{I t:
x

r. Wumwmwwmmmq

(
w.

.

fly} 4
Uri: A
.i.

165%"

 

15.9.. a
4:.

0

05-1,}...

glans}:

. A"...
$ , .

ii 9].
bghfiu...
. i... 5.3...

131.3551.

 

 

iv
safiucmgrw. .L

 

 

 

LIBRARY
MiChi9am State _ , _
University Thé?s§e‘§aififi'flr§33ééhe

 

THE FUNCTION AND DESIGN OF DIS-ACTING ENHANCER
ELEMENTS REGULATED BY SHORT-RANGE
TRANSCRIPTIONAL REPRESSORS: GRAMMAR STUDIES
FROM Drosophila melanogaster

presented by

Meghana Manohar Kulkami

has been accepted towards fulfillment
of the requirements for the

Doctoral degree in Genetics

 

 

M

Major Professor’s Signature
I? . IO - 0 3

Date

 

 

MSU is an Affirmative Action/Equal Opportunity Institution

 

-.- -n-a-o-o-o-c-u-----o----

 

PLACE IN RETURN BOX to remove this checkout from your record.
TO AVOID FINES return on or before date due.
MAY BE RECALLED with earlier due date if requested.

 

DATE DUE

DATE DUE

DATE DUE

 

 

 

 

 

 

 

 

 

 

 

 

 

 

6/01 c:lC|RC/DateDue.p65-p.15

 

 

THE FUNCTION AND DESIGN OF CIS-ACTING ENHANCER ELEMENTS
REGULATED BY SHORT-RANGE TRANSCRIPTIONAL REPRESSORS:
GRAMMAR STUDIES FROM Drosophila melanogaster

By

Meghana Manohar Kulkami

A DISSERTATION

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

DOCTOR OF PHILOSOPHY
Program in Genetics

2003

ABSTRACT
THE FUNCTION AND DESIGN OF CIS-ACTING ENHANCER ELEMENTS

REGULATED BY SHORT-RANGE TRANSCRIPTIONAL REPRESSORS:
GRAMMAR STUDIES FROM Drosophila melanogaster

By

Meghana Manohar Kulkami

Given that the DNA of (most) all cells in an animal is identical, how do different
cells acquire the unique morphologies and fimctional properties to create the diverse
tissues and organs in multicellular organisms? With better understanding of the nature of
genes and the process of gene regulation, the central role of transcriptional regulation in
directing development is becoming evident. These exquisitely orchestrated gene
regulatory programs are encoded in the DNA sequence of cis-regulatory control
elements, whose features are only now becoming apparent. Morphological diversity has
its origins in reshaped developmental processes, and these changes often reflect
alterations in genetic regulatory programs and in particular the transcriptional cis-
regulatory regions. Thus, evolution of diversity is directly related to the evolution of
these cis-regulatory regions or enhancers. To understand this aspect of evolution, we
must first understand the structure-function relationships that apply to developmental cis-
regulatory enhancers. The studies described herein highlight general principles of the
design and function of cis-acting enhancer elements that are involved in the control of
complex patterns of gene expression and should in turn facilitate our understanding of the
regulatory logic underlying morphological complexity and diversity.

Cis-regulatory enhancer elements are thought to function as information

processing centers that integrate the negative and positive transcriptional inputs incident

upon them, in a computer-like fashion, resolving the multiple inputs into a single output
that instructs the basal transcriptional machinery to either turn the linked gene on or off.
In this type of computational model for gene regulation by an enhancer, the basal
transcriptional machinery plays a passive role by simply responding to signals generated
by the enhancer. In contrast, the ‘Information Display’ or ‘Billboard’ model for enhancer
function proposed in this study, demonstrates that instead of integrating multiple inputs,
the enhancer is capable of simultaneously displaying contrasting information that is
interpreted by multiple successive or simultaneous interactions with the basal promoter.
Thus, the basal transcriptional machinery plays an active role in processing regulatory
information presented by the cis-element.

Key structural features of the internal organization of an enhancer control its
ability to convert positional information specified by the transcription factors that bind it
into differential gene activity. Although many functional analyses have indicated the
presence of functional constraints, no work has been carried out to systematically define
the spatial constraints between transcription factor binding sites within enhancers that are
required for their proper function. The experiments described here take a first step
towards systematically deducing the ‘grammar’ rules for an important class of
transcriptional regulators, the short-range repressors. These rules define the internal
organization of a functional module regulated by them.

These studies described here have important ramifications for the biochemistry
and evolution of cis-regulatory elements, and should facilitate the development of more
sophisticated computational algorithms for the identification of cis-regulatory elements

and for the interpretation of their biological function.

To my Dad (Mr. Manohar Kulkami) and Mom (Mrs. Swati Kulkami), without whose
wisdom, love, and encouragement, this thesis would never have seen the light of day.

ACKNOWLEDGEMENTS

I want to take this opportunity to sincerely thank everyone who has at one time or
another helped nurture this project from inception to its final birth. I am indebted to
David Amosti, my thesis advisor and mentor, for unreservedly believing in me and for
helping me explore new scientific perspectives. In his unique way, he has enriched,
enthused, and inspired me. I would like to extend a particularly deep thanks to my thesis
guidance committee: Dr. Steven J. Triezenberg, Dr. Susan E. Conrad, and Dr. Bill Henry
for their many substantive suggestions and for helping me remain focused through the
Ph.D. program. I am especially grateful to the members of the Amosti laboratory, past
(Y ifan Mao, Bethany Strunk, Anumeha Kumar, Jae-Ryeon Ryu, Jelani Thomas, and
Montserrat Sutrias-Grau) and present (Scott Keller, Carlos Martinez, Paolo Struffi,
Martin Buckley, Priya Maui, and Sandhya Payankaulam) for hours of scientific
discussion, criticisms, and emotional support. I also extend my appreciation to other
members of the Amosti lab (Eduardo Femandez-Villatoro, Chu Yin, and Him) without
whose technical support this work would not have been possible. I am especially grateful
to my colleagues from various departments (Craig Hinkley, Heather Hirsch, and Kirsten
Fertuck) for sharing their expertise and for providing an encouraging word. I am also
grateful to the Genetics Program secretary, Jeannine Lee, and members of the
Department of Biochemistry and Molecular Biology, Julie Oesterle, Carol McCutcheon
and Pappan for all their help. My deepest thanks go to my family: my parents, brothers
and sisters-in-law, without whom I would not have come this far; to my husband,

Shailesh, who has so infinitely enriched my life.

