALS:
Place in book drop to
LIBRARIES remove this checkout from
w your record. FINES will

 

 

 

 

be charged if book is
returned after the date
stamped below.

 

nor _2 3t 1995

.gpg,‘ z :1, «not;

.mgzifia «cm

”biflicl‘fooo

mm: min

 

 

 

THE NATURE OF DNA-PROTEIN
INTERACTIONS STUDIED BY POLYACRYLAMIDE GEL
ELECTROPHORESIS

By

John Anthony Ceglarek

A THESIS
Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of
MASTER OF SCIENCE

Department of Biochemistry

1987

 

ABSTRACT

THE NATURE OF DNA-PROTEIN
INTERACTIONS STUDIED BY POLYACRYLAMIDE GEL
ELECTROPHORESIS

By

John Anthony Ceglarek

Polyacrylamide gel electrophoresis is a useful tool for
isolating and studying DNA-protein complexes. Here, properties of
DNA-protein complexes in a gel are compared with those of
complexes in solution with respect to biological function and
thermodynamic parameters. It was found that Escherichia coli RNA
polymerase-lag UVS promoter complexes yield the same transcript
whether in solution or in the gel. In addition, catabolite
activator protein (CAP)-wild type lag promoter complexes display
the same dissociation constant in the gel as they do in solution.
Therefore, complexes isolated by polyacrylamide gel
electrophoresis are functionally the same as those found in
solution. Experiments were conducted to determine how DNA-protein
complexes traverse the gel. The data support end-on migration.

It may also be possible to use DNA fragments cloned in the course
of this work to determine whether RNA polymerase locates promoters

by sliding along nonspecific tracts of DNA leading to the promoter

region.

 

To my wife, Lisa, and my mother, Frances, whose sacrifices
allowed me to be here; and to my father, Victor, and my grandmother,

Anne, who I wish could be here to share this with me.

ACKNOWLEDGMENTS

First of all, I wish to thank Dr. Arnold Revzin for his
continuous support over the last two years, and for his patience and
guidance. Thanks are also due to Dr. Stephanie Shanblatt and fellow
students Donald Lorimer and Jianli Cao for their friendship and
useful suggestions, and to Lloyd LeCureux and Diane Cryderman for
their excellent technical assistance. Finally, I wish to thank

Vicki McPharlin for her help in preparing this manuscript.

iv

 

TABLE OF CONTENTS

Page
List of Tables................... ..... ............ ......... ...vii
List of Figures .............................................. viii
Introduction.. ..... . ..... . ...... ... ...... .......................1
Chapter I. Comparison of Nucleic Acid-Protein Interactions
in Solution and in Polyacrylamide Gels ....... . ...... 3
Materials and Methods ...................... . .............. .7
Materials........... ..... . ....... . .................... 7
Methods ............................................... 8
DNA Fragment Preparation ......................... 8
In-Gel Transcription Reactions ................... 9
Solution Transcription Reactions ..... ..... ...... .9
Sizing of Transcripts ......................... ..10
Determination of Dissociation Constants
in the Gel............. ....... ..................10
Determination of Dissociation Constants
in Solution............. ....... .... ...... .......11
Results and Discussion ............... . .................... 13
Comparison of Transcripts Made in the Presence
or Absence of the Gel Matrix.............. ...... .....13
Effects of the Gel Matrix on CAP-DNA
Dissociation Constants ......................... . ..... 16

 

Chapter II. DNA Fragments Containing Multiple Protein Binding
Sites for Studying the Movement of DNA-Protein

Complexes in Polyacrylamide Gels .................. 21
Materials and Methods .......................... . ...... ....27
Materials...................................... ...... 27
Methods... ...... .. ..... ..............................27

Construction of Fragments.........................27
Construction of lac 211 bp fragment containing

the UVS promoter mutation and the wild type

CAP Site-000.000.00.000 ..... o. oooooooooooooooooooo 27

Preparation of "nonpromoter DNA" .................. 28

Construction of the promoter—containing

fragments P.F. 3 and A............... ........ .....32
Construction of fragments A, B, C, and D.. ........ 35
Construction of 1331 bp fragment ............ ......36

DNA Fragment Preparation..........................36

Electrophoresis of Various DNA Fragments

and Complexes Thereof. .............. . ........ .....36
Transcription Reactions .............. ...... ....... A1
Results and Discussion..................... ........ .......N2

Electrophoretic Mobilities of CAP or RNA
Polymerase Complexed with Fragments A, B, C, and D...A2

Bending of the DNA by RNA Polymerase.................A9

Mechanism of Promoter Search by E. coli
RNA Polymerase .................... ..... ..... . ........ 50

Conclusions....................................................56

List of References ........ ........... ........ . ......... . ..... ..59

Vi

LIST OF TABLES

Page
Dissociation constants of cAMP—CAP-[aszlac promoter
complexes..................................................17
Relative mobilities on A% polyacrylamide gels of
CAP-A, B, C, D Complexes, RNA polymerase-A, B, C, D
complexes and CAP‘RNA polymerase-A, B complexes............48

10.

11.

12.

13.

LIST OF FIGURES

Page
Autoradiograph of RNA made in a polyacrylamide gel
after the electrophoresis of RNA polymerase-lac
UV5 promoter complexes... ...... ................... ........ 1“
Comparison of transcripts made in a polyacrylamide
gel with those made in solution............. ........ . ..... 15
Dissociation of CAP-wild type lac promoter complexes
during electrophoresis in a polyacrylamide gel .......... ..19
Diagram of fragments A-D .................................. 2%

Partial map of the lac operon control region..............30

Flow diagram of the construction of plasmids
containing P.F01and 2.0.0.000...0.00.0000000.000.000.00031

Diagram of DNA fragments used in fragment A-D
constructions......................... .......... ..........3A

Restriction digests of recombinant plasmids

containing fragments A-D... ............................... 38
Diagram of the 1331 bp construct........... ..... ..........AO
Mobilities of CAP complexes with fragments A-D ............ A3

Mobilities of RNA polymerase complexes with
fragments A—DOOOOOOOOOOOOOOOOOOOOOOOOO0.000.... ..... O ..... I46

Mobilities of fragments A and B complexed with
CAP and RNA pOlymeraseOOOOOOOOOO..OOOOOIOOOOOOOI00.0.00000u7

Mobilities of the N93 bp and H75 bp fragments
complexed with CAP and RNA polymerase... ...... ............51

viii

INTRODUCTION

Transcription, the transfer of genetic information from
double-stranded DNA into RNA, is the first step in the pathways
leading to protein synthesis as well as structural and adapter RNA
production. As such, it is a very tightly regulated process; one
can imagine that having too many copies of certain enzymes could
lead to many complications for the cell, perhaps culminating in cell
death. Alternatively, too few molecules of specific proteins
(metabolic enzymes, for example) can render the cell (and possibly
the entire organism) unable to cope with its surroundings.

The intracellular concentrations of proteins are in many cases
regulated at the level of transcription. Strict control requires
specific proteins which recognize and bind to regions of the genome,
and modulate the degree of transcription by interacting with the
DNA-dependent RNA polymerase, either through direct contacts or
perhaps by altering the structure of the double helix itself. To
understand these interactions at the molecular level, it is
necessary to evaluate the stoichiometry of the reactions, the order
of binding of the proteins to DNA, the rates at which these proteins
bind and dissociate, and the strength of the interactions. A
powerful tool for studies of such systems is polyacrylamide gel
electrophoresis. This approach allows one to isolate DNA-protein
complexes for further study, as well as permitting analysis of the

thermodynamic and kinetic parameters involved. In this work

polyacrylamide gel electrophoresis is evaluated as an experimental
tool. It is shown that the gel matrix does not alter the integrity
of complexes within it, and that the matrix does not impose an
artificial stability on complexes migrating through it.
Furthermore, the conformation of DNA-protein complexes during
electrophoresis was studied. It is important to know how complexes
move through gels, because the mobility of complexes is a function
of their shape. A bend in the DNA leads to an altered mobility;
thus if a regulatory protein is found to bend the DNA upon binding,
some insight into the mode of action of that protein may be obtained.
Finally, an attempt was made to use certain cloned DNA fragments to

learn more about the mechanisms by which Escherichia coli RNA

 

polymerase searches for promoter regions on the bacterial

chromosome.

