Activated transcription of the glycolytic enzyme genes of Saccharomyces cerevisiae

MISSING IMAGE

Material Information

Title:
Activated transcription of the glycolytic enzyme genes of Saccharomyces cerevisiae the chromatin structures of TP11 and mechanisms of RAP1P mediated activation
Physical Description:
x, 131 leaves : ill. ; 29 cm.
Language:
English
Creator:
Smerage, Jeffrey Beaumont, 1967-
Publication Date:

Subjects

Subjects / Keywords:
Research   ( mesh )
Chromatin -- ultrastructure   ( mesh )
Saccharomyces cerevisiae -- genetics   ( mesh )
Saccharomyces cerevisiae -- physiology   ( mesh )
Trans-Activation (Genetics)   ( mesh )
Gene Expression Regulation   ( mesh )
Promoter Regions (Genetics)   ( mesh )
Transcription Factors   ( mesh )
Department of Molecular Genetics and Microbiology thesis Ph.D   ( mesh )
Dissertations, Academic -- College of Medicine -- Department of Molecular Genetics and Microbiology -- UF   ( mesh )
Genre:
bibliography   ( marcgt )
non-fiction   ( marcgt )

Notes

Thesis:
Thesis (Ph.D.)--University of Florida, 2000.
Bibliography:
Bibliography: leaves 108-129.
General Note:
Typescript.
General Note:
Vita.
Statement of Responsibility:
by Jeffery Beaumont Smerage.

Record Information

Source Institution:
University of Florida
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
aleph - 026074586
oclc - 52305664
System ID:
AA00022160:00001

Full Text











ACTIVATED TRANSCRIPTION OF THE GLYCOLYTIC ENZYME GENES OF
SACCHAROMYCES CEREVSIAE: THE CHROMATIN STRUCTURE OF TPII AND
MECHANISMS OF RAPIP MEDIATED ACTIVATION










By

JEFFREY BEAUMONT SMERAGE


A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE
UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE
REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY

UNIVERSITY OF FLORIDA


2000














ACKNOWLEDGMENTS

The support of a great many people made the completion of this document and its

associated work possible. I would first like to acknowledge and thank my wife, Lucia,

and my son, Alec, for their love, patience, and support through the completion of my

graduate studies. Without their sacrifice and support, this document might never have

been completed. I would also like to recognize my parents, Glen and Barbara Smerage.

Their love and encouragement have not only helped me through graduate school, but

they are also largely responsible for the development of my curiosity and love of

learning.

I would also like to thank and recognize the lab of Dr. Henry Baker, which has

also been an essential component of my graduate experience. The encouragement and

direction provided by Dr. Baker has been an invaluable resource in my growth and

education. Many current and past members of Dr. Baker's lab have also had a significant

impact on the successful completion of this project. These include Dr. Lucia Eisner

Smerage, Ms. Cecilia Lopez, Dr. Carolyn Drazinic. and Dr. Michael Huie.

My graduate committee has also been instrumental in my graduate experience.

The members include Dr. Henry Baker (Chairman), Dr. Daniel Driscoll, Dr. Alfred

Lewin, Dr. Harry Nick, and Dr. Thomas Yang. The time that they have dedicated to

my education and personal development is greatly appreciated. I have enjoyed the

multitude of discussions, committee meetings, as well as data and document reviews,

which they so generously provided. Their advice and insight into the development of an









academic career will be invaluable as I continue to further my own career and personal

research direction. Dr. Donna Duckworth also deserves special recognition for her

participation in the defense of my dissertation and for her advice and friendship over

these years.














TABLE OF CONTENTS

Page

ACKNOW LDGM EN TS ................................................................................................. ii

LIST OF FIGURES .................................................................................................. vi

KEY TO SYM BOLS AND ABBREVIATIONS ........................................................... vii

ABSTRACT .................................................................................................................. ix

INTRODUCTION .......................................................................................................... I

The Basal Transcription Apparatus .......................................................................... 3
Enhancer Structure and Function ............................................................................ 9
Chromatin and Nucleosomes as Regulators of Transcription ................................. 12
Transcription and Saccharomyces cerevisiae ........................................................ 26
Summ ary ................................................................................................................... 44

M ATERIALS AND M ETHODS ............................................................................... 46

Strains ....................................................................................................................... 46
M edia and Growth Conditions .............................................................................. 46
Chrom atin M apping by Nuclease Sensitivity ........................................................ 46
In vivo Footprinting .............................................................................................. 52
DNA Band-Shift Analysis ..................................................................................... 57

RESULTS ..................................................................................................................... 67

Characterization of the Chromatin Structure at TPIJ in Saccharomyces cerevisiae ... 67
Raplp Binding is Required for Binding of Gcrlp at TPIJ in vivo ......................... 81
Raplp Facilitates the Binding of Gcrlp in vitro .................................................... 85










DISCUSSION ............................................................................................................... 96

Gcrlp Binding Depends on the Presence of Rap Ip ............................................... 96
The N-terminus and C-terminus of Rap lp Both Facilitate the Gcrlp Binding
Specificity ............................................................................................................. 98
Raplp Mediated Gcrlp Binding Is Distance Dependent .......................................... 100
TPI1 Is Located W ithin a Permissive Nucleosome Environment .............................. 101
TPI1 Chromatin Structure Is Not Affected By Mutations in GCRI, GCR2, or
Gal ................................................................................................................. 104
Chromatin Structure At Surrounding Open Reading Frames .................................... 105
M odel of Glycolytic Enzyme UAS Function ........................................................... 105

REFERENCES ........................................................................................................... 108

BIOGRAPHICAL SKETCH ....................................................................................... 130














LIST OF FIGURES


Figure Page

1.1 The glycolytic pathway ................................................................... 28

1.2 Glycolytic UAS elements ............................................................... 32

1.3 Glycolytic UAS elements ............................................................... 34

2.1 Design of RapIp truncation mutants ............................................... 59

2.2 RapIp truncation vectors ................................................................. 60

2.3 Phosphorimager quantification ......................................................... 66

3.1 MNase sensitivity pattern upstream from BsiHKAI ........................ 70

3.2 MNase sensitivity pattern upstream from KpnI ................................ 72

3.3 DNaseI sensitivity pattern upstream from KpnI ............................... 74

3.4 MNase sensitivity pattern downstream from KpnI .......................... 76

3.5 DNasel sensitivity pattern downstream from KpnI ........................... 78

3.6 In vivo footprint analysis at the UASTP ........................................... 83

3.7 Bandshift of Raplp and Gcrlp using native spacing ....................... 89

3.8 Bandshift of RapIp and Gcrlp at +5 basepairs separation ............... 91

3.9 Summary of bandshift data for native and +5 spacing ..................... 94

4.1 Summary of the TPI1 chromatin and promoter structures ................... 103














KEY TO SYMBOLS AND ABBREVIATIONS


A adenine
A600 Spectrophotometric absorbance at 600nm
BMV Brome Mosaic Virus
BSA bovine serum albumin
CaC12 calcium chloride
ChiP chromatin immunoprecipitation
CTD carboxy-terminal domain
D dextrose/glucose
DMS dimethylsulfate
DNA deoxyribonucleic acid
DTT dithiothreitol
E. coli Escherichia coli
EDTA ethylenedinitrilotetraacetic acid
EtBr ethidium bromide
G guanine
GCR1 gene encoding the yeast Glycolysis Regulation protein
Gcrlp Glycolysis Regulation protein
GCR2 gene encoding the yeast Glycolysis Regulation protein number two
Gly/Lac glycerol lactate
GTF general transcription factor
GuHCI guanidine hydrochloride
HCl hydrochloric acid
HDA high-density oligonucleotide array
HEPES 4-(2-hydroxyethyl)- 1 -piperazineethanesulfonic acid
IPTG isopropylthiogalactoside
kb kilobase
KCI potassium chloride
KPO4 potassium phosphate
lacZ E. coli gene encoding B-galactosidase
LB Luria broth
LBamp Luria broth with ampicillin
LTR long terminal repeat
M63 minimal media 63 (media)
MBP maltose-binding protein
Mda megadalton
mg milligram
mg/ml milligram per milliliter
MgC12 magnesium chloride
ml milliliter









MMTV
MNase
MOPS
1t9
jig/ml

NaAc
NaCI
ng
NaOH
NaPO4
NP-40
ONPG
ORF
PBS
PCR
Pol II
poly(dI-dC)
RAP1
Raplp
REBI
Reblp
raplts
SAGE
S. cerevisiae
SDS
SDS-PAGE
TAF
TBE
TBP
TCA
TE
TPI 1
Tris-HC1
UAS
URA3
UV
X-gal
YP
YPD
YP-Gly/Lac


mouse mammary tumor virus
micrococcal nuclease
3-[N-Morpholino]propanesulfonic acid
microgram
micrograms per milliliter
microliter
sodium acetate
sodium chloride
nanogram
sodium hydroxide
sodium phosphate
Nonidet P-40
o-nitrophenyl galactoside
open reading frame
phosphate buffered saline
polymerase chain reaction
RNA polymerase II
polydeoxyinosinic-deoxycytidylic acid
gene encoding the yeast Repressor/Activator protein
Repressor/Activator protein
gene encoding the yeast Ribosomal Enhancer Binding protein
Ribosomal Enhancer Binding protein
temperature-sensitive allele of RAP 1
serial analysis of gene expression
Saccharomyces cerevisiae
sodiumdodecylsulfate
sodiumdodecylsulfate polyacrylamide gel electrophoresis
TATA binding protein associated factor
Tris, borate, EDTA buffer
TATA binding protein
trichloroacetic acid
Tris, EDTA buffer
gene encoding triosphosphate isomerase
tris(hydroxymethyl)aminomethane hydrochloride
upstream activating sequence
gene encoding ornithine transcarbamylase
ultraviolet
5-bromo-4-chloro-3-indolyl-B-D-galactose
yeast peptone (media)
yeast peptone dextrose/glucose (media)
yeast peptone glycerol lactate (media)














Abstract of Dissertation Presented to the Graduate School
of the University of Florida in Partial Fulfillment of the
Requirements for the Degree of Doctor of Philosophy

ACTIVATED TRANSCRIPTION OF THE GLYCOLYTIC ENZYME GENES OF
SACCHAROMYCES CEREVISIAE: THE CHROMATIN STRUCTURE OF TPIJ AND
MECHANISMS OF RAPIP MEDIATED ACTIVATION

By

Jeffrey Beaumont Smerage

December 2000

Chairman: Hemy V. Baker, Ph.D.
Department: Molecular Genetics and Microbiology

Regulation of transcription is a major pathway by which cells maintain a stable

intracellular environment and react to the extracellular environment. This study

addresses mechanisms by which a limited number of transcription factors can regulate

large and complex transcription systems. The glycolytic enzyme genes of

Saccharomyces cerevisiae were used as a model for the study of high-level activated

transcription. High-level expression of these genes is the result of combinatorial

interactions within a conserved set of transcription factors which bind at the Upstream

Activating Sequences (UASs) of each gene. These include the transcription factors

Abflp, Reblp, Raplp, and Gcrlp. In vivo treatment with micrococcal nuclease and

DNase I were used to demonstrate a region of nuclease hypersensitivity over the

promoter element of TPIJ and a series of statistically positioned nucleosomes running

upstream from the hypersensitive region. Nuclease hypersensitive regions are









characteristic of genes which are poised for activated transcription, and statistically

positioned nucleosomes are associated with "boundary" elements which can function to

prevent the formation of repressing nucleosomes over a UAS element. In vivo footprint

analysis of the TPI1 locus was used to demonstrate that binding of Gcrlp at its two sites

in the UAS requires concomitant binding of Raplp at its adjacent binding site. This is a

critical function of Raplp because in the absence of bound Gcrlp transcription is reduced

by 99%. Finally, DNA electromobility gel-shift analysis was used to define domains

within RapIp which are important for the facilitation of Gcrlp binding. The strongest

facilitation was seen in Rapl p species which retain both the deoxyribonucleic acid

(DNA)-binding domain and the asymmetric bending domain, although the asymmetric

bending domain was not absolutely required. Facilitation was observed when either the

amino-terminus or the carboxy-terminus was eliminated but not when both were

eliminated, leaving just the core DNA-binding domain. In summary, this work

demonstrates that the promoter of the TPI1 gene is located within a permissive chromatin

structure, the transcriptional activator Gcr 1 p is directed to its binding site by Rap lp. and

several domains within RapIp appear to be important for the facilitation of Gcr I p

binding.














INTRODUCTION


Molecular biology has become an integral part of modern life. It has changed the

foods we eat. It has revolutionized the way many vaccines and medicines are made. It is

now an important factor in the diagnosis and risk stratification of many illnesses. As a

result, the biotechnology industry has become a major sector of our economy. Many of

these applications have been made possible by our knowledge of gene structure and

transcriptional regulation.

Since the realization that DNA is the repository of genetic information (Avery et

al. 1944; Hershey and Chase 1952), an enormous amount of work has been done

investigating how that information is mobilized. Appropriate and coordinated expression

of genes is important for many reasons, and there are many examples of how either lack

of gene expression or inappropriately activated transcription can result in failure of

cellular processes. Some of the most-studied mammalian mutations lead to inappropriate

expression of oncogenes or lack of expression of tumor suppressor genes resulting in

neoplastic disease. In microbiology, mutagenic analysis has been used to great advantage

in clarifying many details of transcriptional regulation and other cellular functions.

A basic question in the field of transcriptional regulation has asked how a cell can

regulate the complex patterns of gene expression with the specificity required for

successful function, growth, and reproduction. It appears that part of this specificity is

provided by DNA sequences that act as regulatory elements. These sequences function









primarily as protein binding sites, and they can be generally divided into two groups.

First, there is a proximal promoter that serves as a binding area for the transcription

apparatus often referred to as the basal transcription complex (Hernandez 1993; Malik

and Roeder 2000). Second, there are more distal regulatory elements that act as binding

sites for transactivating regulatory proteins (Tjian and Maniatis 1994; Malik and Roeder

2000). In general, these distal elements contain binding sites for multiple proteins. This

study utilizes the glycolytic enzyme genes as a model for the study of these regulatory

elements and their DNA-binding proteins. The evidence presented here demonstrates

that the binding of the transcriptional activator Gcrlp requires the simultaneous binding

of a second DNA-binding protein, Rap lp. It also investigates the domains within Rap I p

that may be required for this interaction. They represent an example of how specific and

appropriate gene expression appears to depend on the combinatorial interaction between

multiple transcriptional activators. An analysis of the chromatin structure surrounding

the glycolytic enzyme gene TPI1 is also presented. Chromatin in this region is found to

have areas of nuclease hypersensitivity and nucleosome phasing, both of which are

consistent with a transcriptionally active gene. These data provide a framework upon

which further investigation can be made into the formation and maintenance of active

chromatin structures. It is through complex interactions such as those seen at TPIJ and

the other glycolytic enzyme genes that a relatively small number of transcriptional

activators can regulate the activity of a large number of individual genes. This also

provides a mechanism by which cells can coordinate the expression of groups of genes in

response to environmental stimuli, intercellular signals, metabolic needs, cell cycle

functions, and cell type functions.









The Basal Transcription Apparatus

The basal transcription apparatus has traditionally been defined as a large multi-

protein complex that consists of the RNA Polymerase II (Pol II), the TATA Binding

Protein (TBP) as a component of TFIID, and additional General Transcription Factors

(GTFs) (Conaway and Conaway 1993; Orphanides et al. 1996). As noted above, this

complex acts at proximal DNA structures including the TATA box and the transcription

initiation site (Avery et al. 1944; Struhl 1989). In addition, a protein complex termed

mediator has been isolated in association with Pol II (Kim et al. 1994; Koleske and

Young 1994). The basal complex is capable of directing transcription of in vitro systems,

and this in vitro activity has been defined as basal transcription. The basal transcription

apparatus, as the target of the activating factors, remains an important component in the

understanding of the in vivo activation of transcription.

RNA Polymerase II and the General Transcription Factors

The heart of the basal apparatus is Pol II. This is one of three eukaryotic RNA

polymerase complexes, and it is responsible for the transcription of most protein-

encoding genes (Sawadogo 1990). Our knowledge of the eukaryotic Pol II has gone

through significant conceptual evolution in the past thirty years. This enzyme was first

isolated as a chromatographic fraction capable of in vitro template-dependent RNA

synthesis (Roeder and Rutter 1969). It has since been isolated from many species

including human (Freund and McGuire 1986), yeast (Sentenac 1985), fruitfly (Greenleaf

and Bautz 1975), mouse (Schwartz and Roeder 1975), frog (Engelke et al. 1983) as well

as others (reviewed in Young 1991). The "core" enzyme is a multi-protein complex that

consists of 10-+2 subunits depending on the species (Sawadogo 1990; Young 1991 ).

They are highly conserved across species (Roeder 1976; Allison et al. 1985; Sentenac









1985; Young 1991). However, this "core" polymerase is unable to direct selective

transcription in vitro (Roeder 1976; Weil et al. 1979).

Selective transcription initiation was achieved by the addition of chromatographic

fractions from crude cellular extracts (Matsui et al. 1980). The components of these

fractions became known as the General Transcription Factors (GTFs) (Sawadogo 1990;

Conaway and Conaway 1993; Orphanides et al. 1996). They have been isolated and

purified from many types of cells including budding yeast, HeLa cells, rat liver cells, and

fruitfly cells (Zawel and Reinberg 1993). The GTFs are represented by six fractions

named TFIIA, TFIIB, TFIID, TFIIE, TFIIF, and TFIIH. As a group, they represent a

collection of approximately 30 proteins (Orphanides et al. 1996). Extensive genetic and

biochemical investigation has defined proposed functions for many of the GTFs. These

include recognition and binding to the TATA box, recruitment of the Pol II enzyme,

melting of the promoter DNA, and initiation of Pol II transcription (Orphanides et al.