TABLE OF CONTENTS

List of Figures vii
Key to Abbreviations ix
CHAPTER I:

Introduction 1

CHAPTER 11:

Information Display by Transcriptional Enhancers 55
Abstract 56
Introduction 56
Materials and Methods 60
Results 62
Discussion 70
References 79

CHAPTER III:

Operating Principle of Short-range transcriptional repressors in Drosophila

melanogaster
Introduction 83
Materials and Methods 90
Results 97
Discussion 120
References 13 1

CHAPTER IV:

Conclusions and Future Directions 139

APPENDIX A:

Composite element representing two different information states exhibits orientation-

independent activity characteristic of enhancers 177

APPENDIX B:

CtBP-dependent and CtBP-independent activities contribute quantitatively to Knirps

repressor function 183

APPENDIX C:

Relative potencies of short-range transcriptional repressors 190
APPENDIX D:
Analysis of other Drosophila transcriptional repressors 195

vi

LIST OF FIGURES
Figure I-l: Sochastic versus rheostat function of enhancer elements 6

Figure I-2: Anterior—Posterior patterning in Drosophila is initiated by a cascade of

transcription factors 11
Figure I-l: Cis-regulatory region of the pair-rule gene even-skipped 15
Figure I-4: Model for the regulation of eve striped 2 18
Figure I-5: Transcriptional repression in Drosophila 22

Figure I-6: Expression pattern of the short-range repressors Giant and Knirps 25
Figure I-7: Model for the regulation of the eve stripes 3 + 7 and 4 + 6 31
Figure II-l: Simultaneous repression and activation from a compact regulatory

element ‘ 67
Figure II-2: Compact regulatory element displays enhancer-like properties of distance
and orientation independence 71
Figure II-3: Conversion of a multiple state element to a binary on or off switch 74
Figure II-4: Enhanceosome versus Information Display enhancer models 78
Figure III-1: Context dependence of short-range repression 102
Figure III-2: Stoichometry between the number of activators and repressors influences
repression effectiveness 105
Figure III-3: Effectiveness of repression correlates with the affinity of activator binding
sites 108
Figure III-4: Repression not dependent on the nature of the activation domain 112
Figure III-5: The arrangement or distribution of short-range repressor binary sites is

critical in dictating repression effectiveness 115

vii

Figure 111-6: The ability to repress depends on enhancer configurations rather than the
nature of the activation domain 118
Figure III-7: Repressor can be distinguished by their differential ability to repress in
different enhancer configurations 121
Figure IV-l: Evolutionary dynamics of transcriptional factor binding sites in a conserved
cis-regulatory element 155
Figure IV-2: Data and model for gap gene circuit 168
Figure A-l: Compact regulatory element displays enhancer-like property of orientation
independence 1 82
Figure B-l: Pattern of endogenous eve expression in embryos expressing full length
Knirps 1-429 and CtBP-independent region of Knirps 1-330 189
Figure C-l: Relative potencies of the short-range repressors Giant, Knirps, and

Kriippel 196
Figure D-l: The activity of the GAL4 activators from three high affinity is not repressed
by Snail 201
Figure D-2: The activity of the GAL4 activators from three high affinity is not repressed
by Engrailed (en), Runt (run), or Even-skipped (eve) 205
Figure D-3: The activity of transcription factors Brinker (brk) and Suppressor of Hairless
(Su(H)) on a reporter with three high affinity GAL4 sites 207
Figure D-4: The activity of the transcription factors Sloppy-paired (Slpl), Tramtrack
(TTK), and Cubitus interruptus (Ci) on a reporter with three high affinity

GAL4 sites 209

viii

A-P:
bp:
brk:
ChIP:
Ci:
CtBP:
dCtBP:
d1:
DPE:

GAL4AD:

GFP:
gt:
HDAC:
Hh:
hkb:

HMGI(Y):

hTBP:
LPN-[3:
IRF:
JAK:
kb:

kr:
kni:
NAD:
NEE:
NFkB:
PCR:
PWM:
RTK:
SAGE:
Slpl:
sna:
Su(H):
Tor:
TTK:
twi:
UAS:
VRR:
ypl, yp2:

KEY TO ABBREVIATIONS

Anterior - Posterior

base pair

Brinker

Chromatin Immunoprecipitation
Cubitus interruptus

C-terminal Binding Protein
Drosophila C-terminal Binding Protein
Dorsal

Downstream Promoter Element
GAL4 Activation Domain

Green Fluorescent Protein

Giant

Histone Deacetylase Complex
Hedgehog

Huckebein

High Mobility Group Protein I (Y)
Human TATA Binding Protein
Interferon - Beta

Interferon Regulatory Factor
Janus Kinase

Kilobase

Kriippel

Knirps

Nicotineamide Adenine Dinucleotide
Neurectoderm Enhancer Element
Nuclear Factor Kappa B
Polymerase Chain Reaction
Position Weight Matrix

Receptor Tyrosine Kinase

Serial Analysis of Gene Expression
Sloppy — paired 1

Snail

Suppressor of Hairless

Torso

Tramtrack

Twist

Upstream Activating Sequence
Ventral Repression Region

Yolk Protein 1, 2

ix

Chapter I

INTRODUCTION

Understanding the basis of the complexity and diverse morphologies of
multicellular organisms remains a formidable challenge. According to the basic dogma of
molecular biology, DNA is the ultimate repository of biological complexity. A clear
revelation of the post-genome era is that organismal complexity does not simply correlate
with gene number (Adams et al., 2000). Morphological complexity is clearly not the
product of single genes. Rather, regulatory information encoded in the genome contains
the key to the differences responsible for morphological complexity and diversity. An
increasingly enormous bank of experimental data now confirms the a priori assumption
that many hundreds, and often thousands, of genes must be differentially expressed as a
function of time and space, in order to create any given tissue, body part, or multicellular
structure. The carefully choreographed progression of temporal- and domain-specific
gene expression is controlled by cis-regulatory elements, which constitute a fraction of
that part of an organism’s genome that does not encode proteins. Cis-regulatory elements
are DNA sequences in the vicinity of each gene that contain sequence specific motifs
bound by regulatory proteins or transcription factors that affect the expression of that
gene. By binding multiple distinct regulatory proteins, the presence of which individually
may depend on signaling events, cell cycle activity, temporal state, lineage, or spatial
position, cis-regulatory elements integrate temporal and positional information to direct
complex patterns of gene expression during development. Thus, the architecture of the

cis-regulatory apparatus constitutes a discretely organized DNA map, which represents in

physical terms the different phases of gene expression that are to be installed throughout
the life cycle, for every gene. Genomic changes that alter cis-regulatory element
architecture have the power to create new developmental processes and thus different
morphological outcomes (Belting et al., 1998; Shashikant et a1., 1998). Thus,
transcriptional regulatory regions themselves appear to be a driving force behind
evolutionary changes that underlie morphological diversity. It follows that analysis of
genomic cis-regulatory elements in terms of their structure and functional organization
holds the key to understanding how genomes encode the properties of organisms. Despite
the importance of the cis-regulatory apparatus in gene regulation, our ability to identify

and uncover the meaning implicit in its DNA sequence is extremely limited.