CHAPTER I

Comparison of Nucleic Acid-Protein Interactions

in Solution and in Polyacrylamide Gels

Polyacrylamide gel electrophoresis has seen widespread use in
the fractionation of nucleic acids and proteins since its
introduction approximately twenty years ago (1,2,3). More recently,
it has been successfully used to separate DNA-protein complexes from
free DNA existing in a sample (A,5). The essence of the latter
approach is to mix a DNA fragment containing a specific binding site
with the protein of interest under appropriate solution conditions,
including any divalent ions or cofactors needed. When binding is
complete, the sample is loaded onto a low ionic strength gel and
electrophoresed for the shortest time required to give the desired
separation of the components. The resulting pattern, which may be
visualized by ethidium bromide staining and/or autoradiography, will
show a diminution of the DNA band and possibly the presence of a
band of DNA-protein complexes. The appearance of a band of
complexes is dependent upon numerous factors: the intrinsic
strength of the interaction, the ionic strength of the gel buffer,
the temperature of the gel during electrophoresis, the total time of
electrophoresis, etc. Therefore, the most rigorous method for
quantifying the degree of complex formation is to measure the level

of unbound DNA rather than the complex to avoid artifacts caused by

dissociation of the complexes which may take place during the
experiment. The measured amount of free DNA will accurately reflect
the level of unbound DNA in the sample loaded onto the gel as long
as the equilibrium is not disturbed during the "dead time" required
for the free DNA to enter the gel (usually a few minutes).
Variations of the above technique are now used to purify eukaryotic
factors which specifically bind to a given DNA fragment by mixing
the DNA with a crude cell extract and electrophoresing the
DNA-protein complexes away from the other components (6,7,8). The
gel binding technique is also used to separate "free" and "bound"
DNA in "interference" experiments which reveal sites where DNA and
protein are in close proximity (9).

As discussed by Garner and Revzin (A) and by Fried and Crothers
(5), accurate thermodynamic and kinetic parameters can be derived
from the gel electrOphoresis technique even if the complexes
dissociate during the run. (What is necessary is that they be
long-lived with respect to the minute or so (see above) required for
the free DNA to enter the gel.) In this respect, the technique
resembles the nitrocellulose filter assay (10) except that free DNA
is normally quantified in the gel assay while complexes are measured
directly by the filter assay. The presence of a band of complexes
in the gel is a bonus since it allows one to determine the
stoichiometry of the DNA-protein interaction. Several investigators

have found a 1:1:1 stoichiometry for the Escherichia coli catabolite

 

activator protein (CAP):cAMP:lac promoter system (11,12,13). In

studies of lac repressor Fried and Crothers (5) observed that

repressor-operator binding in the gel appeared to be stronger than
expected. They proposed that at least a portion of the observed
stability was due to a "caging effect"; that is, the apparent
stabilizing of the complexes by the gel matrix. Their hypothesis
was that the gel polymers would hinder the escape of the DNA
molecule from the protein after dissociation. This would allow the
complex to reform more readily within the gel, leading to an
abberantly slow dissociation. Alternatively, the polyacrylamide
could alter the stability of the complexes by changing the
properties of the solvent in some manner.

In the wake of the conclusions of Fried and Crothers (5) and
the increasing use of the technique, a further evaluation seemed
appropriate. The first section of this work asks whether
perturbations from the gel matrix may affect the properties of the
complexes. Given that the gel assay provides a tool for answering
many experimental questions, including the isolation of DNA-protein
complexes, does one actually isolate the same, viable complexes
which are present in the absence of the gel? Do complexes in a gel
behave as they do in solution? Do parameters determined from gel
experiments accurately describe the DNA-protein interaction under
study?

In this regard it is noteworthy that the gel and filter binding
assays yield the same results when used to assay the same system.
The association rates of RNA polymerase with the A PR promoter were
identical as measured by each technique (1A,15). Furthermore, in

studies on the association of RNA polymerase with the lac UV5 and

the gal P2 promoters, Shanblatt and Revzin (15) found an unusual
biphasic kinetics behavior. As a control all assays were confirmed
using the filter binding technique; the data were superimposable.
Finally, Maxwell and Gellert (16), in work on the interactions of
DNA gyrase with various DNAs, determined affinity constants using
both methods and found them to be identical.

We have studied DNA-protein complexes in gels in the following
ways. First, RNA polymerasejlag UV5 promoter complexes were formed
and electrophoresed into a polyacrylamide gel, then chased with a
nucleotide mixture to see what, if any, transcription products were
formed. If the gel has no effect on the complex, then one would
expect to see the same transcription pattern as is observed in a
solution reaction (without the gel) using the same enzyme and
template. If, on the other hand, the gel is causing some sort of
change in the complex, this would likely lead to an altered product.
Second, the dissociation constants of various cAMP-CAP-Iag
DNA complexes were measured both in the gel and in solution. If
Fried and Crothers' caging hypothesis is correct, then one would
expect to see an increased stability of the complexes in the gel as
opposed to in solution, as they reported for the lag
repressor-operator system. Also, one would predict an increased
apparent stability for complexes formed with longer DNA than with
short fragments, since the longer DNA will be more hindered in
moving away following dissociation of the complex. However, if
caging does not exist, then one would observe no difference between

the properties of the complexes in solution and in the gel.

MATERIALS AND METHODS

Materials

Unless otherwise specified, all reagents were ACS reagent grade
obtained through normal commercial sources and were used without
further purification. (Y-32P)ATP and (a-32P)UTP were purchased from
ICN Radiochemicals. CAP (17) and RNA polymerase holoenzyme (18,19)
were prepared as previously described (A); protein concentrations
were determined spectrophotometrically (A). All protein
concentrations are given in terms of active molecules; RNA
polymerase was about 50% active and CAP about 25% active in a
specific binding assay (A). Restriction enzymes, synthetic linkers,
and DNA modifying enzymes were purchased from Bethesda Research
Laboratories, Inc., New England Biolabs, Inc., Boehringer Mannheim
Biochemicals, or International Biotechnologies, Inc., except EggRI,
which was purified as described in Garner and Revzin (A). The 211
bp wild type lag and mutant L8-UV5 lag DNA fragments were isolated
from recombinant pMB9 plasmids generously provided by Forrest Fuller.
The 789 bp wild type lag fragment was previously cloned into the
EEQRI site of pBR322 in this laboratory. The 6“ bp fragment
containing only the wild type lag CAP site was generated by Alul
digestion of the 211 bp fragment, and was the gift of Roger Wartell.
This fragment was also previously cloned into the EEQRI site of
pBR322. All of the above plasmids had been transformed into one of

several E. coli strains.

Methods

DNA Fragment Preparation

 

E. coli strains containing recombinant plasmids were grown and
the supercoiled plasmid isolated by the method of Clewell (20)
followed by centrifugation in CsCl-ethidium bromide gradients.
Purified supercoils were restricted with the appropriate enzyme,
extracted with phenol-CHCl3 and ether, and ethanol precipitated.
The DNA was resuspended in a convenient volume of "TE" buffer,
2x10"2 M Tris-Cl, pH 8.0 at 23°C, 1x10'“ M EDTA, then 1/10 volume of
a solution of 25% Ficoll, 0.1% bromphenol blue and 0.1% xylene
cyanol (hereafter referred to as 10X loading dyes) was added. The
sample was heated to 60°C for 5-10 minutes and loaded onto a
preparative polyacrylamide gel (30:1, acrylamide:bisacrylamide) in
9x10"2 M Tris base, 9x10"2 M boric acid, 2.5x10"3 M EDTA (TBE). The
gels were run for various amounts of time at 30-40 mA, to give
suitable separation of the fragment from the plasmid and/or other
restriction fragments. The dyes were used as reference points; on a
5% gel, the bromphenol blue comigrates with 65 bp DNA while the
xylene cyanol runs with 260 bp DNA. The gels were stained with
ethidium bromide, and the slice of gel containing the desired insert
was excised. The DNA was recovered by electroelution, then
extracted successively with TE-saturated N-butanol, phenol,
phenol-CHCla, CHCla, and ether. The DNA was precipitated with
ethanol and resuspended in TE. DNA concentrations were determined

spectrophotometrically, using 8260 = 13,000 (M,bp)'1.