1996). However, their true functions within the Pol II "holoenzyme" are still

incompletely understood.

The TFIID subunit has been extensively studied and shown to be the only GTF

with strong DNA-binding activity and the GTF required for sequence specific positioning

of the transcription apparatus (Hahn et al. 1989). TFIID is a multi-protein complex, and

the sequence specificity is provided by a subunit called the TATA-binding Protein (TBP).

The remaining TFIID subunits are referred to as TBP-associated factors (TAFs).

Isolation of TBP was made possible by its high degree of cross-species conservation. It

was initially isolated as a yeast protein that could substitute for TFIID in an in vitro

mammalian transcription system (Buratowski et al. 1988; Cavallini et al. 1988). TBP has









been called the universal transcription factor by many authors (Hernandez 1993; Struhl

1995; Orphanides et al. 1996). It is the most conserved eukaryotic transcription factor,

and it is required for the transcription by all three of the eukaryotic RNA polymerases

(Pol 1, II, and III) (Hernandez 1993; Orphanides et al. 1996). The C-terminal domain

displays over 80% sequence identity across numerous species (Hernandez 1993). TBP

has also been shown to bind with high affinity and specificity to its consensus site

(TATAa/tAa/t) (Hahn et al. 1989). TBP is capable of directing selective but not activated

transcription in vitro (Hernandez 1993). The function of the non-conserved N-terminus

is not yet known.

The structure of TBP and the mechanism by which it binds to DNA may play an

important regulatory role in the ability of the transcription apparatus to bind to the

promoter. The crystal structure of TBP has been solved for both the protein alone as well

as bound to the TATA DNA element (Nikolov et al. 1992; Chasman et al. 1993; Kim et

al. 1993a; Kim et al. 1993b). TBP is an intramolecular dimer with near twofold

symmetry. It binds in the minor groove of the DNA double helix and displays a preferred

orientation. It also induces a dramatic 900 bend in the DNA with resulting unwinding

and superhelical twist at the end of the TATA sequence. This bend induced by TBP has

been proposed to function both to prevent nucleosome formation and to brine distant

regulation elements together through a looping mechanism (Orphanides et al. 1996).

In addition to the TBP, TFIID contains the TAFs, which may contribute further to

the development of an active promoter conformation. TBP creates a footprint over the

TATA box, and the complete TFIID extends this footprint beyond the site of transcription

initiation (Nakajima et al. 1988; Purnell et al. 1994; Sypes and Gilmour 1994). TFIID









may also play an important role in local chromatin structure. Several of the TAFs have

sequence and structural homology to histones (Orphanides et al. 1996)). More

specifically, yeast TAF61, TAF17, and TAF60 have homology to histones H2B, H3, and

H4 respectively. In Drosophila, a crystal structure of the H3- and H4-like TAFs has

revealed a complex similar to the histone (H3-H4)2 heterotetramer (Xie et al. 1996). In

vitro, prebound TFIID can prevent the formation of nucleosomes (Meisterernst et al.

1990). Conversely, it has also been shown that preexisting nucleosomes can prevent

TFIID binding in vitro (Han and Grunstein 1988). As noted above, the TBP causes the

DNA to bend away from the minor groove, resulting in a widened minor groove.

Nucleosomes on the other hand cause the DNA to bend in the opposite direction resulting

in compression of the minor groove (Van Holde 1989). The exact mechanism by which

TFIID affects chromatin is not yet clear. It is possible that TFIID initiates actual

nucleosome disruption, but it is also possible that TFIID simply maintains a structure that

was opened by other mechanisms. In either case, it is unlikely that both TFIID and

nucleosomes can be bound to the promoter simultaneously because of their incompatible

bending effects on the DNA strand.

The TAF components of TFIID have been of significant research interest because

initial in vitro studies demonstrated that they were required for activator-dependent

transcription initiation (Hernandez 1993). There are also many transcription activators

that have been shown to have protein-protein interactions with TBP as well as the TAFs

(Burley and Roeder 1996). However, initial attempts to demonstrate this dependence in

vivo were unsuccessful. Several authors showed that when individual TAFs are mutated

there are no global effects on transcription (Moqtaderi et al. 1996; Walker et al. 1996).









More recent data has shown that selected TAFs are generally although not universally

required for transcriptional activation. In particular, the histone-like TAFs appear to be

important for transcriptional activation (Michel et al. 1998; Moqtaderi et al. 1998;

Natarajan et al. 1998; Lee et al. 2000).

TAF17 is an example of a histone-like TAF that appears to be generally but not

universally required for transcriptional activation. Moqtaderi et al used a copper-

inducible, "double shutoff' construct to produce conditional alleles in TAF 17 (histone H3

homolog), TAF40, and TAF67 (Moqtaderi et al. 1998). This conditional system was

designed such that the addition of copper repressed the expression of each experimental

TAF gene and activated the expression of the ubiquitin degradation pathway. The

experimental TAF gene contained an N-terminal recognition signal that resulted in rapid

ubiquitin-directed depletion of the TAF. Depletion of TAF 17 resulted in the loss of

mRNA species from a broad array of genes, indicating a loss of general transcriptional

activity. This was not universal, however, as even eight hours after the addition of

copper they were still able to induce wild-type Aceip-dependent expression of cupl, and

induce wild-type heat shock-dependent expression of hsp104 and ssa4. The non-histone-

like TAFs, TAF40 and TAF67, were required only for transcriptional activation from

non-consensus TATA box elements. In a prior study, it was also found that two other

non-histone-like TAFs, TAF 145 and TAF 19, were required for transcriptional activation

only if the promoter contained a non-consensus TATA box (Moqtaderi et al. 1996).

The use of high-density oligonucleotide arrays (HDAs) has shown similar results.

Lee, et al (2000) used temperature sensitive alleles of several TAFs to evaluate their

effects on genome-wide expression. They studied seven TAFs, and each individual TAF









was required for transcriptional activation of 2-67% of yeast genes. TAF 17 demonstrated

the widest effects, being required for the expression of 67% of yeast genes. Recent data

has shown that there is a group of TAFs that are contained in a non-TFIID complex

known as SAGA (Grant et al. 1998). Five of the seven TAFs studied by Lee, et al (2000)

are found in both complexes. They evaluated the effects on genome-wide expression by

the depletion of TFIID specific TAFs, SAGA specific proteins, and TAFs that are shared

between the two complexes. They found that the greatest effects on transcription were

seen with the TAFs shared by both complexes. In total, the five shared TAFs were

required for the expression of 70% of the yeast genome. This resulted in the hypothesis

that TFIID and SAGA may perform overlapping functions. Loss of TFIID or SAGA in

isolation only affects transcription from subsets of yeast genes, but loss of both effects

the expression of most yeast genes.

Mediator and the Pol I Holoenzme

Signal transduction from activator elements to the polymerase complex now

appears to be communicated through the mediator complex. The combination of the

mediator and the traditional basal apparatus are now called holoenzyme. The term

mediator was first used to describe a cellular extract fraction required for activated

transcription in vitro using a reaction composed of otherwise purified components

(Flanagan et al. 1991). The mediator has been isolated and characterized by both genetic

(Nonet and Young 1989; Thompson et al. 1993) and biochemical means (Kim et al. 1994;

Koleske and Young 1994). Most of its components have now been identified. It is

composed of approximately 20 proteins, including Gall l p, Sug I p, four Srb proteins

(Kim et al. 1994), Sin4p, Rgrl p (Li et al. 1995), Rox3p (Gustafsson et al. 1997), and









seven Med proteins (Myers et al. 1998). Mediator was discovered in yeast; however

there is evidence that a similar structure is present in mammals (Jiang et al. 1998; Sun et

al. 1998; Gu et al. 1999)

Mediator appears to play a significant role in transcriptional activation. It

stimulates basal transcription 10-fold when added to in vitro transcription reactions

composed of purified components. An additional 30-fold response is seen when activator

proteins are included in the reaction (Kim et al. 1994). This represents a 300-fold

increase in transcriptional activation when compared to the basal apparatus in the absence

of the mediator complex.

Mediator appears to effect its action through the C-terminal domain (CTD) of Pol

II. The CTD is a highly conserved structure that contains tandem repeats of the

consensus amino acid sequence Tyr-Ser-Pro-Thr-Ser-Pro-Ser (Corden 1990). Like many

other components of the transcriptional apparatus, the Pol II CTD is highly conserved

across species, but it does vary in length from species to species. The murine CTD

contains 52 repeats, whereas the yeast CTD contains 26-27 repeats (Allison et al. 1985;

Corden et al. 1985; Nonet et al. 1987). The transition from the polymerase pre-initiation

complex to the actively transcribing polymerase complex correlates with the

phosphorylation state of the CTD (Dahmus 1996). Interestingly, mediator stimulates

activator-dependent phosphorylation of the CTD 40-fold in vitro (Kim et al. 1994).

Enhancer Structure and Function

The activation of the holoenzyme is achieved by the presence of more distal

elements known as enhancers. Enhancers have been defined as DNA elements that

activate transcription when placed at variable distances up- or down-stream from the

proximal promoter element, and their activity is independent of orientation (Struhl 1989;









Tjian and Maniatis 1994). They were also found to be responsive to cell-type specific

signals. It is now known that these enhancer elements represent binding sites for

transcriptional activator proteins (McKnight and Tjian 1986; Guarente 1987).

The first use of the term enhancer was in relation to the SV-40 early region

promoter (Banerji et al. 1981; Benoist and Chambon 1981; Fromm and Berg 1982).

Initial experiments in which segments of the promoter region were deleted identified two

72-base pair repeats that were essential for gene expression in vivo (Benoist and

Chambon 1981). This enhancing activity was further investigated using a construct in

which a variety of SV40 promoter region mutants were placed in front of a rabbit

hemoglobin 01 reporter gene (Banerji et al. 1981). They found that the presence of the

SV40 sequence resulted in a 200-fold increase in expression of the reporter RNA product.

They also found that the SV40 sequence was capable of enhancing transcription when

placed in either orientation and when placed in a variety of positions relative to the site of

transcription initiation. This included placement as far upstream as 1400 base pairs and

as far downstream as 3300 base pairs. Since the discovery of the SV40 enhancing

activity, similar cis-acting enhancer sequences have been found in association with

essentially all genes and in all organisms studied (Guarente 1988). In yeast, these

enhancer elements have been named Upstream Activating Sequences (UASs). A

hallmark property of enhancing elements is that they contain binding sites for a wide

variety of transcription activator proteins (McKnight and Tjian 1986; Guarente 1987).

One of the interesting properties of transcriptional activators is that they have a

modular structure. There are many examples of proteins that contain distinct DNA-

binding domains and activation domains that have been identified by mutation analysis.









For example, the yeast activator Gal4p contains a DNA-binding domain located in the

amino terminus and an activation domain located in the carboxy end of the protein (Brent

and Ptashne 1985). For the activator Gcn4p, the locations are different. The DNA-

binding domain is at the carboxy terminus, and the activation domain is in the middle of

the protein (Hope and Struhl 1986; Hope et al. 1988). In fact, the DNA-binding and the

activation functions can be physically separated. This has been demonstrated by the

creation of hybrid proteins in which the DNA-binding domain of one activator has been

combined with the activation domain of another protein. When the activation domains of

either Gal4p or Gcn4p have been fused with the LexA DNA-binding domain they are

capable of activating transcription from a reporter gene in which a LexA operator has

been placed upstream (Brent and Ptashne 1985; Hope and Struhl 1986). LexA is a DNA-

binding protein that normally represses transcription in the prokaryote E. coli (Brent and

Ptashne 1985).

The function of these transcriptional activators is dependent on the ability of the

DNA-binding domain to bring the activation domain to the enhancing element of the

promoter. If the activator protein is mutated such that the DNA-binding domain is

absent, then the activation domain can no longer activate transcription. This has been

demonstrated well in the "two-hybrid system" (Fields and Song 1989; Chien et al. 1991),

which has been put to dramatic use in identifying protein-protein interactions. In this

experimental system the DNA-binding domain and the activation domain of the Gal4

activator protein are physically separated. The DNA binding domain of Gal4 (amino

acids 1-147) is fused to an unrelated protein. Likewise, the activation domain of Gal4

(amino acids 768-881) is fused to a second unrelated protein. If these two fused proteins









demonstrate protein-protein binding, then this binding tethers the activation domain to the

DNA-binding domain that is bound to the promoter of the reporter gene. If the two

proteins do not interact then this tethering action does not occur, the activation domain is

not brought to the promoter, and transcriptional activation does not occur.

Several families of activator domains have been identified. In mammalian

systems these include acidic domains, proline-rich domains, and glutamine-rich domains

(Struhl 1995). In yeast, only acidic activation domains have been identified (Struhl

1995). Glutamine-rich and proline-rich elements do not appear to function in yeast

(Kunzler et al. 1994; Ponticelli et al. 1995). However, not all yeast activators contain

acidic domains so it is assumed that there are other yet undiscovered motifs in yeast

(Struhl 1995). The mechanisms by which activation domains function are not clearly

known, but there is evidence for several possibilities. These include recruitment of the

Pol II complex, induction of confonnational changes in the Pol II complex, alterations in

chromatin structure, and effects on transcription elongation (Keaveney and Struhl 1998).

However, these do not represent mutually exclusive mechanisms.

Chromatin and Nucleosomes as Regulators of Transcription

In vivo essentially the entire DNA is contained within chromatin. This packaging

system imparts many structural constraints and alterations to the DNA strand as well as to

the individual promoter elements. The nucleosomes and chromatin were initially thought

to be static structures, and they were thought to repress transcription simply by the steric

hindrance of binding by transcriptional activators and RNA polymerase (Thoma et al.

1979). More recently, however, chromatin and nucleosome structure have been found to

be much more dynamic, and they may even play a role in promoter specific









transcriptional regulation (Kingston et al. 1996; Workman and Kingston 1998; Wolffe

and Guschin 2000).

The Nucleosome Core

Chromatin has multiple levels of organization, but at the heart of this structure lies

the nucleosome. This protein complex is an octomer composed of four histones, H2A,

H2B, H3. and H4. Assembly of the nucleosome core involves the formation of H3-H4

heterodimers that then combine to form a tetramer through dimerization of the H3

subunits (Eickbush and Moudrianakis 1978). This (H3-H4)2 tetramer is bound on both

sides by H2A-H2B heterodimers to form the final nucleosome octomer (Arents et al.

1991). The octomer is then wrapped by DNA to form the nucleosome (Hayes et al. 1990;

Hayes et al. 1991).

The histones are another example of proteins that are highly conserved across

species (Wells and McBride 1989, Wells and Brown 1991). The most conserved histones

are H3 and H4 (Wells and McBride 1989; Wells and Brown 1991). The human H3 and

H4 histones are virtually identical (either identical or one amino acid difference) to the

corresponding histone proteins from cow, mouse, frog, trout, and sea urchin. When

mammals are compared to more distantly related metazoa such as wheat, pea, and maize,

the H4 histones only differ at two amino acids. The 13 and 114 histones of S. cerevisiae

vary from human by only 15 and 8 amino acids, respectively. The H3 and H4 histone

lengths are also virtually identical in all species. Histone H4 is always 102 amino acids

in length, and with only one exception, histone H3 is always 135 amino acids. The

histones also have structural conservation, and they have two major structural domains, a

globular core that contains a tertiary structure known as a "histone fold" and an









unstructured N-terminal tail (Arents et al. 1991). The unstructured tail contains multiple

lysine residues that serve as sites of acetylation.

The histones H2A and H2B show more sequence variability, but they continue to

share the tertiary structure of the histone fold and the unstructured tail regions (Wells and

McBride 1989; Wells and Brown 1991). There is some evidence that the H2A-H2B

heterodimers are destabilized when active chromatin structures are created. This has led

to the hypothesis that the variations in H2A and H2B may represent regulatory

differences between different organisms.

Higher Order Chromatin Structure

One of the first views of chromatin revealed a 10-nm filament that has been

described as "beads on a string" (Thoma et al. 1979; De Murcia and Koller 1981). This

structure is characterized by individual nucleosomes separated by lengths of "linker"

DNA. This was consistent with nuclease digestion data in which showed partially

digested chromatin formed a ladder of DNA fragments whose lengths were multiples of a

basic size unit (Lohr et al. 1977a; Noll and Kornberg 1977). More extensive nuclease

digestion demonstrated a 166 bp unit termed a chromatosome (Simpson 1978). The

chromatosome contained DNA supported by both the core histone octomer and one

histone HI. With further digestion, a 145 bp unit was produced that represented the

removal of the linker DNA, leaving only the core DNA (Simpson 1978). The beads-on-

a-string structure is not thought to represent the predominant form of chromatin because

it was seen predominantly in preparations using non-physiologic cation concentrations

(Schwarz and Hansen 1994).









In higher eukaryotes the majority of chromatin is believed to exist in vivo as a

higher order structure, which is 30-nm in diameter (Davies et al. 1974; Paranjape et al.

1994; Wolffe 1998). This 30-nm fiber, as well as even higher order structures, is

virtually devoid of transcriptional activity. As a result, it is believed to be involved with

the epigenetic control of transcription. It is unclear whether a similar 30-nm structure

exists in budding yeast. Several studies utilizing electron microscopy have shown that

yeast chromatin can form a 30-nm filament in vitro (Rattner et al. 1982; Lowary and

Widom 1989). However, yeast do not have cytogenetically detectable heterochromatin

(Rattner and Hamkalo 1981; Bazett-Jones dt al. 1988: Guacci et al. 1993; Elgin 1995).

Early experiments indicated that HI linker histones were required for formation

of the 30-nm fiber (Thoma et al. 1979; Widom 1989). More recent results show that the

30-nm fiber can be formed in the absence of histone H I if appropriate cation

concentrations are present (Schwarz and Hansen 1994). It is now felt that the H I histone

serves to stabilize the 30-nm fiber via a charge neutralization mechanism, in which the

highly basic tail domains neutralize the charge on the polyanionic backbone of the DNA

strand (Clark and Kimuura 1990; Carruthers et al. 1998; Wolffe 1998).