In this work, I describe studies that shed light on how transcriptional regulatory
information encoded within the cis-regulatory enhancer sequence is interpreted by the
cellular machinery; what (if any) are the rules that govern the structural organization of
cis-acting enhancer elements; how these architectural rules influence the biological
output of the elements; and how cis-regulatory ‘grammar’ provides insights into the

dynamics of developmental evolution.

THE REGULATORY APPARATUS ENCODED IN THE DNA

There are several classes of cis-regulatory DNAs that are involved in the control
of gene expression. This includes (a) the basal promoter near the transcriptional initiation
site. There are at least three different sequences (TATA box, initiator element [Inr], and

the downstream promoter element [DPE]) that might be found in the basal promoter that

serve as binding sites for the basal transcriptional machinery and RNA polymerase, and
are involved in the initiation of transcription (Amosti, 2002; Amosti, 2003); (b) Enhancer
elements that contain binding sites for sequence specific transcriptional activators and
repressors, and regulate levels of gene activity in a distance- and orientation-independent
manner (Banerji et al., 1981; Blackwood and Kadonaga, 1998); (c) Silencer elements that
suppress gene expression over long distances mainly through the creation of higher order
chromatin structure (Ogbourne and Antalis, 1998); and ((1) Boundary elements or
insulators, which prevent enhancers/silencers associated with one gene from
inappropriately regulating neighboring genes (Amosti, 2002; Amosti, 2003). These
regulatory elements, enhancers, silencers, and insulators are scattered over distances of
up to 100 kbp in metazoans. This elaborate organization of the regulatory DNAs permits
the detailed control of gene expression. A defining feature of gene regulation in higher
eukaryotes is the use of multiple enhancers, silencers, and promoters to control the

activities of a protein-coding unit.

ENHANCERS

Enhancers represent the most thoroughly analyzed type of cis-regulatory DNA
controlling gene expression. Enhancers were initially identified and characterized for
promoters of mammalian viruses using cell culture assays (Banerji et al., 1981). They
were shown to be DNA sequences that increased the expression of a linked gene in an
orientation- and distance-independent manner (Banerji et al., 1981; Blackwood and

Kadonaga, 1998). The characterization of enhancers in transgenic worms, flies, sea

urchins, ascidians, fish, frogs, chicks, and mice suggests that a typical enhancer is a
discrete element of less than 1 kbp and is composed of multiple binding sites for different
regulatory proteins, both sequence specific transcriptional activators and repressors

(Davidson, 2001).
Stochastic versus Rheostat models for enhancer function

The analysis of transcriptional activation by enhancers on a single-cell basis

indicates that the functional result of enhancer action can be to increase the probability

that a gene will be activated in any particular cell, without influencing the rate of
transcriptional initiation (Walters et al., 1996; Weintraub, 1988). This on or off
(Stochastic) response is in contrast to the Rheostat model in which enhancers increase the
rate 0 f transcriptional initiation (Magis et al., 1996). Both of these effects are observed
With reporter genes in the Drosophila embryo (Figure H). A weak activator drives IacZ
1' ePOrter gene expression in a stochastic manner in which some nuclei fail to show any
expression at all. In contrast, a strong activator drives expression in a more uniform

pattern, with overall higher levels of expression.

T
h 9 Nature of Enhancer activity

Enhancers have been suggested to function mainly through two distinct pathways:

I
Q 11gb remodeling of chromatin (Blackwood and Kadonaga, 1998), thereby facilitating
0:- -
‘1 hhibiting the binding of transcriptional machinery, and through direct interactions
“Vi

1‘ the general transcriptional machinery (Gaudreau et al., 1997). At least two factors

Figure I-l: Stochastic versus rheostat function of enhancer elements.

(A) In the on or off, stochastic model, genes are either in the "on" state or the "oil" state.
Transcriptional enhancers act to increase the probability that their cognate genes will be
transcribed, but do not affect the levels of transcription. The fraction of cells in which the
gene is activated may reflect enhancer strength, which is a function of the type and
number of its associated transcription factors. In the rheostat model, genes are uniformly
activated by enhancers, and the amount of transcription is proportional to the strength of
the enhancer (figure adapted from Blackwood and Kadonaga, 1998).

(B, C) A lacZ reporter gene containing five high affinity Gal4 binding sites is activated
by Gal4-Sp1 (B) or by the activation domain of the Gal4 activator (C). In the case of the
Gal4-Sp1 activator the expression of the reporter gene is switched on in some but not all
nuclei containing the activator giving a punctate pattern indicating a stochastic on or off
effect. In case of the Gal4-AD activator, the pattern of expression is not only more
uniform but is also more intense indicating a rheostat type of effect. LacZ expression is
visualized by in situ hybridization using digoxigenin labeled antisense lacZ mRNA
probes. Embryos are oriented anterior to left and dorsal up; lateral (B) and ventrolateral
view(C).

Images in this dissertation are presented in color.

Figure I-l: Stochastic versus rheostat function of enhancer elements.

A

Stochastic "on or off" model

 

 

 

 

 

 

 

 

 

 

 

 

 

Increasing Enhancer activity )

 

 

 

B
Rheostat model
00000 0.... 0....
OO .0 0.
800800 :::: oo ::::oo

oo
00 0000 '0 a... ‘0 0.0.
000000 000000 0000..

 

 

 

 

 

 

 

 

 

Increasing Enhancer activity

\
VG.

gar-:33, IacZ

Ii-

Gal4 -60

contribute to the activity of enhancer elements: First, the sequence specific positively and
negatively acting proteins that bind directly to sequences within the element and second,
the cofactors (coactivators, corepressors, and chromatin modifying complexes) recruited

by the enhancer bound factors.