In-Gel Transcription Reactions

 

The 211 bp lag L8-UV5 fragment was present at 11.8x10"8 M in a
50 pl sample of 1.11x10"2 M Tris, pH 7.9 at 23°C, 1.3x10"2 M NaCl,
3x10"3 M MgCl,, and 11.8x10'7 M RNA polymerase. (This low salt
buffer supports transcription and is also suitable for
electrophoresis). The solution was incubated at 37°C for 15
minutes, then heparin was added to a final concentration of 1x10’1
g/l to destroy nonspecific DNA-protein complexes and sequester
unbound polymerase. After the addition of 5 pl of 10X loading dyes,
the sample was immediately electrophoresed into a 5% polyacrylamide
gel (30:1, in the same buffer as the sample), and allowed to migrate
under an electric field of 15 V/cm for 30 minutes. Following this,
a 6 pl sample of nucleotides containing 1.6x10"3 M ATP, GTP, and
CTP, 8x10“5 M UTP, and 60‘}Ci of (a-32P)UTP was loaded in the same
lane as the complexes and electrophoresed for 3 hours during which
time the nucleotides passed through the complexes, allowing
transcription to occur. Bands of radioactivity were located by
autoradiography and excised from the gel, crushed in 5x10"l M
ammonium acetate, 1x10'3 M EDTA, and left overnight to elute the RNA.
The solution was then extracted with phenol and phenol-CHCla, and
ethanol precipitated prior to sizing (see below).

Solution Transcription Reactions

 

The 211 bp lac L8-UV5 fragment was made 5x10"8 M in the same
buffer as used for in-gel transcription reactions. RNA polymerase
was added to a concentration of 1x10”7 M active molecules and the

solution was incubated at 37°C for 15 minutes, then 5.P1 of the

10

nucleotide mix (see above) was added. Transcription was allowed to
proceed for 15 minutes, after which 30 pl of a solution of 3x10“1 M
NaOAc, 2x10"2 M EDTA, and 1x10'l g/l tRNA (STOP buffer) was added to
terminate transcription. The samples were phenol extracted and
ethanol precipitated as described above.

Sizing of Transcripts

 

RNA pellets were suspended in 90% formamide and heated to 90°C
for 5 minutes, then chilled on ice and loaded onto a 12%
polyacrylamide gel (20:1) in TBE plus 7M urea (21). Following
electrophoresis at 50 watts (until the bromphenol blue had migrated
to the bottom of the gel), the gel was transferred to Whatman 3MM
paper, dried, and visualized by autoradiography. Molecular weight
markers were derived from pBR322 digested with MpaII and 5'
end-labeled with (Y-32P)ATP as per Maniatis gt al. (22).

Determination of Dissociation Constants in the Gel

 

The appropriate DNA fragment was made 5x10"8 M in 2x10"2 M
Tris, pH 8.0 at 23°C, 3x10"3 M MgClZ, 1X1O—3 M DTT and EDTA, 1X10_1
M KCl (binding buffer), and 2x10'5 M cAMP, and a negligible
concentration of 32p-labeled DNA was added. CAP was added to
1.25x10'7 M, and the solution was incubated at 37°C for 5 minutes;
10X loading dyes were added and the solution loaded onto a 5%
polyacrylamide gel (30:1) in TBE plus 5x10"6 M CAMP. Samples were
electrophoresed for varying amounts of time, after which the gel was
dried on Whatman 3MM paper and autoradiographed. The results were
quantified by cutting out the complex band and the rest of the lane,

and counting them in a scintillation counter in 5 ml of

ll

toluene-based scintillation fluid. The fraction in complex was
determined by (cpm in complex/total cpm in lane). Dissociation
constants were evaluated using the method of least squares and the
equation for a first order process, -ln (% complexes) = kdt. This
technique measures only the dissociation taking place within the gel
itself.

A second approach was used to determine the dissociation
constant of CAP-DNA complexes within the gel. In this method (5)
DNA-protein complexes are electrophoresed into a gel and then chased
by excess unlabeled DNA layered onto the gel at a later time. As
the chase passes through the complexes, any protein which
dissociates is captured by the unlabeled DNA and is no longer
available to interact with the labeled DNA. By comparing the amount
of complexed labeled DNA in two lanes, one with and one without the
chase, one can in principle determine the dissociation constant.
Drawbacks to this technique are the facts that only two data points
are obtained and that it is difficult to determine the exact amount
of time that the chasing DNA is in contact with the complexes.

Determination of Dissociation Constants in Solution

 

The radioactive 211 bp wild type lag fragment was made 5x10‘8 M
in 9x10"2 M Tris base, 9x10"2 M boric acid, 2.5x10“3 M EDTA, and
2x10'5 M cAMP. This is the same buffer used to determine the
dissociation constants within the gel; therefore the results
obtained in solution and in the gel are directly comparable. In
this case enough solution was made to load A lanes (1OO‘PI). CAP

was added as before, and the solution was incubated at 37°C for 5

12

minutes. The solution was then placed at room temperature, a
10-fold excess of unlabeled DNA was added, and 25 Pl samples were
removed at various times and electrophoresed as above. The lanes
were cut up and counted as described for the in-gel dissociation
constants; in this experiment the fraction of DNA in complex is 1-
(cpm in free DNA band/total cpm in lane). This method determines

the amount of complexed DNA at the time the sample was loaded and

 

circumvents problems from dissociation of complexes during the run.

The kd value was determined as above.

RESULTS AND DISCUSSION

Comparison of Transcripts Made in the Presence or Absence of the Gel
Matrix

Beckman and Frankel (23) were able to detect E. coli DNA
polymerase and RNA polymerase activities in a polyacrylamide gel
following electrophoresis of the proteins in the presence of calf
thymus DNA. These studies have been extended here to determine
whether the product of such enzyme activity is the same in the
presence or absence of the gel matrix. The RNA polymerase-DNA
complexes were made and electrophoresed as described in Methods.
Figure 1 shows a typical result following the nucleotide chase and
autoradiography of the wet gel. Three bands are apparent, labeled
1, 2, and 3 in the figure; in control experiments it was shown that
the DNA-RNA polymerase complexes comigrated with band 1. Figure 2
compares these transcription products to those made in solution in
the absence (lane 2) or presence (lane 3) of heparin; in all cases
the expected product is the 69 nucleotide runoff transcript. A
direct comparison can be made since the same buffer was used for the
solution reactions and the in-gel transcription (1.11x10"2 M Tris, pH
7.9 at 23°C, 1.3x10”2 M NaCl, 3x10"3 M MgClz). These data show that
the main product is the same whether or not the gel matrix is
present. While it appears that the RNA associated with band 1 in
Figure 1 (lane 4 in Figure 2) consists mainly of small fragments, in
other experiments the transcripts in this band were similar to those

found in bands 2 and 3 of Figure 1 (lanes 5 and 6 in Figure 2).

13

 

l
l infill

Figure 1: Autoradiograph of RNA made in a polyacrylamide gel after
the electrophoresis of RNA polymerase-lag UV5 complexes. Lane 1
contains complexes, lane 2 is nucleotides only. A separate gel
Showed that the complexes migrate to the same position as band 1.

123456

:I ””11:

"

90- "" z .

76~ ~. '

67' ‘Q .l
0

I

1

34-

‘7
.II I H o 1
~-' .

«.010 ll
.-
O

Figure 2: Comparison of transcripts made in a polyacrylamide gel
with those made in solution. Lane 1; molecular weight markers.
Lanes 2 and 3; RNA made in a solution transcription reaction in the
absence (lane 2) or presence (lane 3) of heparin. Lanes A-6; RNA
eluted from bands 1-3, respectively, of Figure 1.

16

This experiment was done both with and without heparin present in
the gel; therefore whether RNA polymerase can reinitiate is not an
issue. A comparison of lanes 3, 5 and 6 in Figure 2 indicates that
more short transcripts seem to be produced in the gel than in
solution (in the presence of heparin). While this could indicate
that the gel matrix is somehow altering the reaction, a more likely
explanation is that the discrepancy arises from differences in total
nucleotide concentrations, which cannot be absolutely controlled
during electrophoresis due to band spreading during sample loading
and the run. This does not alter the main conclusion, that whatever
else may be happening in a polyacrylamide gel, DNA-RNA polymerase
complexes can transcribe just as they do in the absence of the gel.

Effects of the Gel Matrix on CAP-DNA Dissociation Constants

 

To assess more directly whether the gel matrix perturbs
DNA-protein interactions, the dissociation constant for CAP-wild
type lag DNA complexes was examined as a function of DNA length. If
there is a "caging effect" as suggested by Fried and Crothers (5)
(i.e. if the gel matrix imposes an apparent artificial stability on
the complexes), then one would expect to see a slower dissociation
rate for CAP—DNA complexes in which the DNA is long versus those
made with shorter DNA. That is, if escape of DNA from protein is
limiting, then the movement of longer DNA should be hindered to a
greater extent than that of shorter DNA molecules. In order to test
this hypothesis, CAP-DNA complexes were made and electrophoresed as
described in Methods. Three different size DNA molecules containing

the CAP site were used; a 6A bp fragment, a 211 bp fragment, and a

 

17

Table 1

Dissociation Constants of cAMP-CAP-[azP] lac Promoter Complexes

Size of Promoter Containing

Fragment kd, s"l
6Abp
Trial 1 14.3x10's
Trial 2 7.5x10'5
Average 5.9x10"5
2110p
Trial 1 5.8x10"5
Trial 2 5.6x10"5
Average 5.7x10"s
789bp
Trial 1 7.8x10”5
Trial 2 11.5x10"s
Average 6.2x10"s

18

789 bp fragment. These fragments encompass a twelve-fold range of
DNA sizes. A typical experiment using the 211 bp fragment is shown
in Figure 3. Table 1 summarizes the experimental data. As these
dissociation constants are the same, within experimental error, it
can be concluded that caging is not an important factor in this gel
system.