Histone HI is highly variable across species (Wells and McBride 1989; Wells and

Brown 1991). In fact, it is controversial whether S cerevisiae even possesses an HI

histone. For a long time, it has been believed that S. cerevisiae does not contain a H I

histone. Many attempts to immunoprecipitate an H 1 protein using antibodies against H I

proteins from higher eukaryotes were unsuccessful (Escher and Schaffner 1997).

Additionally, the length of the yeast linker regions is very short, given that the DNA in

the chromatosome is approximately 160 base pairs (Lohr et al. 1977a; Lohr et al. 1977b).









This has been felt to be incompatible with the presence of a linker protein. Recently,

however, the gene HHOJ was identified using data from the S. cerevisiae genome

project. This sequence encodes a protein with a globular domain that has 36% identity to

human HI histone (Landsman 1996). The significance of this protein is currently

uncertain since neither deletion nor over-expression of HHO1 appears to affect activation

or repression of transcription (Escher and Schaffner 1997; Patterton et al. 1998).

Nuclease Hypersensitivity and Statistical Positioning of Nucleosomes

An early observation was that regions of nuclease hypersensitivity were

associated with loci of transcriptional activity, and conceptually they have been thought

of as "open windows" that provide transcription factors access to regulatory DNA

sequences (Wu et al. 1979a; Wu et al. 1979b; Gross 1988). Nuclease hypersensitivity

was first noted in work with the SV40 virus (Scott and Wigmore 1978; Varshavsky et al.

1978). In these studies, DNase I was used to demonstrate that the SV40 chromatin

contained a region of hypersensitivity located near the origin of viral replication. Since

this early discovery, nuclease hypersensitivity sites have been found to be ubiquitous in

eukaryotes, and they are associated with functional elements within the DNA sequence

including centromeres, silencers, recombination sequences, origins of replication,

enhanceriUAS elements, promoter elements, transcriptional terminators, locus control

regions, and A-elements (Gross 1988).

The chromatin structure at glycolytic genes is not well studied. Work at TDH3

has demonstrated a hypersensitive area over the promoter (Gross 1988). Additionally, it

was demonstrated that mutant strains in which the transcriptional activator Gcr 1 p is









deleted are associated with the deposition of two positioned nucleosomes over the

proximal promoter element (Gross 1988).

The PHO5 promoter provides another example of an inducible hypersensitive site

(Schmid et al. 1992). PHO5 encodes an acid phosphatase, and its transcription is induced

by phosphate starvation. The PHO5 promoter has two Pho4p-binding sites as well as a

Pho2p-binding site. When the PHO5 gene is inactive, the UAS region is incorporated

into at least six positioned nucleosomes. The first Pho4p-binding site is positioned

between two nucleosomes in a very small area of DNase I hypersensitivity (Almer and

Horz 1986). The second Pho4p-binding site, along with the Pho2-binding site are

positioned within a nucleosome and are notaccessible for binding (Venter et al. 1994).

Upon phosphate starvation, Pho4p binds to the binding site between nucleosomes

followed by the loss of four positioned nucleosomes as evidenced by the appearance of a

new area of nuclease hypersensitivity (Almer and Horz 1986; Almer et al. 1986). This

area of hypersensitivity included the binding sites for Pho4p and Pho2p. This

nucleosome disruption occurs in the absence of DNA replication (Schmid et al. 1992). It

is during this hypersensitive state that Pho4p and Pho2p are able to bind to their sites,

which were previously incorporated into a nucleosome (Venter et al. 1994).

A second phenomenon observed at transcriptional promoters is nucleosome

positioning. Nucleosomes can be positioned by a variety of mechanisms including DNA-

binding proteins, DNA sequence, and DNA bending (Simpson 1978; Thoma 1992).

Positioned nucleosomes then result in a phenomenon known as statistical positioning

(Fedor et al. 1988), which occurs because the positioned nucleosome acts as a boundary

element. This boundary forces adjacent nucleosomes to assume a specific rotational and









translational positions. Translational position is the location of the histone octomer along

the linear DNA strand. Translational positioning was seen in the PHO5 promoter, as

discussed above, where nucleosomes were positioned along the length of the DNA

resulting in Pho4p access to the first binding site. Rotational position relates to the

orientation of the DNA on the histone octomer. The DNA helix is a three dimensional

structure, and any surface such as the major or minor grooves can be either facing toward

the histone core or it can be facing out away from the core. Thus, if a protein binds a

particular surface of the DNA, this binding site can either be obscured or accessible based

on the rotational positioning.

Nucleosome positioning can have both positive and negative effects on

transcription. When promoter sites are obscured by the placement of a nucleosome, this

can prevent binding and as a result inhibit transcriptional activation. The positioning of

nucleosomes over the Pho4p- and Pho2p-binding sites at the PH05 locus described

above exemplify how transcription factor binding can be inhibited by the presence of a

nucleosome. As noted above, this association was also seen at the TDH3 gene at which

the gcrl deletion mutation was associated with the deposition of positioned nucleosomes

over the proximal promoter (Gross 1988). In the gcrl deletion mutant, transcription from

this promoter is severely reduced (Gross 1988). Alternatively, the mouse mammary

tumor virus (MMTV) provides an example where the presence and exact positioning of a

nucleosome is required for the activation of transcription. The long terminal repeat

(LTR) of the MMTV promoter contains binding sites for multiple protein factors,

including the glucocorticoid receptor, OTF, and NFL. In the absence of glucocorticoid

hormone, the promoter is covered by six positioned nucleosomes (Zaret and Yamamoto









1984; Richard-Foy and Hager 1987; Pina et al. 1990). When the glucocorticoid receptor

binds to its site, both the rotational and translational position of the second or "B"

nucleosome changes. The rotational position allows the NF 1 binding site to become

accessible on the surface of the nucleosome (Lee and Archer 1994; Truss et al. 1995).

The translational positioning also places all the activator binding sites on the curved

surface of the nucleosome. Simultaneous binding of all the transcription factors does not

occur on naked DNA (Bruggemeier et al. 1990). 'Thus, their positioning on a curved

surface is believed to relieve steric constraints and allow concurrent binding of all factors

(Truss et al. 1995).

Promoter-Specific Chromatin Alteration

It was once thought that alterations in local chromatin structure might require

DNA replication to remove pre-existing nucleosomes (Svaren and Chalkley 1990). In

this scenario, transcriptional activator proteins would bind promoters after DNA

replication and prevent the reformation of nucleosomes. This would then allow other

transcription factors and the Pol II apparatus to bind. However, work in yeast has shown

that DNA replication is not required in order for nucleosome clearance/alteration to occur

(Schmid et al. 1992), and there are now many examples of promoter-specific chromatin

modulation. Two families of chromatin modulating complexes have been identified.

One is characterized by ATP-dependent nucleosome alterations, and the other is

characterized by the ability to acetylate nucleosomes.

ATP-dependent nucleosome alteration

There are several ATP-dependent chromatin-modifying complexes. The

prototype is the 2 MDa multiprotein Swi/Snf complex (Cairns et al. 1994). The Swi/Snf









proteins were initially identified in Saccharomyces cerevisiae through mutations that

produced defects in mating type switching (SWI) and sucrose fermentation (SNF)

(Sudarsanam and Winston 2000). The first link to chromatin structure came when

suppressors of Swi and Snf mutations were found to be located within chromatin

components including mutations in the histones H2A, H2B, H3, and H4 (Kruger and

Herskowitz 1991; Hirschhorn et al. 1992; Winston and Carlson 1992). They were then

found to alter nucleosomes in an ATP-dependant manner (Cote et al. 1994). There are

several other members of the Swi/Snf family of ATP-dependent chromatin remodeling

factors and they are found in a number of organisms. An additional yeast complex is

named RSC (Cairns et al. 1996). The related human complexes include hSWI/SNF

(Kwon et al. 1994), NURD (Xue et al. 1998), and RSF (Jaskelioff et al. 2000).

Drosophila complexes include ACF, CHRAC, NURF, and brm (Kingston and Narlikar

1999).

The mechanisms by which Swi/Snf and the other ATP-dependent complex alter

nucleosome structure and transcriptional activation are not well understood. Swi/Snf has

been shown to have a variety of effects on nucleosomes. It has been implicated in

nucleosome movement. This includes "cis-displacement," which is sliding without loss

of DNA binding (Whitehouse et al. 1999). Swi/Snf has also been shown to cause "trans-

displacement" of nucleosomes, which is the actual loss of DNA binding of nucleosomes

as evidenced by the ability to catalyze the transfer of nucleosomes from one DNA strand

to another (Lorch et al. 1999). The degree of cis- versus trans-displacement appears to

be dependent on the relative concentrations of Snf/Swi and nucleosomes. Trans-

displacement requires much higher concentrations of the Swi!Snf complex (Whitehouse









et al. 1999). The binding of Swi/Snf to either naked DNA or chromatin is ATP-

independent (Quinn et al. 1996), and this binding has been show to form loops within the

DNA or chromatin strands (Bazett-Jones et al. 1999). The Swi/Snf alterations in

nucleosome structure have been found to be stable after removal of Swi/Snf (Lorch et al.

1998). Interestingly, Swi/Snf is capable of catalyzing both the forward and reverse

reactions of nucleosome alteration (Lorch et al. 1998; Schnitzler et al. 1998).

There may be more than one mechanism by which the Swi/Snf complex is

brought to the promoter regions. Swi/Snf complexes have been co-immunoprecipitated

with both the yeast and human PollI holoenzymes (Wilson et al. 1996; Cho et al. 1998;

Neish et al. 1998). This finding generated an early hypothesis, which postulated that

Swi/Snf was brought to some promoters by the Poll holoenzyme itself. However, the

intracellular concentration of the Swi/Snf complex is one-tenth the concentration of

holoenzyme (Cairns et al. 1994; Cote et al. 1994), and PollI has not been co-

immunoprecipitated with Swi/Snf (Cairns et al. 1994; Cote et al. 1994; Cairns et al. 1996)

Alternatively, Swi/Snf might be brought to promoters via interactions with

transcriptional activators. In vitro studies performed in yeast demonstrated that Swi/Snf

interacts with Gcn4, Hap4, Ga1-AH, VP-16, and glucocorticoid receptor (Yoshinaga et al.

1992; Natarajan et al. 1999; Neely et al. 1999; Yudkovsky et al. 1999) The strength of

the Swi/Snf to Gcn4 interaction also correlated with the degree of Gcn4 transcriptional

activation (Natarajan et al. 1999). Chromatin immunoprecipitation (CHiP) demonstrated

in vivo that the presence of Swi/Snf at the HO promoter was dependent on the presence of

the transcriptional activator Swi5 (Cosma et al. 1999). The human Swi/Snf (hSwi/Snf)

has been shown to associated with the glucocorticoid receptor in vivo, and this interaction









was required for hormone dependent changes in the chromatin structure at the

glucocorticoid receptor binding site (Fryer and Archer 1998). The transcription factor

EKLF has been shown to recruit hSwi/Snf to the --globin gene (Lee et al. 1999).

Swi/Snf is required for the normal expression of only 6% of yeast genes by

genome wide expression analysis using HDAs (Holstege et al. 1998). Even more

intriguing, loss of Swi/Snf function was associated with decreased expression of 2%

genes and increased expression of 4% of genes (Holstege et al. 1998). Thus, its appears

to function both in transcriptional activation and repression. There is some evidence that

the strength of a promoter may dictate whether the promoter is Swi/Snf dependent.

Swi/Snf was required in vivo to facilitate the binding of Gal4p to weak binding sites

(Bums and Peterson 1997). However, when the binding site was altered to make it a high

affinity site or to place it in a nucleosome-free position, Gap4p binding became Swi/Snf

independent. The HDA data also indicate that Swi/Snf is not significantly involved with

the expression of yeast glycolytic enzyme genes (Holstege et al. 1998). This is consistent

with older data that showed that mutations in swil, swi2/snf2, and swi3 do not affect

TDH3 expression (Yoshinaga et al. 1992), and mutations in swi2/snf2, snf5, and snf6 do

not affect TPIl expression (Hirschhorn et al. 1992).

Histone acetylation

Histone acetylation has been implicated in the regulation of transcription for a

number of years. In the 1960's, while investigators were beginning to look at the role of

nucleosomes in transcriptional repression, it was noted that acetylated histones caused

less repression than non-acetylated histones (Allfrey et al. 1964). Several organisms have

also been noted to contain chromatin structures that correlate acetylation with









transcriptional activity. Tetrahymena contains two nuclei, a macronucleus that is

transcriptionally active and a micronucleus that is transcriptionally inactive (Lin et al.

1989). The macronucleus has been shown to contain acetylated histones, and the

micronucleus has decreased levels of acetylation (Gorovsky et al. 1973; Lin et al. 1989).

In Drosophila the male has a single hyperactive X chromosome that demonstrates

hyperacetylation of histone H4 compared to the female X chromosomes and all

autosomes (Turner et al. 1992). In mammalian females one of the two X chromosomes is

inactivated, and the inactivated X chromosome is "virtually devoid" of H4 acetylation

(Jeppesen and Turner 1993). Other investigators have used organomercurial-agarose gels

to isolate transcriptionally active chromatin(Walker et al. 1990). Chromatin isolated via

this chromatography technique was shown to have elevated levels of histone acetylation.

Antibodies directed against acetylated histones have also been used to purify acetylated

chromatin that was then shown to be enriched for active genes (Hebbes et al. 1988;

Hebbes et al. 1994).

The manipulation of nucleosome acetylation both in vivo and in vitro has

continued to link transcriptional activity with the degree of acetylation. Sodium butyrate

has been used to create hyperacetylated nucleosomes by inhibiting cellular deacetylases.

Treatment of cells with sodium butyrate has been associated with increased DNase I

sensitivity and increased expression from transfected DNA templates (Gorman et al.

1983). Nucleosomes reconstituted in vitro from hyperacetylated histones have

demonstrated increased binding of transcriptional activators (Lee et al. 1993; Vettese-

Dadey et al. 1996). Proteolytic removal of the N-terminal histone tails also results in

increased transcription factor binding that is similar to the result seen with acetylated









histones (Vettese-Dadey et al. 1994; Vettese-Dadey et al. 1996). Acetylation has also

been shown to disrupt higher-order chromatin folding, with an associated enhancement of

transcriptional activity (Tse et al. 1998). Acetylation decreases the number of times the

DNA strand is wrapped around each nucleosome core (Bauer et al. 1994) and results in

DNA that is less constrained by the nucleosome as demonstrated by greater flexibility

(Krajewski and Becker 1998). All of these findings have led to a model where histone

acetylation functions by destabilizing the interaction between the N-terminal tails and

DNA (Wolffe and Guschin 2000). Acetylation reduces the ionic charge on the histone

tails, and as a result acetylation would weaken the histone-DNA interaction proposed by

the charge neutralization hypothesis discussed above.

The connection between histone acetylation and promoter specific transcriptional

activation came with the identification of the Tetrahymena histone acetylase p55, which

was subsequently found to be homologous to the yeast transcriptional activator Gcn5p

(Brownell et al. 1996). GCN5 was initially identified through two independent mutation

screens. One was looking for mutations that effected the function of Gcn4p, which is a

transcriptional activator required for the expression of genes activated during amino acid

starvation (Georgakopoulos and Thireos 1992). It was also identified under a different

name, ADA4, during a screen looking for suppressors of Gal4-VP 16 toxicity (Marcus et

al. 1994).

Gcn5p was found to be associated with two protein complexes, the 0.8 MDa Ada

complex and the 1.8 MDa SAGA (_pt-Ada-Gcn5-acetyltransferase) complex (Grant et al.

1997). In the original isolate, the SAGA complex contained Gcn5, four Ada proteins,

four Spt proteins, and additional unidentified proteins (Grant et al. 1997). SAGA is also









now known to contain at least five TAFs (Grant et al. 1998). There are additional

acetylase complexes found in yeast including NuA3 (nucleosomal acetyltransferase of

H3) and NuA4 (nucleosomal acetyltransferase of H4),(Eberharter et al. 1998; Allard et al.

1999). Even TFIID has been found to contain a histone acetylase, TAF 145 (Mizzen et al.

1996). Gcn5p and SAGA also appear to have cross species conservation. The human

acetylase PCAF, which has homology to Gcn5p (Yang et al. 1996), is also found within a

multi-protein complex, which like SAGA also contains histone-like TAFs (Ogryzko et al.

1998). These are all examples of transcriptional activators that are associated with

acetylase activity. The opposite also occurs. There are several examples of

transcriptional repressors that are associated with deacetylase activity (Taunton et al.

1996; Pazin and Kadonaga 1997; Wolffe 1997).

Nucleosome acetylation appears to be a phenomenon generally associated with

transcriptional activation. However it is unclear whether complexes such as SAGA play

general roles in transcription regulation. When Gcn5p or other SAGA-specific proteins

are inactivated, the expression of at most 12% of genes are significantly affected, as

assessed by genome wide expression analysis using HDAs (Lee et al. 2000). As an

individual gene, the deletion of Gcn5p affected expression of only 4% of yeast genes.

However, it appears that SAGA has overlapping function with TFIID (Lee et al. 2000).

When the TAFs that are shared between the two complexes were mutated, they

individually affected the expression of 10-67% of yeast genes. As a group they were

required for expression of 70% of yeast genes. This effect is not specific to the

acetylation function because when the two acetylases TAF145 and Gcn5p are both

depleted at the same time, expression is only affected in 25% of the genes. Of note,









neither TFIID, SAGA, nor the combination appeared to significantly affect transcription

of glycolytic enzyme genes.