Identification of enhancers was achieved originally through genetic approaches
such as the classic mutations that altered the expression of the HOX genes in the bithorax
complex (Lewis, 1998). The focus on the identification of regulatory elements for

individual genes has included several other experimental approaches: the generation of
deletion constructs to determine the minimal sequences necessary for transcription in
ceil-culture-based systems; DNase I hypersensitivity studies to identify sequences
potentially available for transcription factor binding; and in vitro approaches, such as
DNA footprinting and gel shifts, to determine sequences that bind various regulatory
proteins. Screens to identify cis-regulatory elements have also been carried out in
transgenic mice, albeit in an extremely laborious and low-throughput manner. In addition,

a limited number of large-scale promoter and enhancer trapping studies have been done
(Asoh et al., 1994; Durick et al., 1999; Fukushige and Ikeda, 1996). Most of these gene
regulatory studies have consisted of largely unguided searches of genomic sequence for
those with gene regulatory properties. Much of our knowledge of cis-regulatory enhancer
eléments that function in the context of developmental specification at present derives
fig. 311 DrOSOphila, where sophisticated tools for characterizing the features of enhancer

el
§ Inents are available.

THE FUNCTION OF CIS-REGULATORY ENHANCER ELEMENTS DERIVES

FROM ITS STRUCTURAL ORGANIZATION
Modularity

A general feature of cis-regulatory architecture in higher eukaryotes is that the
diverse phases of expression of developmental genes are frequently mediated by multiple
cis-regulatory elements, referred to as regulatory ‘modules’ present in the DNA flanking
the gene or in its introns. Cis—regulatory modules are discrete DNA sequences of a few

hundred base pairs (100-1000 bp) that contain multiple binding sites for multiple distinct
transcription factors. Each module drives a discrete portion of the overall expression
profile of the linked gene. Modules can be moved from their native context and still
recapitulate a portion of the native expression pattern, thus acting as autonomous units
(Davidson, 2001). Thus, a gene may have many modules in its cis-regulatory region.
Modularity allows each element to fimction independently of the other and therefore,
each module signals the basal promoter independently. At a given moment, in a given
nucleus, one module in a complex regulatory region may be inactive or silent, while at
the same time another adjacent module may be actively firing the promoter. In this study
(Chapter 2), I demonstrated that a similar kind of autonomy occurs within the tight
confines of a single module. Modular regulatory organization provides the organism with

t . . . .
he servrces of a given gene in multiple developmental contexts.

The modularity of cis-regulatory regions is puzzling, because modular structure
e
at:- be all‘gued to be less optimal (precise) than nonmodular (elements that are

int .
§rconnected and not autonomous) structures. Modules convey an advantage in

situations where the developmental specifications change from time to time, as in the
evolution of diverse species. New modules can be easily constructed from existing, well-
tested modules and can be readily configured to adapt to new conditions. Changes in
regulatory sequences within individual elements or modules may subtly affect the level,
timing, or spatial pattern of gene expression, and may do so very selectively in terms of
the tissues and stages of development involved. These quantitative, temporal, and spatial
changes in the deployment of regulatory genes may affect the level, timing, and spatial
expression of other developmental genes. A nonmodular element, in which every
component is optimally linked to every other component, is effectively frozen and cannot
evolve to meet new optimization conditions. Thus, modularity in cis-regulatory systems
is not only essential to their developmental function, but also facilitates the evolution of
developmental processes because individual elements can evolve independently (Carroll,

3. B, 2001).

One of the most extensively studied modular cis-regulatory systems is that
controlling the expression of the pair-rule gene even-skipped] during the segmentation
process early during the embryogenesis of the fruitfly Drosophila melanogaster.

Anterior—posterior patterning in Drosophila is initiated by a cascade of transcription
factors, which culminates in the establishment of segment boundaries (Figure I-2).
PO sitional information along the A-P axis of the syncitial blastoderm is encoded in a
suchssion of different ways during development. A major determinant of anterior

p . . - . . . . . .
o§rtronal Information is the roughly exponential gradient of Brcord protein imposed by

M

ital 1 ~ 2“? the “Se of new Drosophila genetic nomenclature throughout this document: all gene names are
leu§crzed and Written in all lower case letters. Protein names are in regular font and start with an upper case
I.

Figure I-2: Anterior-Posterior patterning in Drosophila is initiated by a cascade of
transcription factors.

Initially, maternally derived morphogenetic gradients of Bicoid, Hunchback and Caudal
transcription factors provide gene regulatory inputs to activate the expression of the
zygotic gap genes giant, knirps, krz'ippel, hunchback and tailless. Each of these zygotic
genes is expressed in one or two broad domains and together with the maternally derived
transcription factors regulate the expression of the pair-rule genes including even-
skipped.The pair-rule genes together with the gap genes then further regulate the
expression of the segment polarity genes that define the borders of the future segmental
compartments of the adult body plan.

10

Figure I-2: Anterior-Posterior patterning in Drosophila is initiated by a cascade of
transcription factors.

hmchback [l‘b]

,even-sldpped (eve
8r fusH inrazu [ftz]

engrdled [en]

 

11

the mother fly, along with anterior maternal hunchback expression. Posterior positional
information is provided by a similar gradient of maternal Caudal protein. These maternal
gradients provide gene regulation inputs that directly or indirectly regulate the expression

of the gap class of zygotic segmentation genes (kriippel, knirps, giant, hunchback and
tailless), each of which is expressed in a simple pattern made up of one or two broad
expression domains. Together with the maternal morphogens themselves, the gap genes
act as more localized gradients to establish the A-P body plan. They directly or indirectly
control the expression of the pair-rule class of segmentation genes, each of which is
expressed in a series of seven transverse stripes that lie along the A-P axis and precisely
define the segmental compartments of the embryo, which in turn correspond to the future
compartments of the body of the adult organism. The primary pair-rule genes, including
even-skipped, encode transcription factors. The location of each stripe of expression

depends on the cis-regulatory system controlling these genes (Rivera-Pomar and Jackle,

l 996)-

eve encodes a homeodomain protein that is expressed in a series of seven

transverse stripes (each of which is 4-5 nuclei in width) along the length of the embryo
and plays a key role in the establishment of the metameric body plan (Frasch et al., 1987;

F 1‘ Busch and Levine, 1987; Harding et al., 1989; Harding et al., 1986; Macdonald et al.,

1 9 86; Macdonald and Struhl, 1986). The transcriptional regulation of eve in a seven-
Stli P3 pattern is complex, in that the eve cis-regulatory region is modular and contains
Sebarate stripe enhancer elements or modules that control the expression of individual

8 \
t5 3365 or pairs of stripes (Goto et al., 1989; Harding et al., 1989; Small et al., 1992).