To confirm the above results, the dissociation constant of the
CAP-211 bp fragment was determined in solution as described in
Methods. In this approach polyacrylamide gel electrophoresis is
used only to separate the free DNA from the complexes existing at
the time of the loading of the gel. The value obtained by this
method was 7.8x10"5 8“, within experimental error of that
determined in the gel itself (Table 1).

As a final confirmation, another approach (5) was used to
determine the dissociation constant within the gel. This procedure,
although far from ideal (see the section in Methods), nevertheless
gave a value of 1.4x10’“ 3“, also in reasonable agreement with the
previously mentioned values.

The caging model of Fried and Crothers was proposed to explain
the finding that specific lag repressor—DNA complexes were
apparently more stable in a gel than in solution. Our results are
not consistent with this model; if escape of DNA from protein in a
gel is not a limiting factor (as these data imply), then the gel
matrix should not affect dissociation of complexes of either
repressor or CAP with DNA. The resolution of this issue may reside

in reference (5). It appears as though the dissociation rates being

 

 

Figure 3: Dissociation of CAP-wild type lac promoter complexes
during electrophoresis in a polyacrylamid3_gel. Lane 1 was
electrophoresed for 120 minutes, lane 2 for 90 minutes, lane 3 for
60 minutes, and lane A for 30 minutes. Data were analyzed as
described in the text.

20

compared there are for different buffers. The shorter half-life
reported for repressor-operator complexes in solution appears to be
for a buffer containing 50 mM KCl, while the rate in the gel
reflects the electrophoresis buffer, which contains no KCl. The
discrepancy is in the expected direction; dissociation would likely
be faster at the higher salt concentration. Therefore it is not

clear that a caging effect was actually observed.

CHAPTER II

DNA Fragments Containing Multiple Protein Binding Sites For
Studying the Movement of DNA-Protein Complexes in

Polyacrylamide Gels

Given that polyacrylamide gel electrophoresis is a useful way
to isolate viable, native complexes, the next question to address
was in what conformation DNA and DNA-protein complexes migrate
during an electrophoresis experiment. The prevailing model for
migration of DNA through a polyacrylamide gel is end-on migration in
which the DNA molecule orients with the electric field and moves in
a worm-like fashion from pore to pore (24,25). This "snake in the
grass" model has been referred to as primary reptation (25,26).
Lumpkin and Zimm (27) have derived an equation which describes this
model: X=(hX)ZQE/L2F, in which the mobility, X, is proportional to
(hx)2, the square of the component of a fragment's overall length
which is parallel with the electric field (E), and to Q, the charge
on the molecule, and is inversely proportional to (L)2, the square
of the overall fragment length, and F, the translational frictional
coefficient. From this equation, one can conclude that for two DNA
fragments of the same overall length under a given electric field,
the mobilities will depend on the conformation (hx), since both Q
and F depend on L such that Q/F is independent of L (27). Several
authors have made use of this equation in their research; Stellwagen
(28) used it to explain her observation that A-T rich DNA fragments

21

22

of pBR322 migrate anomalously slowly due to a kink or bend in the
DNA caused by the A-T rich region. Such studies were extended by
Diekmann and Wang (29) and Wu and Crothers (30) to include their
findings with restriction fragments of trypanosome kinetoplast DNA
(also found to be A-T rich). Wu and Crothers (30) also showed that
the location of the bend within the fragment has an effect on the
mobility; that is, a bend near the end of a fragment causes a
smaller shift in mobility than does a bend near the center of the
same size fragment. These authors reported that CAP-DNA complexes
exhibit the same behavior (30). A CAP molecule bound in the center
of a fragment causes a larger decrease in mobility than does a CAP
bound near the end of a fragment of the same size. Since no such
effect was seen for lag repressor-operator complexes, Wu and
Crothers concluded that CAP was causing a bend to occur in the DNA
upon binding which was not caused by repressor.

The average pore size of a 5% polyacrylamide gel has been
reported to be approximately 3.6 nm according to Cooper (31 and
references therein). The proteins used in our study have the
following properties: CAP, molecular weight “5,000 (32),
approximate diameter 5 nm (33); RNA polymerase, molecular weight
A60,000 (3A), approximate diameter if spherical, 14 nm (35). The
diameter of the DNA double helix is 2.0 nm (36). It is perhaps
somewhat surprising that DNA-protein complexes are capable of
migrating by primary reptation (or even enter the gel at all); that
is, it seems that the pores of the gel might be too small to allow

for such motion. To learn more about the mode of migration, four

 

 

23

Figure A: Diagram of fragments A-D. The constructions are described
in the text. Solid boxes denote the CAP binding site, and the hatched
rectangles represent the RNA polymerase binding site; the arrow shows
the direction of transcription. The component fragments denoted refer

to Figure 7.

24

 

 

 

 

r0
0)
V
10
|\
<r

 

 

 

 

 

 

 

 

 

 

 

 

 

dd mmm .o EmEoEn. 1%! _ . . E11:
Hm com 39:: HI Eom B E: Hm cow

v M9» “ why A

an mmm .0 EmEmE... T J E I ..1. Val . 4
HI Eom B 9:: E 8m 5 9:: HI Sam

.I N made—2 " m2» 1
dd mmw .m EoEoEu .l HI .IEUIJII
E 8w H: Eom E 9:: E 8m

.11 N mm do: Jr who i.

an we. .4 29:08... T _ El . a

H: can. E am. E man HI solm

25

DNA fragments were constructed. Two fragments contain one CAP or
RNA polymerase binding site, located either in the middle or at one
end. The other two fragments contain two CAP or RNA polymerase
binding sites, located either near each other in the center or one
at each end. These fragments are shown schematically in Figure A;
the heavy regions denote the protein binding sites. By binding the
proteins, either separately or in combination, and examining the
electrophoretic mobilities of the resulting complexes, it was hoped
to learn more about the way in which these complexes migrate. Based
on the relatively small pore size and the fact that it is the DNA
which is responsible for pulling the protein into the gel, it seems
possible that some of these complexes will migrate via different
motifs. We reasoned that primary reptation might require that there
be at least one end of the DNA molecule available to lead the
complex into the gel. Since the dumbbell-shaped complex having a
protein at each end will not possess such free ends, one might
expect thesecomplexes to migrate like a hairpin or a horseshoe with
the DNA in the center pulling the proteins at the ends into the gel.
Thus this complex should migrate more like a circular DNA than a
linear molecule. Because circular DNA fragments have much slower
mobilities than the corresponding linear fragments (37.38.39), there
should be a large difference in the mobilities of this complex
compared to the complex in which the two proteins are both bound in
the center of the fragment. In terms of the equation given above,
the complex moving as a hairpin has a smaller (hX) and therefore a

smaller mobility.

26

Thus work with these four fragments may yield insight into the
way in which DNA molecules and DNA-protein complexes migrate through
polyacrylamide gels. As an added benefit, the fragments lend
themselves to studies on how RNA polymerase searches for promoter
sequences within the E. 39;; genome. A more detailed description of

these experiments will be found in the Results and Discussion

 

section immediately before the pertinent results.

MATERIALS AND METHODS

Materials

All materials were as those described in the Materials section

of the preceding chapter.

Methods

Construction of Fragments

 

All DNA manipulations were performed using enzymes and reagents
as described in Maniatis EE El- (22). Therefore, this section will
deal only with the construction strategy. Recombinant plasmids were
transformed into E. 99;; strain HB101 using the method of Hanahan
(A0); screening was accomplished by the procedure of Birnboim and
Doly (A1) to prepare small quantities of plasmids which were then
restriction mapped.