Transcription and Saccharomyces cerevisiae

The yeast S. cerevisiae provides an excellent model for the study of eukaryotic

transcriptional activation. It provides a eukaryotic environment within an organism,

which like the prokaryote Escherichia coli (E. coli), is generally easy to grow, and has a

rapid growth rate. It is genetically very malleable (Guthrie and Fink 1991), allowing

genetic manipulations not currently possible in mammalian cells. The S. cerevisiae

genome has also been sequenced (Cherry et al. 1997; Cherry et al. 1998). It has also been

shown to have many common features when compared to the mammalian eukaryotic cell.

In addition to overall cellular organization, these include similarities between RNA

polymerases, enhancer/UAS elements, protein structural domains, chromatin structure,

and many transcription factors. Many of these are been described above.

Within the context of S. cerevisiae, the glycolytic enzyme genes provide an

interesting system with which to study activated transcription because they are highly

expressed and because their transcription control mechanisms appear to be both

multifactorial and coordinated. The glycolytic pathway (Figure 1.1) is a dominant system

within this organism. S. cerevisiae, also known as Baker's yeast or Brewer's yeast, has

been selected over centuries to efficiently convert sugars into ethanol. It has been found

to possess very high levels of glycolytic enzyme activity. In fact, it is estimated that

these enzymes make up 30-60% of the cell's soluble protein (Hess et al. 1969; Fraenkel

1982), and the mRNA transcripts constitute a major fraction of total cellular mRNA

(Holland and Holland 1978). More recent analysis using both SAGE (serial analysis of

gene expression) (Velculescu et al. 1997) and HDAs (Holstege et al. 1998) have




















































Figure 1.1. The glycolytic pathway

Glucose is transformed to several end products in yeast, including lactate and ethanol.
The enzymes required for each step of the pathway and the genes, which encode each
of these enzymes, are indicated next to each set of arrows.














Glycolytic Pathway


Glucose

Hexokinase
(HXKI, HXK2)
Glucose 6-phosphate
Phosphoglucose isomerase
(PGJ)
Fructose 6-phosphate
Phosphofructokinase
(PFKI, PFK2)
Fructose 1,6,-bisphosphate
SAldolase
(FBAI)


Dihydroxyacetone phosphate I
Triosephosphate
isomerase (TPII


Glyceraldehyde 3-phosphate
4 Glyceraldehyde 3-phosphate
r) Dehydrogenase
(TDHI, TDH2, TDH3)
1,3-bisphosphoglycerate
Phosphoglycerate kinase
(PGKI)


3-phosphoglycerate

SPhosphoglycerornutase
(GPMI)

2-phosphoglycerate
Enolase
(ENOI, EN02)

Phosphoenolpyruvate
Pyruvate Kinase
(PYKI)
Lactate m Pyruvate
Lactate dehydrogenase t Pyruvate decarboxylase
(LDHI) (PDCJ)

Acetaldehyde
Alcohol dehydrogenase
(ADHI, ADH2)
Ethanol









confirmed the high level of glycolytic gene expression. The SAGE analysis

demonstrated that each glycolytic enzyme was represented by 69-425 mRNA transcripts

per cell. The HDA analysis demonstrated each glycolytic enzyme to be represented by

50-123 mRNA transcripts per cell. The vast majority of yeast genes (approximately

80%) are only represented by 0 1-2 mRNA transcripts per cell (Holstege et al. 1998).

Expression of the glycolytic enzyme genes is primarily constitutive. Possible

exceptions may include PGK, EV02, PYK, PDC1, and ADH1 that show some induction

in response to glucose. However, much of the glucose induction may be the result of

changes in enzymatic activity rather than transcriptional activity. Enzymatic activity was

studied in the hybrid yeast Saccharomycesfragili X Saccharomyces dobzhanskii, and

this work showed a glucose-modulated induction of 3- to 7-fold for most glycolytic

enzymes and 50- and 70-fold inductions for pyruvate decarboxylase and glyceraldehyde-

3-phosphate dehydrogenase, respectively (Maitra and Lobo 1971). However, when the

mRNA levels were assayed in S. cerevisiae it was found that most glycolytic transcripts

were induced less than two-fold. The transcripts for PFK2, PGM, ENO, and PYK

showed the greatest effect with 2.2-, 2.3-, 3.1-, and 3.9-fold inductions, respectively

(Moore et al. 1991).

The first evidence that transcriptional activation at the glycolytic enzyme genes is

regulated by a common mechanism came with the discovery of the Glycolysis Regulation

gene (GCR1). Mutations in the regulatory gene GCR] are associated with a significant

loss of both mRNA-encoding glycolytic enzymes and measurable enzyme activity

(Clifton and Fraenkel 1981; Holland et al. 1987; Scott et al. 1990). GCRJ has been

shown to encode a DNA-binding protein that has binding sites in the Upstream









Activating Sequences (UASs) of the glycolytic enzyme genes, and it has been shown to

be required for activated transcription from glycolytic enzyme genes (Baker 1991; Huie

et al. 1992). Gcrlp primarily effects the expression of glycolytic genes and Ty

transposon elements (Ciriacy et al. 1991; Turkel et al. 1997; Dudley et al. 1999; Lopez

and Baker 2000). Additional proteins, which are known to bind at these UAS elements,

include RebIp, Abflp, and Rap Ip. All three of these additional proteins are known to

effect expression of a wide variety of genes and as a result a wide variety of cellular

processes. Figure 1.2 summarizes the structures of the glycolytic UAS elements. The

similarity of the UAS regions for glycolytic enzyme genes suggests that their

transcriptional activation occurs through a similar mechanism and that this mechanism

provides the basis for the coordinated transcription of this family of genes.

TPIJ Transcriptional Activation

In this study, the gene TPI1 was used as a model for activated transcription of a

glycolytic enzyme gene. TPII encodes the triosphosphate isomerase enzyme, which is

required for the conversion of dihydroxyacetone phosphate into glyceraldehyde-3-

phosphate. It is a single-copy gene, and it has a promoter structure that includes both a

TATA element and a UAS (Baker 1986; Scott et al. 1990). Within this UAS there are

known protein-binding sites for RebIp, Rap Ip, and Gcrlp (Figure 1.3). Each of these

sites appears to play an essential role in full expression of the TPIJ gene. Utilizing

UASTPII elements in which either the RebIp-, the RapIp-, one Gcrlp-, or both GcrIp-

binding sites are deleted, transcription fell to 18%, 9%, 28-38%, and 1% of wild-type

respectively (Scott and Baker 1993).











































Figure 1.2 Glycolytic UAS elements

The glycolytic enzyme genes of Saccharomyces cerevisiae have a highly conserved
set of transcriptional activators that bind at their UAS elements. These include
Raplp, Gcrlp, Reblp, and Abflp. References for the binding sites listed here
include: PGI (Tekamp-Olson et al. 1988), PFK2 (Heinisch et al. 1991; Huie et al.
1992), FBA1 (Schwelberger et al. 1989), TPI1 (Baker 1991 ;Chambers et al. 1989;
Huie et al. 1992; Scott et al. 1990; Scott and Baker 1993), TDH3 (Bitter et al. 1991;
Kuroda et al. 1994; Huie et al. 1992), PGK1 (Chambers et al. 1989; Chambers et al.
1994; Ogden et al. 1986; Huie et al. 1992), GPM1 (Rodicio et al. 1993), ENO1
(Brindle et al. 1990; Buchman et al. 1988; Chambers et al. 1989), ENO2 (Brindle et
al. 1990; Willett et al. 1993), ADHI (Buchman et al. 1988; Chambers et al. 1989;
Tornow and Santangelo 1990), PDC 1 (Butler et al. 1990; Chambers et al. 1989).


















TPII "O



PYKI ....



PGI



PFK2 r



FBA I



TDH3



PGK1 C-



GPM] pp



ENO] psi



EN02 mOO



ADHI op



PDCJ e"




100bp *=Raplp 0 Gcrlp cJ=Reblp =Abflp
= structural gene

































Figure 1.3 The TPIJ gene and surrounding region

A. The organization of the TPI1 promoter includes the UAS binding sites between bases -396 to -371 and a TATA box at position
-170. B. The structural gene is 747 basepairs in length. Several restriction enzyme sites used in this study are indicated including
AvalI/TthIllI, Kpnl, BsiHKAI. The UAS and TATA elements are located as notated. C. The sequence surrounding TPIJ contains
several open reading frames and genes as identified by the yeast genome sequencing project. The locations of DBF4, YDR051C, and
YDRO49W are indicated.

















Rebip Gcrlp Rapip Gcrlp TATA


-396


-371


-170


II I %.
-859 +1 '-'+747
UAS TATA




C.


DBF4 YDR05 C TPI YDRO49W


Reblp Gcrlp Raplp


Gcrlp


TATA










A model for UASTPI, function has been proposed that predicts a sequential

binding process in which the binding of one protein allows the binding of subsequent

proteins (Scott and Baker 1993; Drazinic et al. 1996; Huie and Baker 1996). In this

model, Reblp or Abflp bind first, possibly altering the chromatin structure over the UAS

region. RapIp then binds its site, which is within this nucleosome-free region. The

presence of Rap I p then creates conditions that are appropriate for full binding by Gcrl p

at the Gcrlp-binding sites adjacent to the Raplp-binding site. Finally, Gcrlp binds and

stimulates transcriptional activation either directly or through adapter proteins such as

Gcr2p (Uemura and Jigami 1995) or Gall Ip (Stanway et al. 1994). Thus, there would be

a sequential addition of proteins at the UASTpiI, each necessary to prepare the promoter

for the addition of the subsequent proteins.

This model, if correct, has significant implications for the coordinated control of

transcriptional activation. As will be discussed below, Reblp, Abflp, and Raplp act at

many different genes and appear to perform the role of preparing the regulatory elements

of multiple genes that may need to be coordinately expressed or repressed. Also, Gcr 1 p

acts primarily at glycolytic genes and may be the factor that creates specificity for the

activation of this distinct group of functionally related genes. This provides a nested set

of controls that become progressively more specific. Since the glycolytic genes appear to

be constitutively expressed, it may be that this specific grouping of proteins has evolved

in S. cerevisiae to maintain constitutive expression.









Proteins That Bind at the Upstream Activating Sequences of Glycolytic
Enzyme Genes

Rebl12

Reblp is a highly abundant protein, which is essential for cell viability (Morrow

et al. 1993). Reblp affects transcription at many genes, and as a result it appears to be

important for many cellular functions. This multi-functional nature is further emphasized

by the fact that it affects transcription by both RNA Polymerases I and II. At rRNA

genes, which are transcribed by RNA Polymerase I, Reb 1 p is required for both activation

and termination of transcription (Morrow et al. 1989; Kulkens et al. 1992; Lang and

Reeder 1993). Reblp DNA-binding sites have also been shown at loci transcribed by

RNA Polymerase II including the glycolytic enzyme genes (Figure 1.2), autonomously

replicating sequences (Sweder et al. 1988), and GAL1-GALIO (Chasman et al. 1990).

Such a wide variety of activities suggest that RebIp may provide a very general function

such as preparing promoters for activation by other factors.

The proposal that Reb Ip functions to prepare promoters for transcription

originates from evidence that Reblp cannot activate transcription on its own. Instead it

has been found to potentiate activation by other proteins (Chasman et al. 1990). The

strength of this potentiation was also shown to be distance dependent. Reb I p was once

thought to affect chromatin structure at the GAL1-GALIO locus (Fedor et al. 1988:

Chasman et al. 1990). However, recent data demonstrates that this is not true (Reagan

and Majors 1998)

Abflp

RebIp does not have binding sites in all glycolytic enzyme gene promoters, and in

these cases it appears that the protein Abfl p may supply the Reb 1 p function (Figure 1.2).









Abflp is another highly abundant protein that has been shown to act at multiple

promoters and is required for cell viability (Buchman et al. 1988a; Rhode et al. 1989).

The Abflp binding-site alone, like the Reblp binding-site, is only capable of weak UAS

activity, but when combined with a T-rich region from the DEDI promoter it becomes a

strong activator of transcription (Buchman and Komberg 1990). This similarity of

function has been further implicated by experiments that show Reblp binding-sites can

functionally replace Abflp binding-sites, and Abflp binding-sites can functionally

replace Reblp binding-sites (Remacle and Holmberg 1992). Thus, both proteins may use

the same common mechanism to fulfill their role in transcriptional activation.



Raplp is also a highly abundant protein that is essential for cell viability (Shore

and Nasmyth 1987; Buchman et al. 1988a; Buchman et al. 1988b). Like Reblp, Raplp is

important in a wide variety of cellular activities including transcriptional activation and

repression (Shore and Nasmyth 1987; Buchman et al. 1988a), maintenance of telomere

length (Chasman et al. 1990), and meiotic recombination (White et al. 1993). Rap Ip has

been shown to be required for transcriptional activation at many genes including the

glycolytic enzyme genes (Figure 1.2), ribosomal genes (Huet et al. 1985), and an ATP-

ase gene (Rao et al. 1993).

Current evidence indicates that Rap I p does not function as the direct or final

activator protein, and instead it likely works through additional activating factors. Rap I p

binding sites act only as very weak activators, but when a Raplp site is adjacent to a CT-

box, activation increases over 10 fold (Buchman et al. 1988a; Bitter et al. 1991). CT-

boxes contain the sequence C(A/T)TCC and are known to be the binding site for Gcr 1 p









(Baker 1991). In vivo footprinting of HBY4, a gcrl deletion strain in which glycolytic

enzyme activities and mRNA levels are greatly reduced (Scott and Baker 1993),

demonstrated that while the Rap Ip-binding site remained occupied, the adjacent CT

boxes were unoccupied. This provided further evidence that Rap I p does not activate

transcription alone, and implied that RapIp is necessary to allow transcription, but that an

additional factor(s) is required for activated transcription. The Rap I p-binding site is

required for wild-type activation of transcription at the TPI] promoter, and there are

Gcrlp sites located adjacent to a RapIp site (Huie et al. 1992; Scott and Baker 1993). In

fact, as seen in Figure 1.2, this close proximity between Raplp- and Gcrlp- binding sites

is a hallmark of glycolytic enzyme UAS elements. Thus, Gcr1p presents itself as the

probable Rap I p-associated activation factor. The same results have been seen at the

pyruvate decarboxylase structural gene (PDC1), where Raplp is required for expression

but is not sufficient for wild-type expression (Butler et al. 1990).

At this time the mechanism for an interaction between RapIp and Gcrl p is

unknown, but there are two leading possibilities: protein-protein interactions and changes

in the DNA conformation at the GcrIp binding-site. Both of these mechanisms could

theoretically create an environment at the Gcrlp binding-site that allows for greater

binding specificity. There is evidence for both of these possibilities. First, using epitope-

tagged proteins, Raplp and Gcrlp have been co-immunoprecipitated, indicating that they

are capable of forming protein-protein contacts in vitro (Tornow et al. 1993). Second, it

has been inferred from circular permutation assays that Rap 1 p creates a sharp bend in

DNA when it binds DNA (Vignais and Sentenac 1989). The crystal structure and

scanning tunneling microscopy of the Rap I p binding domain (aa 361-596) has shown a









200 to 290 bend in DNA containing its binding sequence (Muller et al. 1994; Konig et al.

1996). Scanning tunneling microscopy using full-length Rap Ip demonstrated a mean

DNA bend of 52' (Muller et al. 1994). This bend may change the conformation of

adjacent DNA sequences, including the Gcrlp-binding sites. These two mechanisms are

both consistent with the current model in which Rap I p directs Gcr I p-mediated

activation, and they are not mutually exclusive. However, it is possible that Rap I p also

forms one subunit of an activation surface that also contains Gcr Ip.

Gcrl p

The characteristics of Gcr 1 p are distinctly different from Reb I p, Abfl p, and

Raplp. It is expressed at low levels and is not essential for cell viability (Baker 1986).

Its binding sites have been found primarily at glycolytic enzyme gene and Ty element

promoters (Ciriacy et al. 1991; Turkel et al. 1997; Dudley et al. 1999; Lopez and Baker

2000). HDA experiments comparing the transcriptomes from a wild-type strain and a

gcrl deletion mutant strain demonstrated that 53 open reading frames (ORFs) were gcrl

dependent (Lopez and Baker 2000). Approximately 75% of these 53 ORFs represented

either glycolytic enzyme genes or Ty elements, and there were an additional 14 unrelated

ORFs. The functional importance of Gcrlp at glycolytic enzyme genes was illustrated by

deleting the single copy of GCRJ. Strains that have the gcr] deletion mutation are viable

but grow very slowly and form very small colonies These strains also have a dramatic

loss of glycolytic enzyme activity to below 10% of wild-type for each of the glycolytic

enzymes (Clifton et al. 1978; Clifton and Fraenkel 1981; Baker 1986). This loss of

enzyme activity is believed to be the result of decreased transcriptional activation because

there is a parallel loss of mRNA levels (Clifton and Fraenkel 1981; Scott et al. 1990). In

addition, when individual Gcrlp sites are mutated in specific promoters including TPIJ,









EN02, TDH3, and PGK expression from that promoter is reduced (Chambers et al. 1989;

Bitter et al. 1991; Scott and Baker 1993; Willett et al. 1993). This implicates Gcrlp as an

important regulator of transcription at the glycolytic enzyme genes.

Gcrlp appears to be capable of transcriptional activation when used in isolation.

When fused with the lexA binding domain, Gcrlp was able to activate transcription 116-

fold over the lexA binding domain alone (Scott and Baker 1993). These lexA

experiments utilized a lexA operator that does not contain sites for Reb I p or Rap I p.