12

Subsequent comprehensive analysis of the eve locus identified within a 16 kbp locus five
separable elements 5’ and 3’ of the transcription unit, that together create the seven stripe
pattern of eve (Fujioka et al., 1999) (Figure I-3). Stripe 2, stripe 3+7, stripe 4+6, stripe 1
and stripe 5 enhancer elements drive expression of eve in the corresponding stripes and
contain all necessary transcriptional information for their correct spatial distribution in
the embryo. Each stripe enhancer element or module contains target sites for positively
acting transcription factors which cause it to be expressed; and for negatively acting
transcription factors which set the anterior and posterior boundaries within which
expression is allowed. Thus, the general strategy for pair-rule gene expression is the
initial widespread activation under the influence of generally distributed activators,
followed shortly afterwards by repression in the interstripe regions (Frasch et al., 1987;

Frasch and Levine, 1987).

Short-range transcriptional repression

A combination of genetic analyses (F rasch et al., 1987; Frasch and Levine, 1987),

DNA binding experiments (Stanojevic et al., 1989), transient cotransfection assays
(Small et al., 1991; Small and Levine, 1991), and expression assays in transgenic
Drosophila embryos (Small et al., 1992; Stanojevic et al., 1991) have provided evidence
for the following model of eve stripe 2 regulation. The broad Bicoid gradient emanating
fiol‘n anterior regions of the precellular embryo induces a steeper pattern of hunchback
exIDI-ession. Bicoid and Hunchback function synergistically to activate the eve stripe 2

enllancer in the anterior half of the embryo. The stripe borders are formed by two

13

 

F. . 4 1E?

 

Figure I-3: Cis-regulatory region of the pair-rule gene even-skipped.

even-skipped is a pair-rule gene encoding a homeodomain transcription factor that is
expressed in seven transverse stripes across the anterior-posterior axis of the blastoderrn
embryo. eve expression is visualized in the embryo above, by in situ hybridization using
digoxigenin labeled antisense eve mRNA probe. Comprehensive analyses of the eve locus
have identified within a 16 kb region, separable regulatory modules called stripe
enhancer elements 5’ and 3’ of the transcription unit. Each of the stripe enhancer
elements (stripe 3 + 7, stripe 2, stripe 4 + 6, stripe 1, and stripe 5) drives a discrete
portion of the overall expression profile of the gene. Figure adapted from Fujioka et al.,
1999; Sackerson et al., 1999.

14

Figure I-3: Cis-regulatory region of the pair-rule gene even-skipped.

even-skipped

    

Transcription unit

3»

stripe 5
stripe 3 4 7 strips 2 ’ strips 4 + 6 stripe 1

Fujioka at al. (1909) Dnvaioprnont 12:2527 and Sackerson at at. (1299) Dev. ”01:11:30

15

repressor gradients; an anterior Giant gradient (Capovilla et al., 1992; Kraut and Levine,
1991a; Kraut and Levine, 1991b) and a posterior Krfippel gradient (Gaul et al., 1987).
The stripe 2 enhancer (Figure L4) is an approximately 500 bp sequence of DNA located
between -1.6 kb and —l.1 kb upstream of the eve transcription start site and contains a
total of eighteen binding sites, including eight activator sites (five for Bicoid and three for
Hunchback) and ten repressor binding sites (three for Giant, six for Kriippel and one for
sloppy-paired) (Andrioli et al., 2002; Berman et al., 2002; Ludwig et al., 2000; Ludwig
and Kreitman, 1995; Ludwig et al., 1998; Small et al., 1992). Mutations in these factor
binding sites alter the normal expression pattern of a stripe 2-lacZ fusion gene in
transgenic embryos; the abnormal patterns ofien mimic those seen when the wildtype
stripe 2 enhancer is expressed in segmentation mutants (Small et al., 1992; Stanojevic et

al., 1 991).

The model for eve 2 regulation is incomplete in at least two respects. First, the
combined activities of Bicoid and Hunchback do not seem to be sufficient for activation
at the position of eve stripe 2. For example, reporter genes containing three high affinity
Bicoid sites and three Hunchback sites cannot respond to the low levels of these proteins

at the position of eve stripe 2 (Simpson-Brose et al., 1994). In addition, several attempts
to construct artificial stripe 2 enhancers using up to ten Bicoid and/or Hunchback sites
have been unsuccessful (S. Small and M. Levine, unpublished). These results suggest that
Cis~regulatory sequences other than the Bicoid and Hunchback sites are important for

activation. Such sequences may contain low-affinity Bicoid- and Hunchback-binding

l6

 

Figure I-4: Model for the regulation of eve stripe 2.

The minimal eve stripe 2 enhancer element is ~480 bp long and contains binding sites for
the maternally derived transcriptional activators Bicoid and Hunchback, and for
transcriptional repressors encoded by the zygotic gap genes giant and kriippel. The
maternal Bicoid and Hunchback proteins form roughly exponential anterior to posterior
morphogenetic gradients, activating eve stripe 2 expression in the anterior regions of the
embryo. The stripe 2 borders are established by the short-range transcriptional repressors
Giant and Kriippel. Giant expressed in a broad band in the anterior regions of the embryo,
sets the anterior border, while Kriippel expressed in the central region of the embryo sets
up the posterior border. The expression of eve is visualized in the embryo above, by in
situ hybridization using digoxigenin labeled antisense eve mRNA probe. Figure is
adapted from D. Papatsenko (http://homepages.nyu.edu/~dap).

l7

Figure I-4: Model for the regulation of eve stripe 2.

 

anterior posterior

 

 

sites, or sites for other activator proteins. Alternatively, they may simply function to
provide the correct spacing between known activator and repressor sites. Second, the
mechanism(s) that control repression of eve 2 in anterior regions are not well understood.
As the activators (Bicoid and Hunchback) are distributed throughout the anterior half of
the embryo, repressive mechanisms other than that provided by Giant must exist that
prevent activation in all nuclei anterior to the position of the stripe. Andrioli et al., show
that the forkhead domain protein Sloppy-paired (Slpl), which is expressed in a broad
anterior domain, binds to a site in the stripe 2 element and is sufficient for repression of
stripe 2 of the endogenous eve gene. Further genetic experiments identify a separate
repression activity near the anterior pole that is dependent on the terminal patterning gene
torso. Thus, three position-specific activities arerequired for anterior repression of eve 2

(Andrioli et al., 2002).

Similarly a separate enhancer module contained within the complex eve cis-
regulatory region directs the expression of stripes 3 and 7. This enhancer is
approximately 500 bp in length and maps ~ 3.3 kb upstream of the transcription start site
(Frasch et al., 1987; Frasch and Levine, 1987; Goto et al., 1989; Harding et al., 1989;
Small et al., 1993). The stripe 3 + 7 enhancer is regulated by one or more ubiquitously
diStributed activators, including components of a JAK-Stat pathway (Yan et al., 1996),

which can switch on the gene along the entire length of the early embryo. The two-stripe
p alitern is defined by multiple tiers of repression mediated by the gap proteins Knirps and
Hunchback, which delimit the ubiquitous activation. The stripe 3 + 7 enhancer has at
least five binding sites for Knirps and eleven binding sites for the Hunchback repressor.