Construction of lac 211 bp fragment containing the UV5 promoter

 

mutation and the wild type CAP site: There were two species of

 

cloned E33 211 bp fragments available in the laboratory. The wild
type promoter requires CAP to be bound before RNA polymerase can
form transcriptionally competent complexes at its primary binding
site (A2). The other fragment contains two mutations; the L8
mutation inhibits CAP from binding (A3), while the UV5 mutation
allows RNA polymerase to bind and transcribe efficiently in the
absence of CAP (AA). It was desirable to construct a "hybrid"
promoter region containing the wild type CAP site and the UV5 RNA

27

28

polymerase binding site. Ultimately, this was used to construct
fragments to which either CAP or RNA polymerase (or both) will bind
strongly. A partial map of the Egg promoter is shown in Figure 5;
the L8 mutation is within the CAP site at position -66 and the UV5
mutation lies within the RNA polymerase binding site at positions -9
and -8. The fragment in the figure is shown without synthetic
linkers attached. The hybrid was made by digesting each fragment
separately with EngI and separating the products on a

polyacrylamide gel. Following elution of the DNA, that portion of

 

the wild type fragment extending from -1A0 to -20 was mixed with the
-19 to +63 segment from the L8-UV5 fragment, and with pBR322 which
had been digested with EggRI and calf intestinal phosphatase. DNA
ligase was added, and the resulting plasmid was then used to
transform E. 99;; strain HB101 as described above.

Preparation of "nonpromoter DNA": To make the constructions

 

described in the Introduction, it was necessary to find and prepare

 

for use a sequence of DNA which did not bind either CAP or RNA
polymerase. One such fragment exists between the EEEHI and §Ell
sites of pBR322, as evidenced by the lack of specific complexes in a
binding gel assay (A). Since this sequence must be used twice in
each of the constructs shown in Figure A, and since inverted repeats
are not stable in E. 99;; strain HB101 or other Rec A“ strains (A5),
it was necessary to prepare this fragment in two ways such that the
restriction sites at the ends of the fragments were reversed. The
preparation of these fragments, termed Preliminary Fragments (P.F.)

1 and 2, is shown in Figure 6. To construct P.F. 1, pBR322 was

29

Figure 5: Partial map of the lac operon. The fragment has EcoRI ends.
Internal restriction sites pertinent to the constructions described in
the text are also shown.

 

 

30

t0
cow

 

 

9. mm_. L
. . I L... a
+ 1.. :3: man. oi-
<20
mi 88295 omucoezod <zm mv-
82,8

n30

_ , Barn HI
5.011
Blunt with
Klenow Fragment

Recircularize
Ec RI

Bam HI
Sal I

811

{——

Blunt
Attach him 111 Linkers
mg III
Recircularize
Egg RI

_B_g_rr_t HI

fliniIII

P.F. I Construct

Figure 6:
P.F. 1 and 2.

  
 

Blunt with
Klenow Fragment

Attach HindIII Linkers
Hind III

Recircularize

Ji'msiIiI

__ $3111

v§2l1
Blunt with
y/Klenow Fragment

VAttach Bgm HI Linkers
vBam HI

Recircularize
sl/

1mm

EoR
— .Bam HI

RF. 2 Construct

 

Flow diagram of the construction of plasmids containing
See text for a description of the steps.

 

32

digested with EEEEIII, blunted with Klenow fragment, and
recircularized by blunt end ligation. This step destroys the
existing EEEEIII site, and was necessary to prevent subsequent loss
of the fragment of interest. A new EEEEIII site was created at the
pBR322 Egll site by digesting the new plasmid with Egll and blunting
with Klenow fragment. Synthetic EEEEIII linkers were attached to
the blunt ends, the product was restricted with fllflQIII' and the
molecule recircularized with DNA ligase following gel purification
to remove the excess linkers. The recircularized plasmid was
transformed into E. 9911 strain H8101. Restriction of this plasmid
with EEEHI and EEEEIII yields P.F.l.

P.F. 2 was constructed similarly by cleaving pBR322 with REEHI
and converting this site to EEEQIII’ then cutting with Egll and
changing that site to BEEHI: Transformation into H8101 followed
this step. Schematic diagrams of P.F. 1 and 2 are shown in Figure
7.

Construction of the promoter-containing fragments P.F. 3 and A:

 

Since the promoter fragment must also be used twice in making
constructs C and D (Figure A), it was necessary to prepare two
promoter-containing fragments with the restriction sites at the ends
reversed, as done with P.F. 1 and 2 above. To construct P.F. 3 (see
Figure 7), the 211 bp hybrid Egg fragment (wild type CAP site, UV5
promoter) described above was digested with EXEII and the blunt end
converted to EEEEIII by the use of synthetic linkers. The linkers

only attach to the blunt PvuII end, leaving the EcoRI site unchanged.

 

33

Figure 7: Diagram of DNA fragments used in fragment A-D constructions.
Preliminary Fragments 1 and 2 and "modified" 2 originate from pBR322;
the arrows show their original orientation clockwise from the plasmid
EggRI site. The solid square in Preliminary Fragments 3 and A and in
the A75 bp and A93 bp fragments represents the CAP binding region,
while the hatched section denotes the RNA polymerase binding site. The
arrows indicate the direction of transcription. The A75 bp and A93 bp
fragments are described more fully in the text.

 

34

 

 

 

 

 

 

Till c um i N mm A
EmEooi an mmv
.Illa . J
E clam. 54:“ E 5'8
Eoanm dnmhv Tlmud . _ an i.
W E . J

E a w B 5: E sum

as mam .N E 8582

 

 

 

 

 

rl lllllllllllllllll U
HI 6.5 E molm
an mow .c mm
n pvt-FIE]
Bus: E g
.3 cm. .m E
q E‘ J
E claim. E Us:
.3 hmm .N mm
II” IIIIIIIIIIIIIIIIIII IIL
HI £8 E E:

.2 .mm ._ CE Essen boséaca

35

This fragment was cloned into pBR322 between the EggRI and EEggIII
sites.

To construct P.F. A, several steps were required. The
recombinant plasmid containing P.F. 3 was cleaved with either
EEggIII or EggRI to generate linear plasmid. The EEggIII cut
plasmid was recircularized after changing the EEggIII site to EggRI,
resulting in a Egg fragment which contained EggRI sites at each end.
Likewise, the EggRI cleaved plasmid was recircularized by changing
the EggRI site to EEggIII, resulting in a Egg fragment flanked by
EEggIII sites. These two steps result in two Egg fragments which
are identical except that one fragment has EggRI sites at each end
while the other has EEggIII sites at both ends. Now by the same
method as used above to construct the hybrid 211 bp fragment, the
region from -19 to +63 from the EEggIII clone was attached to the
~123 to -20 fragment from the EggRI clone. The resulting fragment
was cloned into pBR322 as was P.F. 3; both new plasmids were
transformed into HB101.

Construction of fragments A, B, C, and D (refer to Figure A for

 

a schematic): P.F. 1 and 3 were joined at the EEggIII site, to
yield a A75 bp fragment; P.F. 2 and A were similarly joined, making
a A93 bp fragment. Fragments A and B were constructed by mixing the
A93 bp fragment with P.F. 2 in which the EEggIII site had been
changed to EggRI (see Figure 7), along with pBR322 digested with
either EggHI (for Fragment A) or EggRI (for Fragment B). Fragments
C and D were constructed by mixing the A75 bp fragment, the A93 bp

fragment, and the same vectors used above. All four new plasmids

36

were transformed into 88101. Figure 8 shows a gel from which
restriction maps of the fragments were deduced, to verify the
correctness of the constructs.

Construction of 1331 bp fragment Fragment C was found to be

 

oriented in its plasmid such that the promoters transcribed
clockwise with respect to the standard pBR322 map. The attachment
of nonspecific DNA was accomplished by partially restricting the
recombinant plasmid containing Fragment C with EggHI followed by
complete restriction by EEgI. The digestion products were
electrophoresed, and the fragment extending from the EEgI site to
the far EggHI site (with the near EggHI site intact) was purified by
electroelution. This fragment was blunted with Klenow fragment and
EgEI linkers were attached; the fragment was then cloned into the
Eggl site of pBR322 and transformed into 88101. Figure 9 shows a
schematic diagram of this fragment.

DNA Fragment Preparation

 

All DNA fragment preparations were done as described in the
preceding chapter.