Thus, the Gcrlp activity was isolated from the other UASTpII binding factors. Further

evidence for Gcrlp's role as an activator comes from its association with a region of

nuclease hypersensitivity. In a wild-type strain, the promoter of TDH3 is associated with

a large region that is hypersensitive to DNase I, and this hypersensitivity extends from -

560 to -40 relative to the transcription start codon (Pavlovic and Horz 1988). This region

includes a UAS element and a Gcrlp binding site (Bitter et al. 1991) as well as the TATA

box. In a gcr! deletion mutant the size of the hypersensitive region shrinks. The area

from -560 to -370 remains hypersensitive, but the region from -370 to -50 becomes

protected from nuclease digestion (Pavlovic and Horz 1988). They suggested that this

protected region was the result of two nucleosomes that formed over this region, which

includes the TATA box, in the absence of Gcrlp. Whether these positioned nucleosomes

prevent transcription or are just associated with a lack of transcription activity is yet to be

determined, but it has been suggested that the presence of nucleosomes over a TATA box

inhibits binding of the transcription apparatus (Komberg and Lorch 1992).

The DNA-binding characteristics of Gcrl p are also consistent with the sequential

binding model. Gcrlp has a high binding affinity for DNA but a relatively low









specificity for its specific DNA-binding site (Huie and Baker 1996). It has an apparent

dissociation constant (Kd) of 2.9 X 10O for the Gcr I p-binding sequence found at position

-375 of the TPIJ locus. Additionally, there was only a 33-fold difference between the

specific and non-specific binding affinities in competition experiments. For comparison,

the specific and non-specific binding of Rap I p differ by almost six orders of magnitude

(Huie and Baker 1996). Thus, it is hypothesized that Gcrlp requires factors in addition to

DNA sequence to direct binding specificity.

The spatial context of the Gcrl p-binding site relative to the Rapl p-binding site is

important for Gcrlp binding and transcriptional activation (Drazinic et al. 1996). Using a

TPII::lacZ reporter gene, it was shown that if the Rap I p- and Gcrl p-binding sites are

separated by five base pairs, transcriptional activation was virtually eliminated. If the

sites are separated by ten base pairs then transcriptional activation is 50% of that seen

with wild-type spacing. As the sites are further separated at 5 base-pair intervals, there is

a continual decrease in transcriptional activation. The degree of decreased activity is

greater for separations that are a multiple of five and less for those that are a multiple of

ten. At a +30 base-pair interval, activation is approximately 5% of wild-type, which is

still higher than the level of activation seen at the +5 base-pair interval. In vivo

footprinting was also used to investigate the protein binding at these promoters (Drazinic

et al. 1996). In the wild-type construct there was normal protection over the Rap lp- and

Gcrlp-binding sites. In the +5 base-pair construct there was protection at the Rap Ip-

binding site but no protection at the Gcrlp-binding site. In the +15, +20, +25, and +30

constructs there was incomplete and decreasing protection over the Gcr I p-binding site.









Thus, the facilitation of Gcrlp binding is dependent on both the distance from the Rap I p-

binding site and on its rotational positioning relative to Rap Ip.

The close spatial relationship between the RapIp and Gcrlp binding-sites

implicates RapIp as the factor that provides the additional specificity for Gcrlp binding.

There are two possible mechanisms of interaction. First, RapI p could alter the DNA

conformation at the Gcrlp site, and second there may be physical contact between RapI p

and Gcrlp. As noted above, Raplp does cause changes in DNA conformation by

bending the DNA at its binding-site. In addition, there is evidence that Gcrlp itself

causes a small bend when it binds to the DNA (Huie and Baker 1996). This Gcrlp-

induced bending could be significant because it has been shown that proteins that bend

DNA bind to bent DNA with a higher affinity than to unbent DNA (Kahn and Crothers

1992). The second possibility, protein-protein contacts, has been implicated by the co-

immunoprecipitation experiments discussed above (Tomow et al. 1993). The two

mechanisms, DNA conformational changes and protein-protein contacts, are not mutually

exclusive and may both be present.

Non-DNA-binding Proteins That Affect Expression of Glycolytic Enzyme Genes

Although Gcrlp is capable of activating transcription, it is not known whether

Gcrlp interacts with the transcription apparatus directly or whether it operates through

adapter proteins. The genes GCR2 and GAL] encode non-DNA-binding proteins that

are required for wild-type expression of many glycolytic genes (Nishizawa et al. 1990;

Uemura and Fraenkel 1990). As a result these proteins present themselves as possible

adapter molecules. At the present time, the mechanism of their activity is unknown.

Strains that contain deletions of these genes are viable and thus can be used to investigate

their role in transcription.









Gcr2p

GCR2 was identified in a screen for mutations that affect expression from the

ENO] promoter (Uemura and Fraenkel 1990). Deletion of GCR2 results in a profound

decrease in glycolytic enzyme activity that is similar to the defects found in gcr] mutant

strains (Uemura and Fraenkel 1990). The mechanism for the glycolytic enzyme defects

appears to be due to defects in transcriptional activation as evidenced by significant

reduction in mRNA levels seen in Northern analysis (Uemura and Fraenkel 1990).

However, the gcr2 mutants do not have as severe a growth defect as gcr] mutants when

grown on glucose containing media at 30C (Uemura and Fraenkel 1990).

Current evidence indicates that Gcr2p acts as a non-DNA-binding adapter protein,

possibly through interactions with Gcr lp. Evidence for an interaction between Gcr l p

and Gcr2p comes from genetic studies (Uemura and Jigami 1992, 1995) utilizing the

"two-hybrid" method of Fields and Song (1989). In addition a Rap I p-Gcr2p fusion

protein partially complements the defects of the gcr] mutation (Uemura and Jigami

1992). Currently there is no evidence that Gcr2p binds DNA. In vivo footprinting in

wild-type, gcri, and gcr2 mutant strains did not provide evidence for a Gcr2p-binding

site within the UASTPIJ element (Scott and Baker 1993; Huie and Baker 1996).

Gall p

Gall lp appears to be a general transcription factor that has been implicated in the

positive and negative regulation of several loci. It is a positive regulator of the GAL

system, MRal, MA Tal, and MA Ta2, and it is a negative regulator of Ty element (Fassler

and Winston 1989). There is also evidence that it interacts with zinc-finger proteins









including Abflp (Stanway 1991; Sakurai et al. 1993). It has also been isolated in

association with the basal RNA polymerase II transcription apparatus (Kim et al. 1994).

It has been suggested that Gal lIp acts as an adapter protein at the UAS elements

of glycolytic enzymes. The primary evidence comes from the PYK1 and PGKJ loci. In a

gall] mutant transcriptional activation from a UASpYKI construct is reduced seven fold

when compared to wild-type (Nishizawa et al. 1990). In addition, mRNA levels for the

PGKJ transcript are reduced two fold in a gall mutant (Stanway et al. 1994).

Summary

The process of eukaryotic transcriptional activation is a complex process that

requires interactions between a large number of proteins and protein complexes including

histones, chromatin modifying complexes, transcriptional activators, transcriptional

repressors, mediators, as well as the polymerase complex. The UAS elements of the

glycolytic enzyme genes provide a model of eukaryotic transcriptional activation. More

specifically, they represent an example of a transcription system in which the UAS

elements direct appropriate transcriptional expression from a group of functionally

related genes. This study presents the chromatin structure surrounding the TPIJ gene.

This structure includes an area of distinct nuclease hypersensitivity over the regulatory

promoter elements. It also demonstrates that the transcription factor Gcrl p requires the

simultaneous binding of a second DNA-binding protein, Raplp. This combinatorial

interaction would provide increased specificity to transcriptional activation. Experiments

are presented that investigate which domains within Raplp are required for this

interaction.









The current evidence supports a stepwise preparation of the promoter region,

which then allows the final activator to bind and direct transcriptional activation. The

UAS elements in yeast glycolytic enzyme genes are structurally similar to the UAS

elements in other yeast genes. They are also similar to the enhancers of higher

eukaryotes. Thus, it is likely that these mechanisms will be present at other non-

glycolytic yeast genes as well as at genes in higher eukaryotes. With greater knowledge

of the UAS/enhancer function it may become possible to modulate the activity of specific

genes or gene families for research, industrial, and medical benefit.














MATERIALS AND METHODS


Strains

The genotypes for all strains of S. cerevisiae and E. coli used in this study are

described in Table 2.1.

Media and Growth Conditions

Yeast cultures were grown in yeast-peptone (YP) medium (Rose et al. 1990)

supplemented with 2% glucose (YPD) or 2% lactate and 2% glycerol (YP-Gly/Lac).

Unless otherwise noted, liquid yeast cultures were grown at 30'C with shaking. In

experiments involving temperature-sensitive strains, yeast cultures were grown at the

permissive temperature of 24'C.

Unless otherwise noted, E. coli cultures were grown in Luria Broth (LB) medium

(Miller 1972) and, as appropriate, supplemented with ampicillin (LBamp). E. coli

cultures used for growth of phage Ml 3 were grown in Minimal Media 63 (M63)

supplemented with one microgram per milliliter (jig/ml) thiamine hydrochloride, and

25jig/ml amino acids as required (Miller 1972). All E. coli cultures were grown at 37C

with shaking.

Chromatin Mapping by Nuclease Sensitivity

The nuclease sensitivity experiments were based on a method published by Kent

et al. (1993). Nuclease sensitivity patterns have been used most commonly to map the

positions of nucleosomes. Unfortunately, nucleases such as DNase I and micrococcal

nuclease are unable to enter intact cells because they are too large to diffuse through the









Table 2.1. Strains


Genotype


Strain

E. coli

KK2186


MC1061


DH5alpha



S. cerevisiae

S150-2B

HBY4

DFY641

DFY642

YDS485

YDS487

W303-1A

R884-1C


supE, sbcB 15, hasdR4, rpsL, thi, A(lac-proAB) F' [traD36,
proAB+, lacq, lacZAm 15]

hsdR, mcrB, araD139, A(araABC-leu)7679 AlacX74, galU,
galK, rpsL, thi

080dlacZAM 15, A[lacZYA-argF] U 169, deoR, recA 1, hsdR 17,
supE44, thi-1, hyrA96, relA1




MATa, leu2-3, 112, his3A, trp1 -289, ura3-52

MATa, gcrlA::HIS3, leu2-3,112, his3A, trpl-289, ura3-52

MATa, gcr2A::URA3 leu2-3,112 ura3-52

MATa, leu2-3,112, ura3-52

MATt, ade2-1 his3-11,13 trpl-1 ura3-1

MATa, ade2-1 his3-11,13 trpl-1 ura3-1 rapl-2ts

MATa, ade2-1 trpl-1 canl-100 leu2-3,112 his3-11,15 ura3052

MATa, ade2-1 trpl-1 canl-100 leu2-3,112 his3-11,15 gall 1-
313










cell wall and membrane. This had previously limited their use to the treatment of in vitro

systems and to the treatment of isolated nuclei. The method of Kent et al. (1993) utilized

unlysed yeast cells. To perform the nuclease digestion in vivo it was necessary to

spheroplast and permeabilize the cells. Yeast possesses a cell wall composed primarily of

3(1-3)glucan, (1-6)glucan, and mannoprotein (Guthrie and Fink 1991; Johnston 1994).

This cell wall can be enzymatically disrupted with zymolase, which hydrolyses the poly

3(1-3-glucose) of the cell wall glucan (Guthrie and Fink 1991; Johnston 1994). The lipid

bilayer of the cell membrane was then permeabilized but not disrupted using the non-

ionic detergent NP-40. Permeabilization allowed for rapid sample processing and for less

manipulation of the cell and nucleus than occurred with older methods that utilized

isolated nuclei.

Permeabilization and Nuclease Treatment of Cells

The various strains S. cerevisiae were grown at 30'C with shaking to an

exponential growth phase and harvested at an A600 of 1.0 by centrifugation. Cells were

then washed three times with ddH20. The wet weight of the pellets was measured. The

cells were then resuspended in one ml/g of High DTT Solution (50 mM Tris-

hydrochloric acid (HC1) pH 7.5, 8 mM MgCl2, I M sorbitol, 30 mM dithiothreitol (DTT))

and incubated at room temperature for 5 minutes and then pelleted by centrifugation.

The DTT functions to reduce disulfide bridges in the mannoprotein layer allowing

increased access to the underlying 13(1-3)glucan layer by yeast lytic enzyme (Guthrie and

Fink 1991; Johnston 1994). The cells were then resuspended in three volumes of Low

DTT Solution with yeast lytic enzyme (50 mM Tris-HC1 pH 7.5, 8 mM MgCI2, 1 M









sorbitol, 5 mM DTT, and 200 u/pl yeast lytic enzyme) and incubated at 37C for 30 to 45

minutes. The adequacy of spheroplast formation was determined by observing rapid lysis

when 1 p was mixed with 5 ml ddH20. The spheroplasts were then pelleted by

centrifugation and washed twice in 1 M sorbitol. The spheroplasts were then

resuspended in one volume of Nuclease Treatment Buffer (I M sorbitol, 50 mM NaCI, 10

mM Tris-HCl pH 7.5, 5.3 mM MgC12, 1 mM calcium chloride (CaCI2, 0.3 mM

spermidine, and 1 mM i-mercaptoethanol). The spheroplasts were then aliquoted into

100 jtl volumes that were treated with either MNase or DNase I as described below.

Dilutions of MNase and DNase I were made in Nuclease Treatment Buffer

containing 0.05% Nonidet P-40 (NP-40) (v/v). The MNase concentrations used were 0,

0.006, 0.02, 0.06, and 0.2 units/pl, and the DNase I concentrations used were 0, 0.005,

0.01, 0.02, and 0.04 units/pl. Note that the concentrations of nuclease and NP-40 in each

reaction were actually half of the values described above because in a subsequent step the

nuclease samples were diluted 50% when they were added to the spheroplast samples.

After the nuclease dilutions were prepared, both the spheroplast aliquots and the nuclease

aliquots were pre-heated in a 37'C heat block for 2 minutes. Then the nuclease dilutions

were then added to the spheroplast tubes, gently mixed, and incubated at 37C for 10

minutes.

The nuclease reactions were stopped by adding 11 tl of 0.5 M EDTA, 11 tl 20%

SDS, and 2.5 pl Proteinase K (20 mg/ml) sequentially. It was not possible to premix the

stop solutions due to precipitation of the SDS. The samples were then incubated at 370C

for 30 minutes followed by centrifugation to pellet the cellular debris. The supematants









were extracted once with phenol and then once again with phenol:chloroform (50:50).

They were then ethanol precipitated.

Several experiments were performed to determine the appropriate concentrations

of NP-40 and nuclease. The degree of nuclease digestion is directly proportional to the

concentrations of both the NP-40 and the nucleases, and the actual concentrations were

made empirically. For a given concentration of NP-40 a series of nuclease dilutions were

tested for their average fragment size when viewed on ethidiumbromide (EtBr)-stained

agarose gel.

Electrophoresis and Southern Transfer

The DNA samples were loaded onto a 1.5% agarose Tris-borate-EDTA (TBE) gel

25 cm in length. The gel was then run at 160 V until the bromophenol blue had migrated

a distance of 21 centimeters. The gel was stained with EtBr and photographed.

The DNA was then transferred to Hybond-N+ nylon membrane by the method of

Southern (Elder et al. 1983; Elder and Southern 1983). The gels were soaked in 0.25 M

HC1 for 30 minutes. They were then soaked in 0.5 M NaOH, 1.0 M NaC1 for 30 minutes.

The transfer buffer used in the Southern apparatus was 0.025 M NaPO4 (pH 6.5). After

transferring overnight for 12-14 hours the nylon membrane was removed and irradiated

in a UV-Stratalinker-1800 (Stratagene) for a total of 1200 X 102 joulese.

Probes, Probe Synthesis, and Hybridization Conditions

The probes used for indirect end-labeling were created by primer extension on

M13 templates. The M13 strains jbsTPImpl 8.2-21 and jbsTPImpl8.2-24 both contain a

4.4 kilobase (kb) HindII fragment from pHB51 (Scott et al. 1990), which contains the

TPII structural gene with surrounding sequence. Strain jbsTPImpl 8.2-21 contains the

non-coding strand, and strain jbsTPlmpl 8.2-24 contains the coding strand. Probes









produced with jbsTPImpl 8.2-21 and HB149 allow visualization of the region upstream

from the BsiHKAI restriction site that is located at nucleotide + 189 relative to the

translation start site of TPIJ. Probes produced with jbsTPImpl 8.2-21 and HB 152 allow

visualization of the region upstream from the KpnI restriction site that is located at

nucleotide +514 relative to the translation start site of TPI1. Probes produced with

j bsTPImp 18.2-24 and HB 149 allows visualization of the regions downstream from the

KpnI restriction site that is located at nucleotide +512 relative to the translation start site

of TPIJ (Figure 1.3).

Primer extension utilizing Klenow fragment was carried out using 2.5 I-tg of

template and 5 pmol of HB54 in a reaction buffer of 20 mM DTT, 10 mM MgC12, 50 mM

Tris-HCl (pH 8.0), 200 mM NaCl, 0.24 mM dNTPs (minus dATP), and 100 [iCi a 32P

dATP. The reaction was incubated at 37C for 45 minutes. The products were then

denatured at 100C and separated on a small preparative sequencing gel. The location of

the single-stranded, radioactive product was localized by brief autoradiography. The

appropriate area of the gel was excised from the gel and manually crushed. The crushed

gel fragment was suspended in 6 ml of hybridization solution (1% BSA, 1 mM EDTA,

[0.5 M Na]HPO4, 7% SDS).

Hybridization was carried out in a rolling drum hybridization chamber. The

membrane was pre-hybridized in hybridization solution at 60'C for 60 minutes. The

solution used for pre-hybridization was poured off, and the 6 ml volume of probe was

added. The membrane and probe were then incubated at 60'C for at least 12 hours. The

hybridized membrane was washed 3-4 times in wash solution (1 mM EDTA, [40 mM

Na]HPO4, 1% SDS) at 60-C.