The zinc finger repressor Hunchback is responsible for establishing the anterior border of

19

stripe 3 and the posterior border of stripe 7; Knirps, a member of the nuclear receptor
family of transcription factors, establishes the posterior border of stripe 3 and the anterior

border of stripe 7 (Small et al., 1996).

Thus, transcriptional repressors are critical in establishing localized patterns of
gene expression during Drosophila embryogenesis that is essential for the development
of this multicellular organism. A striking feature of the stripe enhancer elements is that
their action is autonomous, such that the repression of one element does not lead to the
general repression of the entire locus. For example, repression of the eve stripe 2
enhancer in the central regions of the embryo by Kriippel does not prevent the more distal
eve stripe 3 element from activating the promoter. Key to the functional autonomy of the
stripe enhancer modules is the short-range of the transcriptional repressors that bind and
regulate them. The short—range transcriptional repressors that are critical in the regulation
of the even—skipped pair-rule gene are the products of the gap genes, giant, knirps and
krr'ippel. The short-range of the repression activity plays an important role in the

regulation of complex modular promoters (Figure I-5), by preventing inappropriate
‘cross-talk’ between enhancers and is essential in specifying the sharp borders of even-
skipped expression. The stripe 2 and stripe 3 + 7 enhancers in the eve locus are separated
by a 1.7 kb spacer sequence. The removal of this spacing caused repression signals from
one enhancer to interfere with the activity of the adjacent enhancer to generate abnormal
patterns 0f expression in the early embryo (Small et al., 1993). Transgenic embryo assays
caJ‘rying endogenous enhancers linked to the lacZ reporter gene have been used to study
the range of this class of repressors. The zygotically active repressors Giant, Knirps and

1<riippfil were shown to repress either activators within enhancers or basal promoter

20

 

Figure 1-5: Transcriptional repression in Drosophila.

Transcriptional repressors in Drosophila are classified as Short-range repressors or Long-
range repressors. The zygotic gap genes giant, knirps, kriippel and the mesoderrn
determinant snail encode the short-range repressors that are active during early
embryogenesis. These proteins are capable of repressing the activity of enhancer
elements when located within 100 bp of key activator sites, or of basal promoter elements
when cognate sites are introduced close to the start site of transcription. The short-range
of their activity allows multiple enhancers in a complex promoter to function
autonomously such that the repression of one element does not lead to the general
repression of the entire locus. In contrast long-range repressors, such as Hairy, can
mediate repression over distances of > 1000 bp and thus function in a dominant manner
to silence multiple enhancers in a complex promoter. Short-range repressors represent a
flexible form of gene regulation as they show both enhancer-specific and promoter-
specific effects depending on the location of their binding sites.

21

Figure I-5: Transcriptional repression in Drosophila

Short-range repressors Long-range repressors
giant h“ii'Y
knirps DOPSG'
repression
kruppel » complexes
snail

£25 @ mfl

5‘ ii 3 g I IZ’ $43 $4! [j'

gloobp >1000bp

 

22

elements when bound within ~100 bp of their apparent targets (Amosti et al., 1996a;
Amosti et al., 1996b; Gray and Levine, 1996; Gray et al., 1994; Hewitt et al., 1999;
Keller et al., 2000). In contrast to the short-range repressors, long-range repressors such
as Hairy, which are also active in the early embryo, can block enhancer elements over
distances of more than 1 kb away (Figure I-5), leading to the dominant repression of
multiple enhancer complexes (Barolo and Levine, 1997). Thus, short-range repressors
clearly provide a more precise, tunable form of repression that can be used to direct

complex patterns of gene expression during development.

The short-range transcriptional repressors, Giant, Knirps and Kriippel, are the
products of the gap genes that are among the first zygotic genes to be expressed during
the early development of Drosophila melanogaster (Figure I-6). Loss of gap gene
function leads to missing segments, or gaps, in the embryo body plan. The gap gene
product Giant has been characterized as a repressor of other gap genes, including kriippel
and knirps, as well as pair-rule genes such as eve (Eldon and Pirrotta, 1991; Kraut and
Levine, 1991a; Kraut and Levine, 1991b; Small et al., 1992). The Giant protein contains
a C-terminal dimerization/DNA-binding domain of the basis/leucine zipper class
(Capovilla et al., 1992; Vinson et al., 1989). giant mutants exhibit abdominal segment
defects and loss of head structures, a pattern consistent with the regions in which it is
expressed (Gergen and Wieschaus, 1986; Petschek et al., 1987). The Drosophila Knirps
protein is a member of the nuclear receptor family of transcription factors and is
expressed in the abdominal regions of the precellular embryos and anterior regions of the
presumptive head (Rothe et al., 1989). The Knirps protein plays an essential role in the

segmentation process, both by refining the expression patterns of other gap genes and by

23

Figure I-6: Expression pattern of the short-range repressors Giant and Knirps.

(A) giant is expressed in two broad domains in the anterior and posterior regions of the
early blastoderm embryo.

(B) knirps is also expressed in two broad domains in the head region and in the
presumptive abdomen in the early blastoderm embryo.

Expression of giant and knirps in the embryos above are visualized by in situ
hybridization using digoxigenin labeled antisense mRNA probes. Embryos are oriented
anterior to left; dorsal up.

24

Figure I-6: Expression pattern of the short-range repressors Giant and Knirps.

 

25

establishing pair-rule stripes of gene expression (Nusslein-Volhard et al., 1987; Pankratz
et al., 1989; Small et al., 1996). Mutations in the knirps gene are embryonic lethal,
showing a characteristic gap phenotype of the larval cuticle. Knirps also plays important
roles in tracheal and wing formation later in development. kriippel expression is regulated
by the maternal Bicoid gradient (Hoch et al., 1992) and is restricted to the central regions
of the embryo corresponding to the presumptive thorax and anterior abdomen (Gaul et
al., 1987) and abutting the domains of Giant and Knirps repressor proteins. Kriippel is a
zinc finger transcriptional repressor (Licht et al., 1993; Zuo et al., 1991) that regulates the

expression of the pair-rule gene even-skipped.