Electrophoresis of Various DNA Fragments and Complexes Thereof

 

All binding reactions contained 5x10'° M DNA which had been 5'
end-labeled (33), and 1x10"7 M RNA polymerase or 2.5x10"7 M CAP and
2x10~s M cAMP in 2x10“2 M Tris, pH 8.0 at 23°C, 3x10"3 M MgClz,
1x10"3 M DTT and EDTA, and 1x10"1 M KCl. Solutions were incubated
at 37°C for 15 minutes; if RNA polymerase was used the solutions
were then made 1x10"l g/l in heparin. A 1/10 volume of 10X loading

dyes was added, and the samples were loaded onto a polyacrylamide

37

Figure 8: Restriction digests of recombinant plasmids containing
fragment A (lanes 1-3), fragment 8 (lanes A-6), fragment C (lanes
8-10), and Fragment D (lanes 11-13). Lanes 1 and 8: EggRI, lanes A and
11: EggRI, lanes 2, 5, 9, and 12: EggHI and EggRI, lanes 3, 6, 10, and
13: EggRI, EggRI, and HindIII. Lane 7 shows molecular weight markers:
from top to bottom; 1300, 789, 622, 527, A03, 305, 2A2, 238, 217, 201,
190, 180, 160, 1A7, 122, 110 bp. The 375 bp fragment seen in lanes 2,
5, 9, and 12 results from plasmid cleavage, as does the 3A6 bp fragment
seen in lanes 3, 6, 10, and 13. These digests were used to verify that
the correct constructs were obtained.

 

38

l

 

39

Figure 9: Diagram of the 1331 bp construct. The arrows denote the
direction of transcription from the promoters.

40

as .3.

Ha

 

 

 

 

B as: E mom. B E: E cm H E.
1 "IF “A o o . J
.P o EmEooi _ mmm l.—

41

gel (percentage and electric field varied; see Results and

 

 

Discussion). After electrophoresis of the xylene cyanol to the
bottom of the gel, the gels were dried on Whatman 3MM paper and
autoradiographed.

Transcription Reactions

 

The DNA fragment was made 5x10"e M in 25 pl of binding buffer.
RNA polymerase was added at varying concentrations and allowed to
react for 15 minutes, then a solution of ATP, CTP, GTP, UTP, (a-32P)
UTP, and heparin was added. Final concentrations were: ATP, CTP,
and GTP, 2x10‘“ M each; UTP, 2.5x10‘5 M; heparin, 1x10"1 g/l.
Transcription was allowed to proceed for 15 minutes and was
terminated with 30 pl of STOP buffer. The samples were extracted
with phenol-CHCI3 and precipitated with ethanol, resuspended in 90%
formamide, heated to 90°C and loaded onto a 5% (20:1) polyacrylamide
gel in TBE plus 7M urea (21). After electrophoresis at 50 watts
(until the bromphenol blue reached the bottom), the gel was
transferred to Whatman 3MM filter paper and dried. Following
autoradiography the transcripts were cut from the gel and counted in
5 ml of toluene-based scintillation fluid. Relative numbers of
transcripts were determined by dividing the number of counts/minute

by the number of uridine residues in the transcript.

RESULTS AND DISCUSSION

Electrophoretic Mobilities of CAP or RNA Polymerase Complexed with

 

Fragments A, B, C, and D

 

As mentioned in the Introduction, if the reported pore sizes

 

for polyacrylamide gels are accurate, it seems improbable that a
complex of fragment D and a protein would be able to migrate by
primary reptation, since there would be no free end of DNA to get
the complex started into the gel. Rather, it is conceivable that
this type of complex would migrate in a shape reminiscent of a
hairpin, with the protein molecules being dragged into the gel.
This conformation would have a pronounced effect on hx, the fraction
of the overall length of the DNA which is oriented with the field.
If this were the case, one would expect to see quite different
mobilities for complexes of C and D with either CAP or RNA
polymerase.

In any interpretation of the data, allowance must be made for
the bending phenomenon reported for CAP; CAP has been shown to bend
the DNA molecule when it interacts with its specific target site at
the Egg promoter (30), and the position of the bend has an effect on
mobility as predicted by the equation of Lumpkin and Zimm (27).

That is, a bend in the center of a DNA molecule retards its

migration more so than a bend near one end. This phenomenon can be
seen clearly in Figure 10, which shows CAP complexed with fragments
A, 8, C, and D. The complex formed with fragment A (lane 2 from the

left) migrates much more slowly than the complex formed with

42

43

 

Figure 10: Mobilities of CAP complexes with fragments A-D. In each
pair of lanes, the left lane is DNA only, while the right contains
complexes.

44

fragment 8 (lane A), due to the positions of the CAP binding sites.
This same effect can be seen with fragments C and D (lanes 6 and 8
in Figure 10). Similarly, Figure 11 depicts the result of RNA
polymerase binding to fragments A, 8, C, and D. The effects seen
with fragments A and 8 suggest some possible bending of the DNA
molecule by RNA polymerase. There are exactly 200 bp separating the
CAP sites (and also the RNA polymerase binding sites) in fragment C,
which corresponds to precisely 19.0 helix turns. This means that
the orientation of the proteins on the DNA will tend to enhance the
effect of bending on electrophoretic mobility. It would be
interesting to change the spacing to 19.5 turns (by inserting a
linker, for example) and compare the migration of this new fragment
complexed to either CAP or RNA polymerase to that of fragment C
complexed with the same protein. It is suspected that such a change
would lead to an increased relative mobility of the complex, since
the proteins would be binding on opposite sides of the helix and
would no longer be working in conjunction with each other to bend
the DNA. The binding of CAP and RNA polymerase simultaneously to
fragments A and 8 is depicted in Figure 12. We see by comparing
lanes A and 8 that bending by CAP occurs in the presence of RNA
polymerase.

Most experiments were done on A% gels (30:1) at 15 V/cm;
similar results were seen on 5% gels (30:1) and no difference was
seen using the A% gels over a range of electric field strengths

between 1 V/cm and 20 V/cm.

45

Table 2 summarizes all the mobility data. It can be seen from
this Table that there are no drastic mobility differences between
fragments A and B or fragments C and D with either CAP or RNA
polymerase bound. In particular, fragments C and D move similarly
when two RNA polymerase molecules are bound (Figure 11). This
result was unexpected; it was thought that fragment D would not
migrate by primary reptation, since there is very little overhanging
DNA to guide the complex into the gel. However, there are about 80

bp of overhanging DNA on one end of fragment D and about A0 bp on

 

the other end. It is possible that these overhangs are large enough
to allow the complex to enter the gel end-on. A construct similar
to fragment D having 25 bp overhangs behaves similarly to fragment D
(data not shown). The ideal case of no overhang is probably not
obtainable; recent results with the ggE P2 promoter suggest that
there is a minimum amount of overhanging DNA required for RNA
polymerase to bind (Dr. S.H. Shanblatt, personal communication).

All in all, our results suggest that the original hypothesis is in
error. Perhaps all of the constructed fragments are capable of
migration via the primary reptation motif. However, the question
then remains as to how a DNA-protein complex is capable of moving
through a pore which is barely large enough to allow the DNA
molecule to pass. The resolution to this problem may be that the
generally accepted pore size for polyacrylamide gels (31) is
incorrect. Ruchel and Brager (A6) report pore sizes of 1-2x103 nm,
approximately 3 orders of magnitude larger than the 3.6 nm indicated

in reference 31. This suggests that perhaps neither of these

46

 

Figure 11: Mobilities of RNA polymerase complexes with fragments
A-D. For each pair of lanes, the left lane is DNA only, while the
right contains complexes.

47

 

Figure 12: Mobilities of fragments A and B complexed with CAP and
RNA polymerase. Lanes 1-A; fragment A alone, with CAP, with RNA
polymerase, and with both proteins. Lanes 5-8 are the same for
fragment 8.

48

 

 

.>Ho>uuooamoc .>u«afinos o>HumHoL

undamas use .oumdooscoucfl .umoon on» no“: moxoaaeoo mafiafinoe on» on some; «L030; new ‘.8304 .coqa: .p—
oasmfim cH .wcficmoe 0: mm: «L030; uo>onm Boswmoc Luzon can some: mm mam .Lozoq can Load: .0, ocsmfim :H «

o>HpmHoL umocmH: new umozoa on» no“: moxoaqsoo ago was @0289

zmp.o
>:P.o
p©—.o

ommgoezaom «2m
new mac mafia .

sz.o .mwm.o .omm.o 1--
m_=.o .zmm.o .mmm.o It-
NN:.o .emm.o .smm.o ---

mom.o ._mm.o
mom.o .m_m.o
msm.o .mmm.o

>mm.o
omm.o
=0:.o

smm.o
s=m.o
som.o

ommcoeaaom
42m mafia P

.>Ho>fiuooamoe .>uHHHnoe

Lozoq was some: on» .—P com o— mossmfim :H .

owm.o
.mmw.o
.mow.o

woo.o .mmm.o
moe.o .zmm.o
omo.o .mpm.o

>m>.o
om>.o
wa>.o

:©©.o
owo.o
mmo.o

aeo mafia ,

ommco><
m Hague
— Hosea

«wcozoq.~cozoq .Loaas o

ommco><
N Hosea
F Hodge

.Lozoq .Loqaa o

ommeo><
m amass
P amass
m
ommno><
m Hodge
_ amass

<

acoemmcm

moxoaaeoo m .<rommcoe>Hom <zmrm¢o can .moxoaasoo a .o .m .<rommeoe>aom <zm
.mmxmflaeoo a .o .m .<-a<o co mama meuemaseoasaoa a: co mmHSHHHnoz o>fiumaom

N QHDMH

49

measurements is correct. Before primary reptation can be completely
confirmed, the pore size must be more rigorously determined. As a
means of addressing this issue, one could conceive of producing
scanning electron microscopy photographs as shown in reference A6 in
which DNA could actually be seen within the pores of the gel.