Autoradiography

The Southern blots were exposed to Kodak AR film both with and without a

Lightning Plus (Dupont) intensifying screen.

In vivo Footprinting

In vivo footprinting utilizes one of the four chemical sequencing reactions

originally described by Church and Gilbert (Maxam and Gilbert 1977; Church and

Gilbert 1984). Dimethylsulfate (DMS) methylates DNA at the N7 position of guanine

(G) residues and the N3 position of adenine (A) residues. Because DMS can penetrate

the cellular and nuclear membranes, it was possible to perform this methylation in vivo.

When a protein binds a DNA sequence that contains G residues this methylation will

often be inhibited. Thus, analysis of the methylation patterns produced by DMS

treatment often allows the detection of protein binding at specific DNA sequences. The

procedure described here has been published previously (Huie et al. 1992; Lopez et al.

1993). Additionally, there are several other published descriptions of in vivo footprinting

protocols (Becker and Schutz 1988; Becker et al. 1993; Homstra and Yang 1993).

In vivo Dimethylsulfate Treatment and Preparation of Saccharomvces
cerevisiae DNA

The strains YDS485 and YDS487 were grown at 24C in 500 ml of yeast peptone

dextrose (YPD) with shaking to an optical density (A600) of about one. The cultures were

then concentrated 1 00-fold into 5 ml of clarified YPD. The concentrated cells were

incubated in the clarified growth media under three conditions: 24C, 24C with 10

gg/ml cycloheximide, and 37C for 30 minutes. DMS was then added at a concentration

of either 50mM for the 24C samples or 10mM for the 37C samples. The lower

concentration of DMS used at 37C was used to compensate for the increased









methylation activity of DMS at the higher temperature. Incubation times with DMS are

indicated in the text. The methylation reactions were quenched by the addition of 40ml

ice-cold phosphate buffered saline (PBS) (137 mM sodium chloride (NaCI), 2.7 mM

potassium chloride (KCl), 4.3 mM sodium phosphate (NaPO4), 1.4 mM potassium

phosphate (KPO4), pH 7.4). The cells were then pelleted by centrifugation. The cells

were washed three times with 35 ml ice-cold PBS to dilute out the DMS. The washes

were accomplished by resuspending the cells in ice-cold PBS, pelleting the cells by

centrifugation, and decanting the supernatant in a fume hood.

The genomic DNA was extracted from the DMS-treated cells with a Qiagen-tip

500 following the manufacturer's recommended protocol, as described below. The

washed cell pellet was resuspended in 12 ml Qiagen buffer Y1 (1 M sorbitol, 100 mM

ethylenedinitrilotetraacetic acid (EDTA), 14mM jP-mercaptoethanol) containing 0.4

milligrams per milliliter (mg/ml) yeast lytic enzyme and incubated at 37'C for 60

minutes. Spheroplast formation was confirmed by checking for osmotic lysis in ddH20.

The spheroplasts were then pelleted by centrifugation at 4 krpm for 5 minutes. The

spheroplasts were then resuspended in 12 ml Qiagen buffer G2 (800 mM guanidine

hydrochloride (GuHC1), 30 mM EDTA pH 8, 5% Tween-20, 0.5% Triton X-100) plus

250 microliters (Rt) 10 mg/ml RNase A and 20tl 20 mg/ml Proteinase K. This was

incubated at 370C for 2 hours. The lysate was then poured over a Qiagen-tip 500 that had

previously been equilibrated with Qiagen buffer QBT (750 mM sodium chloride (NaCI),

50mM 3-[N-Morpholino]propane sulfonic acid (MOPS), 15% ethanol, 0.15% Triton X-

100, pH 7.0). The column was then washed twice with 30 ml Qiagen buffer QC (1.0 M









NaC1, 50mM MOPS, 15% ethanol, pH 7.0). The genomic DNA was then eluted with 15

ml of Qiagen buffer QF (1.25 M NaCI, 50mM Tris-HCL, 15% ethanol, pH 8.5).

During the isolation and subsequent processing of the DMS treated DNA, care

was taken that it not be exposed to temperatures in excess of 37C. At temperatures

higher than 37C non-specific cleavage occurs at both the methylated G and methylated

A as a result of uncontrolled apurination (Becker and Schutz 1988). Preferential cleavage

occurs at the G residues during treatment with piperidine as will be described below.

In vitro DMS Treatment of Plasmid DNA

In order to identify areas that were protected from methylation, the DNA that was

methylated in vivo needed to be compared to similar DNA that had been methylated in

the absence of bound proteins. This was accomplished by treating the plasmid pHB51

(Scott et al. 1990) with DMS in vitro. The plasmid pHB51 has bases -3079 to +826 of

the TPIJ locus cloned into the HindIlI site of pUC 18.

Twenty micrograms ofpHB51 were precipitated with ethanol and resuspended in

200ptl of DMS buffer (50mM Na-cacodylate, I mM EDTA pH 8.0). DMS was added to

a concentration of 50mM (1 p ) and incubated at room-temperature for 60 seconds. The

reaction was then stopped by precipitation with of 50 ptl DMS stop solution (1.5 M

sodium acetate (NaAc), I M P-mercaptoethanol) and 750 ptl cold absolute ethanol. The

methylated DNA was pelleted by centrifugation, and the excess DMS was removed by

washing the pellet with 70% ethanol.

Restriction Diaest and Cleavage of Methylated Guanine Residues

Both the genomic DNA samples and the plasmid samples were digested with

Avail, which cleaves at a unique restriction site located 859 base-pairs upstream from the









TPIJ structural gene (Figure 1.3). The Avail-digested DNA was then cleaved at the

methylated G residues by incubating in 2001 1 M piperidine at 100'C for 30 minutes.

Under these conditions the cleavage occurs more rapidly at methylated G residues and

relatively slowly at methylated A residues (Maxam and Gilbert 1977; Becker and Schutz

1988). The piperidine reaction was stopped by precipitation with ethanol. The DNA was

pelleted by centrifugation, and the excess piperidine was removed during washes with

70% ethanol and vacuum drying of the pellet.

Electrophoresis

The fragments resulting from restriction digestion and piperidine cleavage were

separated on a sequencing gel (7% acrylamide, 0.23% bisacrylamide, 8 M urea, 0.5 X

TBE. The plasmid-derived G ladder was loaded at 4 nanograms (ng) per lane, and the

genomic samples were loaded at either 2 tg or 5 pg per lane as indicated. The gel was

pre-run at 50 V/cm for 30-60 minutes. The samples were denatured at 95'C for 3

minutes in formamide sequencing dye and then loaded onto the gel. After loading the

samples, the gel was run at 50 V!cm until the xylene cyanol had migrated 30 cm.

Transfer of DNA to Nylon Membrane

After electrophoresis was complete, the glass plates were separated, and a piece of

Hybond-N+ (Amersham) was placed on top of the gel. The membrane was used dry and

not pre-wetted. The gel, which adhered to the nylon membrane, was easily lifted off the

glass plate. It was placed membrane down onto two pieces of dry 3MM paper

(Whatman), covered with plastic wrap, and placed on the wire mesh of a Savant gel dryer

(Model SGC4050; Savant Instruments). It was then blotted under vacuum without heat

for 45-60 minutes. The gel sandwich was then placed plastic wrap side down on a 312









nm transilluminator (Fischer Scientific) and irradiated with ultraviolet (UV) light on high

for 10 minutes. The gel was removed from the membrane using ddH20.

Probes, Probe Synthesis, and Hybridization Conditions

The probe used for indirect end-labeling was created by primer extension on an

M13 template. Probes for the AvalI restriction site located at -487 with respect to the

translational start site of TPI1 were prepared from the M13 clone mesTPImpl 8B (Scott

1992), which contains the 5' non-coding region of the TPI1 locus. The primer used,

HB54, is complementary to the sequence at the AvaIl restriction site such that upon

annealing, its 5' end is located at the AvalI site (Huie et al. 1992). Primer extension

utilizing Klenow fragment was carried out using 2.5 lag template and 5 pmol of HB54.

The reaction buffer was composed of 20 mM DTT, 10 mM magnesium chloride (MgC12),

50 mM Tris-HCI (pH 8.0), 200 mM NaCl, 0.24 mM dNTPs (minus dATP), and 100 jiCi

&2p dATP.

The probe was then purified by electrophoresis on a 6% polyacrylamide, 8.3 M

urea, 0.5X TBE gel. The location of the single-stranded, radioactive product was

localized by brief autoradiography. The appropriate area of the gel was excised from the

gel and manually crushed. The crushed gel was suspended in 6 ml of hybridization

solution (1% bovine serum albumin [BSA], 1 mM EDTA, [0.5 M Na]HPO4, 7%

sodiumdodecylsulfate [SDS]).

Hybridization was carried out in a rolling drum hybridization incubator (Robbins

Scientific, Model 310). The membrane was pre-hybridized in hybridization solution at

60'C for 60 minutes. The solution used for pre-hybridization was poured off, and the 6

ml volume of probe was added. The membrane and probe were then incubated at 60'C









for at least 12 hours. The hybridized membrane was washed 3-4 times in wash solution

(1 mM EDTA, [40 mM Na]HPO4, 1% SDS) at 60'C.

Autoradiography

The hybridized membranes were exposed to Kodak AR film with one Dupont

Lightning Plus intensifying screen.

DNA Band-Shift Analysis

DNA band-shift analysis, also known as electrophoretic mobility gel-shift

analysis, is an in vitro method used for the study of protein-DNA interaction. The

procedures used in these studies were based on those by Fried and Crothers (1981) and

Garner and Revzin (1981).

Expression Vectors for RAp] and RAPJ-Truncations

Full-length Raplp and several carboxy-terminal and amino-terminal truncations

of RapIp were produced by in vitro transcription and translation from plasmid expression

vectors utilizing the SP6 RNA polymerase promoter. The plasmid pSP56RT7 contains

the entire RAP] coding region and its construction has been described previously

(Chambers et al. 1989). The plasmids pSPRAP361-596 and pSPRAP361-827, contain

truncated forms of RAP! (Figures 2.1 and 2.2). The RAPI truncations were produced by

polymerase chain reaction (PCR). The PCR primers were designed to introduce start and

stop codons at the proper ends of the coding sequences and to introduce restrictions sites

for cloning into the pSP19 multiple-cloning site (McCracken 1985). The 5' end of the

primers was also designed for optimal restriction activity since digestion with restriction

enzymes at the ends of DNA strands is often very inefficient. The cloning enzymes were

chosen by both their presence in the pSP19 polylinker as well as their ability to cut at the









































Figure 2.1 Design of RapIp truncation mutants

The RapIp mutations were created with custom oligonucleotides used as primers for
PCR amplification. A. The RAP] structural gene is displayed. It encodes a protein
that has several functional domains including a DNA-binding domain, an asymmetric
bending domain, an activation domain, and a silencing domain. The sites of
complementary annealing for the oligonucleotide primers are indicated. B. The
primers used for amplification contain either an EcoRI or a KpnI site as indicated.
Translation start and stop codons were designed into the sequence as indicated. The
genetic code triplets are bracketed above the sequence, and the associated amino acid
and native position within the wild-type Raplp are indicated. C. The products
produced by PCR amplification were cloned into the polylinker of the expression
vector PS 19. The SP6 promoter and the relative position of the polylinker are
indicated.
















HB 166


HB163
1 4.--


HB165


=_ Asymmetric DNA-bending domain, aa44-274
= -DNA-binding domain, aa361-596
-Activation domain, aa630-695
= Silencing domain, aa665-827


B.

Coding-strand primer


Start A
EcoRI Codon 361
HB165 5'aaaacg gaattc atg gct


S F T
362 363 364
tct ttt aca


D E
365 366
gat gag g 3'


Noncoding-strand primers


HB163 5'cggataacaatttcacacagg 3'


Stop A
Kpnl Codon 596
HB166 5'aaaaaa tgcag tca ggc








;2
0
U


A S
595 594
ggc aga


N Y N
593 592 591
gtt ata att gc 3'


A.
I IIIIIt ftIIIIIIIIItM tlI


I


| .....................






60




pSP56RT7 StuI
1 827 XbaI
1-183 1-127

SP6
promoter BglI
1-596
pSPTPI361-827
361 827 XbaI
SP6 1"2

promoter BglI
1-596


pSPTPI361-596
Kpn1
361 596

SP6
promoter


m = Asymmetric DNA-bending domain, aa44-274
= DNA-binding domain, aa361-596
= Activation domain, aa630-695
= Silencing domain, aa665-827






Figure 2.2 RapIp truncation vectors

The Raplp and Raplp truncations were produced by in vitro transcription followed by in
vitro translation. The transcription reaction utilized the SP6 polymerase. There were
three vectors. The plasmid pSP56RT7 contains the entire RAPI coding sequence. An
mRNA encoding full-length Raplp, aal-827, was produced by cleavage with Xbal
followed by run-off transcription. Similarly, mRNAs encoding aal-583, aal-596 were
produced run-off transcription after cleavage with Stul and BglII, respectively. Plasmid
pSPTPI361-827 was created using primers HB 165 and HB 163 as described in Figure 2.1.
An mRNA encoding aa361-827 was produced by cleavage with XbaI followed by run-of
transcription. Plasmid pSPTPI361-596 was created using primers HB165 and HB166 as
described in Figure 2.1. An mRNA encoding aa361-596 was produced by cleavage with
KpnI followed by run-off transcription.









ends of DNA strands as described by New England Biolabs (1996). Additional bases

were included to move the restriction sites away from the very ends of the PCR products.

The PCR template was pSP56RT7, the same plasmid used to express full-length Rap Ip.

The 5' cloning junctions were all confirmed by sequencing with HB 167, and the 3'

cloning junctions were confirmed by sequencing with HB163.

In vitro Transcription and Translation of Rap I p Constructs

Ten micrograms of each DNA template were linearized by incubation with

appropriate restriction enzyme. For pSP56RT7, cleavage with XbaI, Stul, BglI, BstBI, or

BglII resulted in transcripts encoding amino acids 1-827 (full-length), 1-583, 1-596, 1-

677, or 1-701, respectively. The plasmid pSPRAP361-596 when linearized with PstI

resulted in a transcript encoding amino acids 361-596. The plasmid pSPRAP361-827

when linearized with XbaI resulted in transcripts encoding amino acids 361-827,

respectively.

The templates were then transcribed with SP6 RNA polymerase. The reaction

conditions were 40mM Tris-HCl (pH 7.5), 6 mM MgC12, 2 mM spermidine, 10mM NaCl,

10mM DTT, 0.1 mg/mI BSA, 1 mM ATP, 1 mM CTP, 1 mM UTP, 0.1 mM GTP, 0.5

mM m7G(5")ppp(5')G, 50 units Rnasin, and 2.7 units SP6 RNA polymerase. After

incubating at 370C for 60 minutes, the samples were treated with DNase I for 15 minutes

at 370C, followed by extraction with phenol/chloroform and precipitation with ethanol.

The samples were resuspended in 20gl of ddH20 and incubated at 650C for 15 minutes.

The transcription products were stored at -700C.

The transcription products described above were translated in a Nuclease Treated

Rabbit Reticulocyte Lysate (Promega) with [35S]methionine to radiolabel the resulting

proteins. Two microliters of the in vitro transcribed RNA was combined with 18pil rabbit









reticulocyte lysate, 1 pl Rnasin, 1 pl 1 mM amino acid mixture minus methionine, 2itl

[35S]methionine 10 mCi/ml (Amersham), and 4jpl ddH20. The positive control

translation reaction used 1 p1 of Brome Mosaic virus (BMV) RNA (Promega).

Transcription, Translation, and Purification of Gcrl p(63 1-785)

The Gcrlp binding domain was produced as a malE::Gcrl fusion protein. As

previously described (Enea et al. 1975; Scott and Baker 1993), the E. coli strain TB1 was

transformed with pMAL::GCR1(631-785). This plasmid contains sequence encoding a

malE::GCR1 (631-785) fusion protein that is cloned downstream from a Ptac promoter. A

500 ml culture was grown at 37C to an A600 of 0.5. Then IPTG was added to a final

concentration of 2 mM to induce expression of the malE::GCR1(631-785) fusion gene.

The cultures were grown for an additional 2 hours at 37C and then harvested by

centrifugation. The cells were then resuspended in 3 ml TEN lysis buffer (50 mM Tris

[pH 8.0}, 1 mM EDTA, 50 mM NaCl) per gram of cells. The bacteria were lysed using a

French pressure cell at 20,000 lb/in2 (Clifton et al. 1978). The cellular debris was

pelleted by centrifugation at 17,000 x g for 20 minutes at 4'C. The supernatant was

retained as the crude protein extract.

The MBP-Gcrlp(631-785) was purified by affinity chromatography utilizing an

amylose column. A 2.5 x 10 cm column was packed with 7 cm3 of amylose resin. The

column was washed with three volumes (25 ml) of column buffer (10 mM phosphate, 0.5

MM NaCI, 1 mM azide, 1 mM DTT, 1 mM EGTA) containing 0.25% Tween 20. The

crude protein extract was diluted 1:5 with column buffer containing 0.25% Tween 20.

The diluted lysate was then passed through the column, after which the column was

washed with three volumes of column buffer containing 0.25% Tween 20. The column

was then washed with three volumes of column buffer without Tween 20. The MBP-









Gcrl (631-785) fusion protein was then eluted from the column using 10 nM maltose in 3

ml aliquots. Aliquots of each fraction was visualized via SDS-PAGE with Coomassie

blue staining. The fractions that contained the fusion protein were pooled and

concentrated by ultrafiltration in a Centriprep-30 concentrator (Amicon).