Context-dependent multiple repression activities of short-range repressors

A common property of short-range transcriptional repressors is their interaction
with the evolutionarily conserved corepressor CtBP (C-terrninal Binding Protein). This
transcription factor was originally identified in human cells through its interaction with
the adenovirus ElA oncoprotein (Chinnadurai, 2002; Schaeper et al., 1995; Turner and
Crossley, 2001). CtBP proteins are similar in sequence to NAD-dependent D-
hydroxyacid dehydrogenases, and possess very similar overall structures to these
enzymes (Kumar et al., 2002; Nardini et al., 2003). Recent studies have shown the CtBP
has a weak dehydrogenase activity in vitro, although the physiological substrates of CtBP
as well as the significance of this enzymatic activity in transcriptional repression remain
unknown (Balasubramanian et al., 2003; Kumar et al., 2002). CtBP is recruited to

promoters through interactions with a short PXDLS peptide motifs found in short-range

26

transcriptional repressors (and in other interacting proteins) where it mediates repression
through mechanisms not currently understood. CtBP has been shown to interact with
chromatin modifying factors, including histone deacetylases and histone
methyltransferases (Shi et al., 2003; Subramanian and Chinnadurai, 2003; Sundqvist et
al., 1998). Thus, the major activity of CtBP might be to serve as a bridging molecule,
recruiting chromatin or transcription factor modifying enzymes to the promoter. CtBP
itself might possess an alternative enzymatic activity similar to the NAD-dependent Sir 2
repressor protein, which requires NAD to mediate deacetylation of histone proteins

(Marrnorstein, 2002).

Although CtBP-mediated repression is critical for full activity of short-range
repressors, Drosophila short-range repressors also possess CtBP-independent repression
activities (Keller et al., 2000; La Rosee et al., 1997; La Rosee-Borggreve et al., 1999;
Nibu et al., 2003; Strunk et al., 2001). A part of the CtBP-independent activity of these
repressors might be attributed to competition for activator binding sites but they also
possess CtBP-independent repression activities that does not require direct competition
(Keller et al., 2000; La Rosee-Borggreve et al., 1999; Nibu et al., 2003; Strunk et al.,
2001). Studies on Giant repression activity in our lab demonstrated that the dCtBP
cofactor is required for Giant repression of some, but not all target genes (Strunk et al.,
2001). The results indicate that Giant can repress via both dCtBP-dependent and —
independent pathways and dCtBP requirement can vary on a gene-to-gene as well as on

an enhancer-to-enhancer basis.

Previous work in our lab has identified two repression regions of the Knirps

Protein (Keller et al., 2000). The C-terminal region from amino acids 202-358, appears to

27

mediate repression through dCtBP and contains the dCtBP-binding motif PMDLSMK.
The N-terminal region, from amino acids 139-330, does not bind dCtBP, and was shown
to repress the eve stripe 2-lacZ in a dCtBP mutant background (Keller et al., 2000). Also,
the loss of maternal dCtBP does not affect Knirps repression of eve stripe 3 + 7 enhancer,
but abolishes repression of eve stripe 4+6 enhancer element (Keller et al., 2000). La
Rosée et al. have shown that the other short-range repressor Kriippel can repress the hairy

stripe 7 enhancer in the absence of dCtBP (La Rosee et al., 1997).

Thus, a growing body of evidence demonstrates that in addition to the short-range
of activity, many, or perhaps all, short-range repressor proteins exhibit multiple
repression activities. Multiple repression activities may allow for quantitative or
qualitative effects on gene expression and may be context-dependent. Thus, qualitatively,
a repressor may function selectively in a tissue-specific manner or in an activator-specific
manner (Postigo and Dean, 1999) or in different promoter contexts (Lunyak et al., 2002).
Quantitatively, dual activities may increase the overall level of repression. Consistent
with a quantitative model, the CtBP-dependent and CtBP-independent repression
activities of Knirps have been found to exhibit striking functional similarities in cell
culture assays, indicating that they might utilize similar mechanisms of repression (Ryu
and Amosti, 2003). Additionally, we have shown that increasing the dose of the Knirps
repressor may be sufficient to overcome the requirement for the dCtBP cofactor and that
multiple repression activities within a single protein represent quantitative effects on gene
expression (Paolo Struffi, Maria Corado, Meghana M. Kulkami and David N. Amosti.
“Quantitative Contributions of CtBP-dependent and —independent repression activities of

Knirps”. Manuscript in revision).

28

Other studies have indicated that different concentrations of Giant, Knirps and
Krilppel are important for the proper regulation of zygotic genes (Kosman and Small,
1997; Wu et al., 1998). An example is the regulation of eve stripes 3, 4, 6 and 7 (Figure I-
7). As previously described, transcriptional repressors Knirps and Hunchback delimit
expression borders of the even-skipped stripes 3,4, 6 and 7. Two corresponding cis-
regulatory enhancer modules, the stripe 4+6 and stripe 3+7 enhancers, encode all
sufficient transcriptional information for these four stripes. The eve stripes 4 and 6 are
formed in the embryo zones with lower concentration of Hunchback and higher
concentration of Knirps, conversely the eve stripes 3 and 7 are formed where Hunchback
concentration is greater than that of Knirps. Sequence analysis of the eve gene indicates
that there are more high affinity Knirps binding sites within the eve stripe 3 + 7 enhancer
than in the 4 + 6 element (Berman et al., 2002; Papatsenko et al., 2002). This is also
consistent with our observations that the eve 3 + 7 enhancer is more sensitive to Knirps

repression than the eve 4 + 6 element (Struffi et al., manuscript submitted).

These observations suggest that in addition to spacing constraints, the relative
affinity and the number of binding sites in the regulatory region are critical parameters in
dictating short-range repression effectiveness. To analyze enhancer function in a setting
in which activator-repressor stoichiometry and spacing can be exactly defined, I
constructed chromosomally integrated, compact synthetic enhancer modules containing
binding sites for endogenous short-range repressors Giant, Knirps, or Kriippel and
chimeric Gal4 activators (Chapter 3). The construction and functional assessment of
different synthetic elements allowed us to accurately quantify not only the exact distance

requirements, but also parameters such as relative affinities, number and arrangements of

29

Figure I-7: Model for the regulation of the eve stripes 3 + 7 and 4 + 6.

The expression of eve is visualized in the embryo above (A), by in situ hybridization
using digoxigenin labeled antisense eve mRNA probe. (B) Two corresponding regulatory
elements the 3 + 7 and 4 + 6 enhancers encode all the necessary transcriptional
information for the precise expression of even-skipped in stripes 3, 7, 4 and 6. The
transcriptional repressors Hunchback and Knirps, for which the enhancer elements
contain binding sites, establish the stripe borders. The eve stripes 4 and 6 are formed in
embryo zones with lower concentration of Hunchback and higher concentration of
Knirps. Conversely, the eve stripes 3 and 7 are formed where Hunchback concentration is
greater than that of Knirps. Figure is adapted from D. Papatsenko
(http://homepages.nyu.edu/~dap).