Bending of the DNA by RNA Polymerase

 

The presence of three bands of complexes in the case of
fragment D bound to RNA polymerase (lane 8 in Figure 11) likely
results from DNA bending caused by the protein. Titrating fragment
D with RNA polymerase shows that each of the two lower bands
contains one enzyme molecule. At low concentrations of RNA
polymerase the lower two bands are present in about equal amounts,
and upon increasing the protein concentration these bands diminish,
giving rise to the uppermost band, which contains complexes of two
polymerase molecules bound to one fragment. It is suspected that
the doublet results from the fact that the RNA polymerase binding
sites are not symmetrically placed within the fragment (see Figure
A). Consequently, the bends caused by RNA polymerase binding lead
to different configurations in the two complexes. As previously
discussed, there are about 80 bp of overhanging DNA on one end and
about A0 bp on the other. This difference is apparently large
enough to cause the observed effect. Support for this idea comes
from the results of studies of CAP and/or RNA polymerase binding to
the A75 bp and A93 bp fragments (Figure 13). Comparing lanes 3 and
A shows that the CAP-A93 complex has a higher relative mobility than

the CAP-A75 complex, since the CAP site is more centrally located in

50

the A75 bp fragment (see Figures 5 and 7). Furthermore, as expected
from the locations of the promoter sites, the polymerase-A75 complex
has a higher relative mobility than its counterpart containing the
A93 bp DNA (lanes 5 and 6). These migration shifts are in the
direction predicted if the distance of the bend from the end of the

fragment is important.

Mechanism of Promoter Search by E. coli RNA Polymerase

 

Another interesting question which can be addressed using
fragment C and its derivatives is whether E. ggEE RNA polymerase
finds promoter regions by a sliding process, in which the protein
first binds to nonspecific DNA and moves by facilitated
unidirectional diffusion toward its target site. von Hippel and
colleagues have shown that this mechanism applies for the Egg
repressor-operator system (A7,A8), and it has been shown that the
restriction endonuclease EggRI finds its site of catalysis in this
manner (A9). Modrich and coworkers used an assay in which EggRI
restriction sites were situated at varying positions within a DNA
fragment; the mechanism of site search was probed by adding a
sub-saturating amount of enzyme and analyzing the products on a
polyacrylamide gel. It was found that a restriction site was more
likely to be cleaved if it had a run of nonspecific DNA leading to
it. In other words, in a linear fragment with two restriction
sites, a site in the center of the fragment was more likely to be
cleaved than a site at one end because there was a greater

probability of the restriction enzyme finding the central site by

51

 

 

Figure 13: Mobilities of the A93 bp and A75 bp fragments complexed
with CAP and RNA polymerase. In each pair of lanes, the left lane
contains the A93 bp fragment, and the right the A75 bp fragment.
Lanes 1 and 2, DNA only; lanes 3 and A, plus CAP; lanes 5 and 6,
plus RNA polymerase; lanes 7 and 8, plus both proteins.

52

sliding. Several lines of evidence point towards E. ggEE RNA
polymerase using such a mechanism as it searches for its promoter.
In 1980 a group of authors (50) presented indirect evidence based on
the salt dependence of the enzyme-T7 promoter association rate. In
1982 (51) Wu and coworkers used an elegant photocrosslinking
procedure to try to locate RNA polymerase moving along DNA. We have
initiated studies to ascertain whether RNA polymerase slides during

its search for a promoter in a "sea" of nonspecific DNA. As

 

described in Methods, a 1331 bp fragment was constructed from

 

fragment C by attaching an additional piece of nonspecific DNA to
one end. The new fragment (see Figure 9) has 3A0 bp of DNA
downstream of the right-hand promoter and 730 bp of DNA upstream of
the left-hand promoter, counting from the edges of the RNA
polymerase binding sites shown in Figure 5. The right-hand promoter
makes a 362 nucleotide transcript, while the left-hand promoter
yields a 562 nucleotide transcript. If sliding is taking place,
then at sub-saturating RNA polymerase concentrations one would
expect to see more of the 562 nucleotide transcript than the 362
base transcript because the promoter giving rise to the 562 base
transcript is more centrally located. That is, the more central
promoter has a greater possibility of being encountered by a
polymerase molecule sliding along the fragment. At saturating
polymerase the ratio of transcripts from each promoter should be
unity, since promoter search is not rate limiting under these
conditions. Since both promoters are centrally located in fragment

C and thus should have about an equal probability of being found by

 

53

RNA polymerase, the ratio would be unity for these two promoters,
even at sub-saturating polymerase concentrations. The transcription
experiments were performed as described in Methods using 5X
(excess), 2X, 1X, 0.5x, and 0.1x (sub-saturating) total polymerase
concentrations, where X is the DNA concentration. The result was
somewhat unexpected: for both fragments, under all polymerase
concentrations tested, the ratio of 562 base transcript to 362 base

transcript was about 0.5. In other words, the end promoter was

 

preferred approximately 2:1 over the more central promoter. This

 

same result was seen at KCl concentrations from 2x10"1 M to 5x10”3
M.

One interpretation of this result is that RNA polymerase
prefers to approach its target from the downstream end of the
promoter. To test this hypothesis the A75 bp and the A93 bp
fragments used to construct fragment C (see Figure 7) were mixed and
transcription was done at the various polymerase concentrations
described above. The A93 bp fragment yields a 362 base transcript,
while the A75 bp fragment makes a 69 base transcript. If the
"preferential direction of approach" hypothesis is correct, one
would expect to see a preponderance of the longer RNA, since the A93
bp fragment has a much longer tract of nonspecific DNA leading to it
from the downstream end. However, the result obtained is just the
opposite- there is about twice as much transcript from the A75 bp
fragment as there is from the A93 bp fragment. Clearly, some
additional work is required. One explanation for this finding is

that transcripts from the more central promoter are, for some

54

reason, prematurely terminated. This could occur if a transcription
complex originating from the left promoter encounters a polymerase
molecule bound at the other promoter region. There is some evidence
for this; a band of short RNA of about the right length is seen in
transcripts from fragment C or from the 1331 bp DNA (data not shown).
That is, a short transcript appears which is about the correct size
to have originated at the central promoter and terminated at the
downstream promoter. If the number of short transcripts is added to

the number of full-length transcripts from the central promoter, the

 

ratio of (central promoter transcripts/noncentral promoter
transcripts) becomes 0.8, closer to unity. A more accurate result
might be obtained by constructing a fragment in which there is a
small segment of unique DNA between the promoters such that a
transcript originating from the left promoter would incorporate this
sequence while RNA made from the right promoter would not.
Transcription products could be electrophoresed, transferred to a
filter, and probed with a sequence specific for the inserted unique
DNA, which would appear only in transcripts originating from the
left, more central, promoter. Another possible explanation for the
unexpected results with the 1331 bp fragment and fragment C is that
there is not enough difference (in bp) between the amounts of
nonspecific target DNA leading to the two promoters. This
possibility is substantiated by the result seen using a fragment
identical to the 1331 bp fragment except that the promoters are
reversed compared to those in the original 1331 bp fragment (see

Figure 9). In this case, the ratio of the short transcripts (from

55

the promoter which RNA polymerase encounters first if it approaches
from downstream, also the one predicted to be preferred if sliding
is occurring) to long is about 2 to 1, but again there is no change
in this ratio between 5X (excess) and 0.1x (sub-saturating) RNA
polymerase concentrations. Another construction may be needed
containing considerably more nonspecific DNA.

In conclusion, the results discussed above are very

titillating, and warrant further study.