Protein SDS-PAGE and TCA Precipitation of Translation Products

The amount and quality of protein produced in the in vitro translation reactions

was analyzed by both sodiumdodecylsulfate polyacrylamide gel electrophoresis (SDS-

PAGE) and trichloroacetic acid (TCA) precipitation. SDS-PAGE utilized a 3%

acrylanide stacking gel and a 10% acrylamide separating gel. One microliter of each

translation reaction was diluted to 5 g1 in loading buffer, denatured at 100C for five

minutes and then loaded onto the gel that was then run at 20 mA until the bromophenol

blue reached the bottom of the gel. The gel was then vacuum-dried onto 3MM paper

(Whatman) and visualized via autoradiography.

For TCA precipitation, 1 jtl of each translation product was mixed with 50I1 0.1

mM sodium hydroxide (NaOH) and incubated at 37'C for 15 minutes. One milliliter of

cold 10% TCA was added to each sample and then incubated on ice for 60 minutes. Each

sample was then applied to a fiberglass filter that was then washed twice with cold 10%

TCA. The filters were then counted in Scintiverse (Fischer) using a scintillation counter

(Beckman, model LS5801).

Probe Synthesis

The probes were derived from the plasmids YIpCD12 and YIpCD 13 (Drazinic et

al. 1996). The plasmid YIpCD12 contains both a Raplp- and Gcrlp-binding site with the

native PYK1 spacing. The plasmid YIpCD 13 is identical except for the addition of 5

base-pairs between the two binding sites. These probes were used to investigate the









binding interactions between the Gcrlp and the various Raplp species. The plasmids

pCD12 and pCD13 were cleaved with Xhol and Hnel to produce 71 base-pair fragments

that were end labeled with Klenow fragment. The probes were then purified by gel

electrophoresis.

Binding Reaction

The binding reactions were done in binding buffer (4 mM Tris [pH 7.5], 12 mM

4-(2-hydroxyethyl)-1-piperazine ethanesulfonic acid (HEPES) [pH 7.5], 60 mM Kcl, 5

mM MgCl2, 0.6 mM EDTA, 60pmm DTT, 0.1 mg/mI polydeoxyinosinic-deoxycytidylic

acid (poly(dI-dC)), 0.3 mg/mI BSA, and 50% glycerol). The poly(dI-dC) functions as a

non-specific competitor to inhibit non-specific binding by Gcrlp or the Rap I p constructs.

The DNA probes were diluted to 1 kcpm/Ll in binding buffer. The proteins were diluted,

using binding buffer, such that each sample produced approximately the same amount of

binding activity. Combinations of the probes and proteins, as described in the Results,

were incubated in total volumes of 20til of binding buffer for 10 minutes at room

temperature before loading onto the polyacrylamide gel.

Electrophoresis

The reaction mixtures were fractionated by molecular weight using

electrophoresis through a non-denaturing 5% T (82:1) polyacrylamide gels using a 0.5X

TBE running buffer, utilizing the Vertical Gel Electrophoresis System (Bethesda

Research Laboratories, Model V16). Before loading, the gels were pre-run at 100V (6.67

V/cm) for 2 hours. After loading the voltage was increased to 150V (10 V/cm). To

monitor the extent of electrophoresis, loading dye (0.25% bromophenol blue, 0.25%

xylene cyanol, and 25% ficoll-400) was added to an empty lane. Band-shifts using the 72

base-pair probes derived from pCD12 or pCD13 were run until the bromophenol blue had









migrated 14 cm through the 15 cm gel. The gels were then vacuum dried onto 3MM

paper (Whatman).

Autoradiography and Phosphorimager Analysis

Dried band-shift gels were exposed to using two Dupont Lightning Plus

intensifying screens. The screen separating the gel and the film served the additional

purpose of blocking most of the 35S signal from the labeled protein.

Phosphorimages were produced using a Phosphorlmager (Molecular Dynamics).

Using the peak finder function it was possible to calculate relative volumes of radioactive

emission for each visible band. Control titrations were also included with each exposure

to demonstrate that the screen maintains a linear absorption over the range of counts

present in the gel (see Figure 2.3).




















a a- ii


LINE- 1I_
Sensitivity Noise Min. Area Max. Area Auto Nols Baseline Kernel
64 0.005 0 0 No Automatic 12
Peak # Area Height Percent Vaiance m]ex(mm Separatn
1 26318.24 1946.834 53.259 2.74E+05 19.12 W
2 12233.92 728.794 24.757 4.82E+04 38.8 W
3 5951.062 440.456 12.043 1.63E+04 59.3 W
4 2713.516 192.548 5.491 3.77E+03 78.95 W
5 1237.54 82.024 2.504 &45E+02 98.77 W
8 590.323 39.924 1.196 198E+02 118.77 W
7 232.827 15.475 0.471 3.34E+01 139.3 W
8 83.973 6.212 0.17 4.31E+00 158.95 W
9 36.748 2.997 0.074 .68E-01 178.95 W
10 17.315 1.88 0.035 1.IOE-01 199.5 W


Figure 2.3 Phosphorlmager quantification


The band shift gels were exposed to an Phosphorlmager (Molecular Dynamics) screen.
The image was then decoded and the density of signal was calculated from a line

centered over the signal and with width encompassing the signal. A. This is an example
of a titration of which was used as a control exposure for linearity of the phosphorescent
screen. Each dot represents a progression of 50:50 dilutions. The solid line was place
manually, and dotted lines represent the width of the line in which data were calculated.
B. The signal density for each of the dots is represented graphically. The calculated area
under each curve is displayed in the table. A similar method was used for calculating the
signal density for the bands in the bandshift gels.














RESULTS

Characterization of the Chromatin Structure at TPIJ in Saccharomyces cerevisiae

Chromatin structure has been shown to be an important factor in the

transcriptional activity of many eukaryotic genes. The importance of chromatin structure

should not be different for the glycolytic enzyme genes. As noted in the Introduction, a

hypersensitive site has been demonstrated for TDH3, and this structure changed in the

context of a gcrl mutant. The following experiments were performed in order to

elucidate the wild-type chromatin structure at TPIJ and to see if the gcrl mutation had a

similar effect on the TPI1 chromatin structure as seen previously at the TDH3 locus.

TPIJ was selected to serve as a model in which the promoter/UAS regions were defined

in great detail. In addition, two other mutations that have been shown to effect glycolytic

enzyme gene expression, gcr2 and gall1, were evaluated for alterations in chromatin

structure.

Nuclease Sensitivity Patterns at TPJJ in Purified DNA

Since both micrococcal nuclease and DNase I show varying degrees of sequence

specificity, it is necessary to know the cleavage patterns in the absence of nucleosomes or

other bound proteins. Otherwise interpretation of the in vivo pattern can be difficult.

Samples of purified genomic DNA were treated with escalating concentrations of

micrococcal nuclease or DNase I. They were then cleaved at the unique sites BsiHKAI

or KpnI for indirect end-labeling and separated by electrophoresis in a 1.5% agarose gel.

BsiHKAI cleaves within the TPI1 coding sequence at position +185 relative to the









translation start site, and KpnI cleaves within the TPIJ coding sequence at position +511

relative to the translation start site. The DNA that had been fractionated by

electrophoresis was then transferred to a nylon membrane, and indirectly end-labeled

with a radioactive DNA probe. Nuclease treatment of purified "naked" genomic DNA

from the wild-type strain S150-2B produced cleavage patterns at the TPIJ locus that were

consistent with the cleavage characteristics of micrococcal nuclease and DNase I. The

pattern produced by micrococcal nuclease digestion of purified DNA is labeled "wild-

type" in Figure 3.1, lanes 1-3 and Figures 3.2, 3.3, lanes 1-4. The pattern produced by

DNase I digestion of purified DNA is labeled "wild-type" in Figures 3.3 and 3.5, lanes 1-

4. Micrococcal nuclease produced a distinctive pattern that results from this enzyme's

higher affinity for dAodT-rich duplex DNA sequences (Gross 1988). DNase I produced a

more uniform smear, which results from its lower sequence specificity (Gross 1988).

The same nuclease digestion patterns were seen when purified DNA from the strains

HBY4, DFY64 1, and R884-1 C were treated with micrococcal nuclease and DNase I.

These strains contain gcrl, gcr2, and gall1 mutations, respectively. This sensitivity

pattern was used for comparison when evaluating the in vivo treated samples.

Statistically Positioned Nucleosomes are Present Both Upstream and Downstream
from the TPIJ Structural Gene

The wild-type strain S150-2B was treated in vivo with DNase I and micrococcal

nuclease after spheroplasting and simultaneously with permeabilization with dilute NP-

40. They were treated with escalating concentrations of each nuclease. They were then

lysed and their DNA purified. The DNA samples were cleaved at the unique restriction

sites BsiHKAI or KpnI. The DNA fragments were separated by electrophoresis in a

1.5% agarose gel, transferred to a nylon membrane, and indirectly end-labeled using a































Figure 3.1. MNase sensitivity pattern upstream from BsiHKAI


Micrococcal digestion looking upstream from BsiHKAI restriction site. Spheroplasts
permeabilized with NP-40 were treated with increasing concentrations of micrococcal
nuclease. The strains used were S150-2B, HBY4, DFY-41, and R884-IC. They
represent wild-type (lanes 4-7) and the gcrl (lanes 8-11), gcr2 (lanes 12-15), and
gall1 (lanes 16-19) mutants respectively. Purified DNA from the wild-type strain
S 150-2B was used to display the digestion pattern that results from the sequence
specificity of micrococcal nuclease. Lanes 1-3 contain purified genomic DNA that
was treated with increasing concentrations of micrococcal nuclease. Lane 20 contains
a lambda HindIll size marker. The samples, which were treated in situ, are displayed
in groups of four lanes for each strain treated. From left to right each group of four
lanes represents treatment with 0.012, 0.04, 0.12, and 0.4 units of micrococcal
nuclease respectively.

The nuclease sensitivity pattern is interpreted to the immediate right of the
autoradiograph. The ovals are located at regions of nuclease protection and they
represent statistically positioned nucleosomes. The vertical striped box represents an
area of nuclease sensitivity characterized by a digestion pattern identical to that of
purified DNA.

Genetic landmarks are illustrated to the right of the nuclease sensitivity illustration.
The open reading frames for DBF4, YDR051C and TPIJ are represented by black
arrows. The UASTpuJ, TATA, and transcription start site (Tx start) are indicated by a
black box and two black lines respectively.













nuclease treated chromatin
a wild-type gcrlA gcr2A galll-313

[ -- ]Ir -- -- I ------ I
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 1819 20




0
~DBF4












O YDR051C


0







UAS





TATA





Tx start



TPIJ


























Figure 3.2. MNase sensitivity pattern upstream from KpnI


Micrococcal nuclease digestion looking upstream from KpnI restriction site.
Spheroplasts permeabilized with NP-40 were treated with increasing concentrations
of micrococcal nuclease. The strains used were S 150-2B, HBY4, DFY64 1, and
R884-1C. They represent wild-type (lanes 5-8) and the gcrl (lanes 9-12), gcr2 (lanes
13-16), and gall1 (lanes 17-20) mutants respectively.

Purified DNA from the wild-type strain S 150-2B was used to display the digestion
pattern resulting from the sequence specificity of micrococcal nuclease. Lanes 1-4
contain genomic DNA that was treated with increasing concentrations of micrococcal
nuclease. Lane 21 contains a lambda-HindlII size marker.

The samples, which were treated in situ, are displayed in groups of four lanes for each
strain treated. From left to right each group of four lanes represents treatment with
0.012, 0.04, 0.12, and 0.4 units of micrococcal nuclease respectively.

The nuclease sensitivity pattern is illustrated to the immediate right of the
autoradiograph. The ovals are located at regions of nuclease protection, and they
represent statistically positioned nucleosomes. The vertical striped box represents an
area of nuclease sensitivity characterized by a digestion pattern identical to that of
purified DNA. The vertical empty box represents continued nuclease sensitivity, but
the degree of sensitivity in this area differs in the samples treated with DNasel
(Figure 3.3).

Genetic landmarks are illustrated to the right of the nuclease sensitivity illustration.
The open reading frames for DBF4, YDR051C and TPIJ are represented by black
arrows. The UASrp11, TATA, and transcription start site (Tx start) are indicated by a
black box and two black lines respectively.










nuclease treated chromatin
0 wild-type gcrlA gcr2A galll-313 .

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 1819 20 21




I DBF4






S YDR051C

*0



*UAS


0- TsATA


h -Tx stat


TPI1



























Figure 3.3. DNaseI sensitivity pattern upstream from KpnI


DNaseI digestion looking upstream from KpnI restriction site. Spheroplasts
permeabilized with NP-40 were treated with increasing concentrations of DNaseI.
The strains used were S150-2B, HBY4, DFY641, and R884-1C. They represent
wild-type (lanes 5-8) and the gcrl (lanes 9-12), gcr2 (lanes 13-16), and gall] (lanes
17-20) mutants respectively.

Purified DNA from the wild-type strain S 150-2B was used to display the digestion
pattern that results from the sequence specificity of micrococcal nuclease. Lanes 1-4
contain genomic DNA that was treated with increasing concentrations of DNase.
Lane 21 contains a lambda HindIII size marker.

The samples that were treated in situ are displayed in groups of four lanes for each
strain treated. From left to right each group of four lanes represents treatment with
0.01, 0.02, 0.04, 0.08 units of DnaseI respectively.

The nuclease sensitivity pattern is illustrated to the immediate right of the
autoradiograph. The ovals represent statistically positioned nucleosomes. The
vertical striped box represents an area of nuclease sensitivity, which is characterized
by a digestion pattern identical to that of purified DNA. The vertical empty box
represents continued nuclease sensitivity, but the degree of sensitivity in this area is
less than that seen in the are represented by the striped box.

Genetic landmarks are illustrated to the right of the nuclease sensitivity illustration.
The open reading frames for DBF4, YDR051C and TPIJ are represented by black
arrows. The UASTPIj, TATA, and transcription start site (Tx start) are indicated by a
black box and two black lines respectively.












-< nuclease treated chromatin
wild-type gcrlA gcr2A gall 1-313 -
II-3
I 1 -- I I
1 2 3456 78 9101112131415161718192021














YDR051C







I UAS


TATA



Tx start










TPI'






























Figure 3.4. MNase sensitivity pattern downstream from KpnI


Micrococcal nuclease digestion looking downstream from KpnI restriction site.
Spheroplasts permeabilized with NP-40 were treated with increasing concentrations
of micrococcal nuclease. The strains used were SI 50-2B, HBY4, DFY64 1, and
R884-1C. They represent wild-type (lanes 5-8) and the gcrl (lanes 9-12), gcr2 (lanes
13-16), and gall 1 (lanes 17-20) mutants respectively.

Purified DNA from the wild-type strain S150-2B was used to display the digestion
pattern, which results from the sequence specificity of micrococcal nuclease. Lanes
1-4 contain genomic DNA that was treated with increasing concentrations of
micrococcal nuclease respectively. Lane 21 contains a lambda-HindlIl size marker.

The samples that were treated in situ are displayed in groups of four lanes for each
strain treated. From left to right each group of four lanes represents treatment with
0.012, 0.04, 0.12, and 0.4 units of micrococcal nuclease respectively.

The nuclease sensitivity pattern is illustrated to the immediate right of the
autoradiograph. The ovals represent statistically positioned nucleosomes. The
vertical striped boxes represent areas of nuclease sensitivity which were characterized
by a digestion pattern identical to that of purified DNA.

Genetic landmarks are illustrated to the right of the nuclease sensitivity illustration.
The open reading frames for YDRO49 Wand TPI1 are represented by black arrows.
The asterisk represents the location of a putative poly(A) signal sequence.







nuclease treated chromatin
Z wild-type gcrlA gcr2A gall1-313
1 2 3 4 5 6 7 8 9 10 11 12 1314 15 16 17 18 19 20 21








tI
i! fo



,6
YD09


WpI'






























Figure 3.5. DnaseI sensitivity pattern downstream from KpnI


DNaseI digestion looking downstream from KpnI restriction site. Spheroplasts
permeabilized with NP-40 were treated with increasing concentrations of DNaseI.
The strains used were S150-2B, HBY4, DFY641, and R884-1C. They represent
wild-type (lanes 5-8) and the gcrl (lanes 9-12), gcr2 (lanes 13-16), and gall 1 (lanes
17-20) mutants respectively.

Purified DNA from the wild-type strain S 150-2B was used to display the digestion
pattern, which results from the sequence specificity of DNaseI. Lanes 1-4 contain
genomic DNA that was treated with increasing concentrations of DNasel. Lane 21
contains a lambda-HindIII size marker.

The samples, which were treated in situ, are displayed in groups of four lanes for each
strain treated. From left to right each group of four lanes represents treatment with
0.01, 0.02, 0.04, 0.08 units of DNaseI respectively.

The nuclease sensitivity pattern is illustrated to the immediate right of the
autoradiograph. The ovals represent statistically positioned nucleosomes. The
vertical striped boxes represent areas of nuclease sensitivity which were characterized
by a digestion pattern identical to that of purified DNA.

Genetic landmarks are illustrated to the right of the nuclease sensitivity illustration.
The open reading frames for YDRO49W and TPIJ are represented by black arrows.
The asterisk represents the location of a putative poly(A) signal sequence.











nuclease treated chromatin
wild-type gcrlA gcr2A gall 1-313
.... 1 .. m _
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 1819 20 21













YDRO49W







: I


TPI1









radioactive DNA probe. Both enzymes produced distinct ladders of nuclease protections

with intervening areas of nuclease sensitivity. The micrococcal nuclease pattern of S 150-

2B is seen in Figure 3.1, lanes 4-7 and Figure 3.2, lanes 5-8, and the DNase I pattern is

seen in Figure 3.3, lanes 5-8.

Extending upstream from the 5' end of the UASTpu element there is a ladder of

alternating areas of nuclease sensitivity and protection. There are seven protected areas.

After the seventh area of protection the micrococcal nuclease pattern changes subtly, and

then it continues for what appears to be an additional four areas of nuclease protection.