 

CONCLUSIONS

The main conclusions which can be drawn from this work are: 1)
that polyacrylamide gel electrophoresis as an analytical tool does
not give artifactual results, and 2) that to the extent which these
experiments can determine, there is no evidence against primary
reptation being the mode by which DNA and DNA-protein complexes
migrate through a polyacrylamide gel.

The in-gel transcription and dissociation rate experiments
address the first point. The transcription data show that a DNA-RNA
polymerase complex is able to function within the gel matrix just as
it does in the absence of the gel. This experiment is an indirect
assay of complex integrity, and is based on the premise that since
transcription is such an intricate process it would be easily
disturbed by any extreme conditions caused by the gel matrix. No
such disturbance is evident. The dissociation rate experiments were
designed to address the question of "caging", or the "artificial
stability" of DNA-protein complexes possibly induced by the gel
matrix. The premise here was that if such a cage exists, then 1)
the dissociation rate of a DNA-protein complex would be slower in a
gel than in solution, and 2) the dissociation rate of a DNA-protein
complex would be dependent on the size of the DNA, with larger
fragments having a slower dissociation rate from a given protein.
The experiments show that the dissociation rates of CAP-Egg DNA
complexes are the same whether in solution or in the gel, and that

the in-gel dissociation rate of CAP-lac DNA complexes is not

56

57

dependent on the size of the DNA fragment, supporting the idea that
caging does not exist under our electrophoresis conditions. These
results taken together lead to the conclusion that the complexes
seen, quantified, and in some cases isolated from polyacrylamide
gels, are the same as those found in solution.

The relative mobility experiments described here do not
contradict the notion that primary reptation is the means of
movement for all complexes tested. This result is not consistent
with our original hypothesis, in which we proposed that complexes of
CAP or RNA polymerase with fragment D would migrate in a
conformation similar to a hairpin, thereby causing the complex to
migrate much slower than the complexes formed using fragment C.
However, it is not possible to reach firm conclusions without
additional data on the pore sizes of polyacrylamide gels.

Several interesting findings were uncovered in the course of
this work. It appears that RNA polymerase bends the DNA in a manner
similar to CAP, although perhaps not as strongly. Also, it was
found that CAP bends the DNA even in the presence of RNA polymerase.
These two facts may speak to the mechanism by which CAP stimulates
transcription at the Egg operon, a subject still under intensive
study.

In addition, the beginnings of a system through which the
question of the mechanism of promoter search by RNA polymerase may
be answered was established. While more work is required to make
the system fully usable, such experiments promise to add to our

knowledge of the processes which work to control gene expression.

58

Polyacrylamide gel electrophoresis, then, remains a valuable
tool for use in separating and isolating DNA, proteins, and
DNA-protein complexes. In addition to qualitative uses, one may use
the technique for analytical purposes with confidence that the
results obtained will be accurate and reflect the true biological

nature of the system under study.

LIST OF REFERENCES

LIST OF REFERENCES
1. Raymond, 8., and L. Weintraub; Science 129 711 (1959).
2. Ornstein, L.; Ann. N.Y. Acad. Sci. EEE 321 (196A).
3. Chrambach, A., and D. Rodbard; Science 11E AAO (1971).
A. Garner, M.M., and A. Revzin; Nucl. Acids Res. 2.30A7 (1981).
5. Fried, M., and D.M. Crothers; Nucl. Acids Res. 2 6505 (1981).
6. Strauss, F., and A. Varshavsky; Cell E1 889 (198A).

7. Carthew, R.W., L.A. Chodosh, and P.A. Sharp; Cell 3; A39
(1985).

8. Schneider, R., I. Gander, U. Muller, R. Mertz, and E.L.
Winnacker; Nucl. Acids Res. E3 1303 (1986).

9. Hendrickson, W., and R. Schleif; Proc. Natl. Acad. Sci. USA 93
3129 (1985).

10. Jones, O.W., and P. Berg; J. Mol. Biol. EE 199 (1966).
11. Garner, M.M., and A. Revzin; Biochemistry El 6032 (1982).
12. Fried, M.G., and D.M. Crothers; Nucl. Acids Res. 1 1A1 (1983).

13. Kolb, A., A. Spassky, C. Chapon, 8. Blazy, and H. Buc; Nucl.
Acids. Res. El 7833 (1983).

1A. Roe, J.H., R.R. Burgess, and M.T. Record, Jr.; J. Mol. Biol.
176 A95 (198A).

15. Shanblatt, S.H., and A. Revzin; Nucl. Acids Res. 1% 5287
(198A).

16. Maxwell, A., and M. Gellert; J. Biol. Chem. 259 1AA72 (198A).

17. Boone, T., and G. Wilcox; Biochim. Biophys. Acta 5A1 528
(1978).

18. Burgess, R.R., and J.J. Jendrisak; Biochemistry 13 A63A (1975).

59

 

19.

20.
21.

22.

23.

2A.
25.
26.

27.
28.
29.
30.
31.

32.

33.

3A.

35.

36.

37.

38.

39.

60
Lowe, P.A., D.A. Hager, and R.R. Burgess; Biochemistry EE 13AA
(1979).
Clewell, D.; J. Bacteriol. 110 667 (1972).
Maxam, A.M., and W. Gilbert; Methods Enzymol. E2 A99 (1980).

Molecular Cloning-A Laboratory Manual (T. Maniatis, E.F.

 

Fritsch, and J. Sambrook) c. 1983 Cold Spring Harbor
Laboratories, Cold Spring Harbor, NY.

Beckman, L.D., and G.D. Frankel; Nucl. Acids Res. E 1727
(1976).

Lerman, L.S., and H.L. Frisch; Biopolymers EE 995 (1982).
Fisher, M.P., and C.W. Dingman; Biochemistry E9 1895 (1971).
Serwer, P., and J.L. Allen; Biochemistry E; 922 (198A).
Lumpkin, 0.J., and 8.8. Zimm; Biopolymers EE 2315 (1982).
Stellwagen, N.C.; Biochemistry EE 6186 (1983).

Diekmann, S., and J.C. Wang; J. Mol. Biol. Egg 1 (1985).

Wu, H., and D.M. Crothers; Nature 292 509 (198A).

The Tools of Biochemistry (T.G. Cooper) 0. 1977 John Wiley &

 

Sons, Inc. New York, NY.

Anderson, W.B., A.B. Schneider, M. Emmer, R.L. Perlman, and I.
Pastan; J. Biol. Chem. 2A6 5929 (1971).

Kumar, S.A., N.S. Murthy, and J.S. Krakow; FEBS Letters 109 121
(1980).

Lowe, P.A., D.A. Hager, and R.R. Burgess; Biochemistry EE 13AA
(1979).

Simpson, R.B. in Promoters: Structure and Function (R.L.
Rodriguez, M.J. Chamberlin, Eds.) pp. 16A 0. 1981 Praeger
Press, New York, NY. ‘

 

Genes (B. Lewin) c. 1983 John Wiley & Sons, Inc. New York, NY.

Shore, D., and R.L. Baldwin; J. Mol. Biol. 170 983 (1983).

Horowitz, D.S., and J.C. Wang; J. Mol. Biol. 173 75 (198A).

Zivanovic, Y., I. Goulet, and A. Prunell; J. Mol. Biol. 192 6A5
(1986). -—-

A0.
A1.

A2.

A3.

uu.

A5.
A6.
A7.

A8.

A9.

50.

51.

61

Hanahan, D.; J. Mol. Biol. 166 557 (1983).
Birnboim, H.C., and J. Doly; Nucl. Acids Res. 1 1513 (1979).

Peterson, M.L., and N.S. Reznikoff; J. Mol. Biol. 185 535
(1985).

Ippen, K., J.H. Miller, J. Scaife, and J. Beckwith; Nature 217
825 (1968).

Silverstone, A.E., R.R. Arditti, and 8. Magasanik; Proc. Natl.
Acad. Sci. USA 99 773 (1970).

Leach, D.R.F., and F. Stahl; Nature (London) 305 AA8 (1983).
Ruchel, R., and M.D. Brager; Anal. Biochem. EE, A15 (1975).
Winter, R.B., and P.H. von Hippel; Biochemistry E9 69A8 (1981).

Winter, R.B., O.G. Berg, and P.H. von Hippel; Biochemistry 29
6961 (1981).

Terry, B.J., W.E. Jack, and P. Modrich; J. Biol. Chem. 260
13130 (1985).

Belintsev, B.N., S.K. Zavriev, and M.F. Shemyakin; Nucl. Acids
Res. E 1391 (1980).

Park, C.S., F.Y.-H. Wu, and C.-W. Wu; J. Biol. Chem. 257 6950
(1982). “‘

"Iiilililililiiliillili