In the DNase I samples, the area, which had the pattern change, demonstrated a small

region of nuclease hypersensitivity (Figure 3.3). This transition is spatially related to the

5' end of the open-reading frame YDR05 1 C and the 3' end of the DBF4 gene (Chapman

and Johnston 1989; Cherry et al. 1998). It is noted that the ladder of nuclease protections

lies over the coding sequence for open-reading frame YDR05 1 C. The average repeat

length for the seven areas of protection is 170 base-pairs. This is slightly longer than the

known repeat length of 160 base-pairs for S. cerevisiae (Bernardi et al. 1991; Thoma

1992).

There is an area of hypersensitivity that extends from the 5' end of the UASTp/1

element to the transcription start site. This is best seen in the DNase I samples (Figure

3.3, lanes 5-8), but it is also seen in the micrococcal nuclease samples (Figure 3.2, lanes

5-8). This area of hypersensitivity is consistent with a nucleosome-free region.

There is also some nuclease protection located over the TP11 structural gene.

Both micrococcal nuclease and DNase I sensitivity patterns over the coding sequence are

identical to that seen in the "naked" DNA samples, but the intensity of these bands is less









in the in vivo treated samples. This implies a modest, although incomplete nuclease

protection. However, the intensity of the bands are less than that seen over the upstream

promoter region. This protection over the structural gene is also more evident in the

DNase I samples (Figure 3.3).

Extending downstream from the TPIJ gene, a second ladder of nuclease

protections is identified by micrococcal nuclease digestion (Figure 3.4, lanes 5-8). It is

also seen faintly in the DNase I digested samples (Figure 3.5, lanes 5-8). This ladder of

nuclease protections with intervening nuclease sensitivity extends for at least nine areas

of protection. After the ninth area of protection, the pattern becomes less clear due to the

sequence specificity of micrococcal nuclease, but the nuclease sensitivity pattern may

continue for another four more cycles before the next KpnI cleavage site prevents further

evaluation. This is consistent with at least nine and possibly 13 positioned nucleosomes

extending away from TPIJ in the 3' direction. These have an average size of 160 base-

pairs. This downstream ladder of nuclease protections overlie an ORF designated

YDR049W (Cherry et al. 1998).

Nuclease Sensitivity Patterns Are Not Affected By Mutations In gcrl. gcr2, or gall I

The strains containing the mutations gcrlA, gcr2A, and gallI-313 were treated in

vivo with micrococcal nuclease and DNase I in a manner identical to that described above

for the wild-type strain S1 50-2B. As described in the introduction, all three of these

mutations have been shown to affect the level of transcription of glycolytic enzyme

genes. There does not appear to be any significant difference in the nuclease sensitivity

patterns seen either upstream or downstream from the TPIJ structural gene (Figures 3. 1,

3.2, 3.3, 3.4, 3.5).









Raplp Binding is Required for Binding of Gcrlp at TPI] in vivo

An in vivo dimethylsulfate (DMS) methylation protection assay was used

investigate the hypothesis that RapIp is required for Gcrl p binding at the UAS TpI. This

assay is commonly referred to as in vivo footprinting. In vivo analysis of the

hypothesized Rap lp/Gcrlp interaction at the TPIJ locus required a conditional mutation

in RapIp because both RapIp and triosphosphate isomerase are essential proteins. For

this study the RAP1 allele rap1-2' was used. It is one of a collection of temperature

sensitive (ts) alleles of RAP] that were isolated and characterized by Kurtz and Shore

(1991). They demonstrated that upon a shift to the non-permissive temperature of 37C,

the rap]-2's protein no longer binds the RapIp DNA-binding site. The isogenic strains

used in this study were YDS485, containing wild-type RAP], and YDS487, containing

the rap1-2's allele.

DMS Footprinting at the Permissive Temperature Demonstrated Protection at All
Known Binding Sites

Previous DMS footprinting of TP11 had been done at 30'C (Huie et al. 1992;

Lopez et al. 1993). Thus, it was necessary to demonstrate that culture growth and DMS

treatment at 24C, the permissive temperature for YDS487, do not affect protein binding.

Cultures of YDS485 and YDS487 were grown at 24C to log-phase at an A600=1,

harvested and treated with DMS at 24C. Purified DNA samples from these treated cell

were digested at the unique site Avail and then treated with piperidine at 95'C for 30

minutes to cleave DNA at the guanine residues methylated by DMS. The DNA

fragments were then separated on a sequencing gel and transferred to nylon membrane.

Indirect end-labeling demonstrated protection over the known binding sites for Reb I p,

Raplp, and Gcrlp (Figure 3.6, lanes 3 and 6).
































Figure 3.6 In-vivo footprint analysis at the UASTPIJ


In-vivo footprint analysis at TPIJ in wild-type and rap]-2ts mutant strains. In vivo
guanine methylation protection assays of the UASTPI1 were performed in the wild
type strain S 150-2B and the rapl-2ts mutant strain YDS485 at permissive and non-
permissive temperatures and while inhibiting protein synthesis with cycloheximide.

Lanes 1,2,8, and 13 contain samples prepared by DMS methylation of purified DNA.
These reference lanes display all the guanine residues in the sequence. Lane 3
contains YDS485 DNA that was treated in vivo with DMS for 5 minutes at 24C
after a 30 minute incubation at 24'C. Lanes 4 and 5 contain YDS485 DNA that was
treated in-vivo with DMS for 5 minutes at 37C after a 30 and 60 minute incubations
respectively at 37'C. Lanes 6 and 7 contain YDS485 DNA that was treated in vivo
with DMS for 2 minutes at 24'C and 37C after 30 minute incubations at 24'C and
37C respectively. Lanes 9 and 10 contain S150-2B DNA which was treated in vivo
with DMS for 5 minutes at 37C after a 30 minute incubation at 37C. Lanes 11 and
12 contain YDS485 DNA that was treated in vivo with DMS for 5 minutes at 24'C
after a 30 minute exposure to 10 pig/ml cycloheximide.

The sequencing reactions and in-vivo footprints were resolved by electrophoresis on
denaturing polyacrylamide gels, transferred to Hybond-N+, and visualized after
indirect end-labeling by autoradiography. The positions of Reb Ip binding sites
(stippled ovals), Gcrlp-binding sites (open ovals), and Raplp-binding sites (closed
ovals) are indicated.









1 2 3 4 5 6 7 8 9 10 11 12 13


4


Gcrlp 8 S
ii*4
8S





Gcrlp 8
R ap lp~ti .....................
'400,




Grelp 0 !iiiil~


G G Rap U UU G U U G


Rap l-2t= RAP]









DMS Footprinting at The Non-permissive Temperature Demonstrated Loss of Protection
at Both the Raplp- and Gcrlp-binding Sites

DMS treatment was performed at the non-permissive temperature to investigate

what effects the loss of Raplp binding would have on the binding of Gcrlp. Cultures of

YDS485 and YDS487 were again grown at 24C to an A600=l. Aliquots of these cultures

were then incubated at 37C, the non-permissive temperature, for 30 minutes and then

treated with DMS at the same temperature. The DNA was then isolated and processed as

described above. Indirect end-labeling demonstrated a loss of protection over both the

Raplp- and Gcrlp-binding sites in the YDS487 sample treated at 37'C (Figure 3.6, lanes

4, 5, and 7). This loss of protection indicated that RapIp and Gcrlp are not bound to

their sites. Notably, deprotection at these sites did not occur when the isogenic wild-type

strain YDS485 was treated under the same 37*C conditions (Figure 3.6, lanes 9 and 10).

This indicates that the deprotection was specific to the rap1-2 mutation and not the

strain background. Previous experiments have shown that RapIp remains bound in the

absence of Gcrl p (Huie et al. 1992). Thus, the in vivo binding of Gcrl p was dependent

on the binding of Raplp at an adjacent site within the UASTpIJ, but the binding of Raplp

is not dependent on Gcrlp.

Inhibition of Protein Synthesis Does Not Affect DMS Protection at the GcrI -
Binding Site

Cycloheximide was used to demonstrate that the loss of Gcr I p-binding was not

due to changes in protein synthesis. The transfer of a YDS487 culture to 37C is an

irreversible and terminal event. If the culture is returned to 241C, it does not resume

growth. Thus, one possible explanation for the loss of Gcrlp binding was that cell death

associated with the shift to the non-permissive temperature resulted in the cessation of









protein synthesis. It was possible that if protein synthesis ceased, the intracellular

concentration of Gcrlp might drop, and Gcrlp dissociation would occur by a shift in

mass action equilibrium rather than a loss of Rap 1 p facilitation. To test this, the wild-

type YDS485 strain was grown at 24'C until a density of A600=l had been reached. The

cells were then treated with cycloheximide for 30 minutes at a concentration (101ag/ml)

adequate to arrest growth. Cycloheximide inhibits protein synthesis by preventing

polypeptide elongation (Ferguson et al. 1967). The standard DMS footprinting protocol

was then performed. Indirect end-labeling of the resulting blot demonstrated that the

binding sites for RebIp, Rap ip, and Gcrlp remained protected (Figure 3.6, lanes 11 and

12). This protection indicates that inhibition of protein synthesis did not affect the

binding equilibrium of these proteins.

The timeline of these experiments suggests that these changes are a direct result

of Rap Ip dissociation rather than secondary to other cellular events such as DNA

replication and the process of cell division. In the experiments presented here the

deprotection was evident after 30 minutes of incubation at the non-permissive

temperature. The shortest incubation time performed was 20 minutes, which also

demonstrated Raplp-dependent protection over the Gcrlp-binding site (not shown). This

interval is much shorter than the two hour doubling time for the wild-type strain grown at

30'C or four hours for the mutant strain when grown at 240C.

Raplp Facilitates the Binding of Gcrlp in vitro

Having shown that RapIp was required in vivo for the binding of Gcrl p, in vitro

electrophoresis mobility gel-retardation analysis was used to characterize this binding

dependence further. The mobility retardation assay, also known as a band-shift, was









utilized to compare Gcrlp binding with and without RapIp present in the binding

reaction. In these experiments a DNA probe was pre-incubated with the proteins of

interest to allow binding. The mixture was then separated by electrophoresis through a

non-denaturing polyacrylamide gel. Electrophoresis separates the bound from the

unbound probe because the protein-bound probe has a higher molecular weight and

migrates more slowly through the gel matrix. If one binding protein is present in the

binding reaction there would be two bands produced: 1) free probe and 2) protein bound

probe. If two binding proteins were present in the binding reaction there would be four

bands produced: 1) free probe, 2) probe + protein-i, 3) probe + protein-2, and 4) a ternary

complex of probe bound by both proteins.

The ability of one protein to facilitate the binding of another can be tested by

comparing the amount of probe shifted by each protein alone to the amount of probe in

the ternary complex, which is produced when both proteins are present in the same

reaction. If the two proteins bind independently and no facilitation occurs then they

should produce a ternary complex whose radioactive signal equals the multiplication

product of the radioactive signals produced when each protein is incubated alone.

Conversely, if facilitation does occur then the radioactive signal of the ternary complex

would be greater than this multiplication product.

The end-labeled DNA probes used in these experiments were modeled after the

sequence of the UAS element of PYK1 (Drazinic et al. 1996). One probe was derived

from pCD12, which includes the binding sites for Raplp, and Gcrlp as found at the UAS

element of PYK1. A second probe was derived from pCD 13 in which the same Rap Ip-

and Gcrlp-binding sites were separated by five base-pairs. Separation of the two sites by









five base-pairs has been shown qualitatively to decrease the binding of Gcr 1 p both in vivo

and in vitro footprinting (Drazinic et al. 1996).

The Gcrlp used in these experiments is a fusion between the Maltose-Binding

Protein (MBP) and the DNA-binding domain of Gcrlp (MBP::GCR1631.785). It was not

possible to utilize full-length Gcrlp because of its instability. This instability has been

seen when MBP-Gcrlp.785 has been synthesized either in E. coli or in rabbit reticulocyte

lysate systems (Huie and Baker 1996), and as a result it has not been possible to purify

full-length Grc I p for use in vitro experiments.

The RapIp species used in these experiments included full-length RapIp as well

as amino- and carboxy-terminal truncations. As displayed in Figure 3.7, the amino-

terminal truncation deletes the asymmetric DNA-bending domain (Muller et al. 1994),

and the carboxy-terminal truncation deletes the transcriptional activation domain (Sussel

and Shore 1991), and the transcription silencing domain (Sussel and Shore 1991). If both

ends are deleted only the minimal DNA-binding domain remains (Henry et al. 1990).

The Gcrlp and RapIp proteins produced discrete, single band-shifts when used

alone, and they produced discrete, single ternary complexes when used in combination.

The DNA probe alone produced a single band (Figure 3.7, panel b, lane 1). The

individual addition of each of the proteins Gcrlp, Rap1 P361-596, Rap1P36i-827, Rapl p 1-596,

and Raplp1.827 produced a single shifted band (Figures 3.7 and 3.8, panel b, lanes 2, 4, 5,

7, 9, and 11, respectively) in addition to the free probe. When Gcrlp was combined with

each individual Rap Ip species an additional band representing the ternary complex was

seen in each case (Figure 3.7, panel b, lanes 6, 8, 10, and 12).








































Figure 3.7 Bandshift of Raplp and Gcrlp using native spacing.

A. The drawing is represents the RapIp protein and highlights its functional
domains. This panel also displays the four RapIp species used in the bandshift
analysis. B. This is an autoradiograph of the in vitro bandshift gel in which Gcrlp
was incubated with and without the four RapIp species. Lanes I and 2 show
unbound probe and probe bound by Gcrlp respectively. Lane 3 contains probe and
BMV proteins translated in vitro using rabbit reticulocyte lysate. Lane 4 is the same
as lane 3 with the addition of Gcrlp. Lanes 5 12 all contain probe and Gcrlp with
the addition of the indicated RapIp species in lanes 6, 8, 10, 12. The bands
representing probe bound by Gcrlp are indicated by the empty arrow to the right of
the autoradiograph, and the bands representing the various Rap I p species are
indicated by filled arrows to the right of the autoradiograph. The ternary complexes
are indicated by black dots on the autoradiograph. C. This is a table that displays the
phosphorimage data collected using a Molecular Dynamics Phosphorlmager. The
data in this table is in phosphorimager units.


















RapIp

44 274 361 596630 695


665 827
11111 =Asymetric DNA-bending domain, aa44-274
.' -DNA-binding domain, aa361-596
Activation domain, aa630-695
Silencing domain, aa665-827


Rap Ip(1-596)
Raplp(361-596)


Rap I p(I-827)



Rap lp(361-827)


Gcrip RapIp RapIp RapIp
(631-785) BMV (361-596) (361-827) (1-5%)
+ r+ -I- + +


Rapip
(1-827)
+


- Raplp (1-827)
- Raplp (1-596)

- Raplp (361-827)
~ Gcrlp (631-785)


Raplp (361-596)


- Free probe


C.
Rapip Rapip Gcrlp Ternary Free Total counts Ternary -Gcrip free Gcrlp RapIp Free Rapip Expected Act./Exp Avg.
construct peak peak peak probe in lane % alone probe % alone probe % Gcrlp
361-596 1464 1878 661 6520 10522 6.3% 2546 9786 26.0% 2042 8552 19.3% 5.0i 1.3 1.5
1739,__1664_ 852 6074_ 8591' 9.9'4 2262 9713 23.3% 2992 862 25.7% 60% 17
527 585 266 2133 3511 7.6, 769 3294 23.4% 756 2870 20.9% 4.9% 1.6
361-827 2533 979. 1985 4145- 9643, 20.6% 2546 9786 26.0%i 3820 7019-352%4 9.2% 2. 2.4
2140 12W60_ 1729) 4530 7518 23.0% 2262 9713 23.3% 344 689 33.3% 7,8% 3.0
522 481 318 2023 33441 95% 769 3294 23,% 711 2617 21.4%1 50%' 1.9
1I-596 1-387: 1889 1910 6 7 12 119 16.1%/ 2546 9786126.0%, 2417T 9541~ 20 2%' 5.3%' 3. 3.3
1162 1578 1890 6273i 9741 19.4% 2262 9713 23.3% 2916 9004 24.5% 57% 3.4
369 535' 5 27 2087 3519, 150% 769 3294 23.4%x 684 2852: 193%:: 4.58 3.3
1 827 1633 1482 1668 5702' 10484 15.9% 2546 9786 26.0% 2620 7619 25.6%/, 6 7/, 2.f4_3.2
1213 1462 1164 6324 8950 13.0 2262 9713 23.3 1W 2 919f 15,.2 3.% 37
363 497 324 2120. 2941 110%1 769 3294 234% 456 2873 13.7%' 32 3.4


'41, -AkA.AA.A







































Figure 3.8. Bandshift of RapIp and Gcrlp +5 basepairs separation

A. The drawing is represents the RapIp protein and highlights its functional domains.
This panel also displays the four Raplp species used in the bandshift analysis. B.
This is an autoradiograph of the in vitro bandshift gel in which Gcrlp was incubated
with and without the four RapIp species. Lanes 1 and 2 show unbound probe and
probe bound by Gcrlp, respectively. Lane 3 contains probe and BMV proteins
translated in vitro using rabbit reticulocyte lysate. Lane 4 is the same as lane 3 with
the addition of Gcrlp. Lanes 5-12 all contain probe and Gcrlp with the addition of
the indicated RapIp species in lanes 6, 8, 10, 12. The bands representing probe
bound by Gcrlp are indicated by the empty arrow to the right of the autoradiograph,
and the bands representing the various Raplp species are indicated by filled arrows to
the right of the autoradiograph. The ternary complexes are indicated by black dots on
the autoradiograph. C. This is a table that displays the phosphorimage data collected
using a Molecular Dynamics Phosphorlmager. The data in this table are in
phosphorimager units.