30

Figure 1-7 : Model for the regulation of the eve stripes 3 + 7 and 4 + 6.

A
iiij‘ rm
i" are
34 67
B
Hb -i -i l- I-Hb
3i- 4l-Kni-16 i7

 

 

31

activator and repressor binding sites, features that could not be dissected in previous

studies of short-range transcriptional repression.
CIS—REGULATORY INFORMATION PROCESSING

The qualitative functional complexity of cis-regulatory enhancer elements is a
manifestation of the complexity of the assemblage of target sites within the module.
Thus, enhancers serve as binding platforms for multiple diverse regulatory proteins, both
positively and negatively acting factors. In addition, each element may contain multiple
target sites for the same or different transregulator thereby serving to increase the local
concentration of regulatory proteins that affect the expression of the linked gene.
Furthermore, the presence of each factor may additionally depend on signaling events,
cell-cycle activity, temporal state, lineage or spatial position. Thus, enhancers can
integrate environmental and developmental information to regulate the expression of
genes in a biologically appropriate manner. Depending on the signals or cell type, the
same enhancer element might activate or repress, and the magnitude of activation signals
can be variable, in the manner of a rheostat (Amosti, 2002; Amosti, 2003; Barolo and
Posakony, 2002; Biggar and Crabtree, 2001; Rossi et al., 2000). The combinations and
arrangement of binding sites within the enhancer dictate the nature of the signal
generated. This property of enhancer elements has led to the analogy of an enhancer as a
molecular logic device or computer. Computational functions of an enhancer have been
ascribed to its ability to receive multiple inputs in terms of the different transcription
factors that bind it, to process these inputs resolving them into a single output, which is

directed to the basal machinery, either turning the gene on or off.

32

The analysis of gene regulation in sea urchins by Eric Davidson and colleagues is
the best known example of the “enhancer as a computer” analogy and showed the
possibility of modeling gene regulation as logic circuits (Yuh et al., 1998). The same
study also demonstrated that the endo 16 cis-regulatory region of the purple sea urchin
Strongylocentrotus purpuratus works remarkably like a tiny analog computer. Like other
cis-regulatory systems that mediate complex developmental patterns of expression, the
endo 16 system is modular in organization, that is, it consists of sub-elements of DNA
sequence, each of which can execute a certain regulatory function such as activation,
repression, synergism, and integration. Separate portions of the regulatory region can be
combined to recreate some or all, of the expression pattern of the gene, a characteristic of
modular elements. To test their understanding of the endo 16 cis-regulatory region, Yuh
and colleagues created a computer model with Boolean elements that simulates these
regulatory interactions. The model made predictions about the consequences of specific
promoter manipulations on transcription levels and when tested experimentally
successfully simulate the output of the regulatory regions. The success of the model
emphasizes the integrative, computer-like processing suggested to be a characteristic of

enhancer elements.

Studies such as those described above suggest that the enhancer functions as the
central information processing unit while the basal transcriptional machinery itself simply
responds to the signals generated by these molecular logic circuits. Such integrative
functions have been ascribed to the human interferon-B (IFN—B) enhancer, which drives

transcription of the IFN-B gene in response to viral infection (Struhl, 2001). The IFN-B

33

enhancer is a small 65 bp region immediately upstream of the core promoter. It contains
binding sites for activators of the NFKB, IRF, and ATF/Jun families as well as target sites
for the architectural protein HMGI (Y) (Kim and Maniatis, 1997; Munshi et al., 2001;
Thanos and Maniatis, 1995). These regulatory proteins assemble through cooperative
interactions into a well—defined nucleoprotein complex called the ‘enhanceosome’.
Assembly of the enhanceosome is absolutely essential for the transcription of the IFN-B
gene in response to viral infection in cells. Individual activators bound to their sites in the
enhancer do not by themselves stimulate transcription. This is because the nucleoprotein
complex provides a stereospecific interface for interaction with the basal transcriptional
machinery, possibly engaging several components of the basal machinery simultaneously
to effect synergistic activation (Carey et al., 1990; Chi et al., 1995). In this structured
element, the presence of each transcription factor binding site and its precise arrangement
within the regulatory element are critical for the various regulatory proteins (sequence-
specific activators and architectural proteins) to assemble through cooperative
interactions into a well-defined nucleoprotein complex called the “enhanceosome” and is
essential in dictating the output of the element (Carey et al., 1990; Chi et al., 1995; Kim
and Maniatis, 1997; Munshi etal., 2001; Thanos and Maniatis, 1995). The enhanceosome
therefore imposes the restriction that the target gene would be activated only when all the
regulatory proteins are present and are functionally active. The IFN-B enhanceosome is a
paradigm for precision and functions as a precise on/off binary transcriptional switch in

response to the appropriate stimulus (Struhl, 2001).

In contrast to the stringent organization of factor binding sites in the IFN-B

enhanceosome some flexibility in organization has been observed for the even-skipped

34

 

stripe 2 enhancer element (Amosti et al., 1996a; Ludwig et al., 2000). A defective stripe 2
enhancer lacking the crucial Bicoid Bl site can be complemented with a high affinity
Bicoid binding site inserted at a new location (Amosti et al., 1996a). However, the same
study also demonstrated that there are some constraints on enhancer organization since
the insertion of the same high affinity Bicoid sequence at another location results in only
a partial restoration of the stripe. In addition to flexibility in the organization of the cis-
element, the authors also observed that there was flexibility in the trans-regulation of the
stripe. Activation does not seem to depend on a particular class of DNA binding protein
(the Bicoid homeodomain can be replaced by the Gal4 zinc finger domain), nor does
expression require a particular type of activation domain (Bicoid activation domain can
be replaced either by the glutamine rich activation domain of Spl or the acidic activation
domain of GCN4). The results from this study suggest that there may be no specific
requirement for particular protein-protein interactions and that eve stripe 2 expression
only requires the binding of a sufficient number of activation domains (Amosti et al.,
1996a). Mutations in individual binding sites further demonstrated that each bound

transcription factor contributes to the overall output of the enhancer.

The functional analysis of the eve stripe 2 enhancer by Kreitman and colleagues
has also suggested a more flexible arrangement of regulatory proteins to be the
predominant pattern for cis-elements that provide diverse patterns in developing systems,
rather than the more constrained architecture afforded by the enhanceosome.

Phylogenetic comparisons among eve stripe 2 enhancer elements from disparate

Drosophilids indicate that this enhancer has undergone some genetic drifi in its internal

35

 

organization. However, the redesign in internal architecture does not affect the overall
output of the element (Ludwig et al., 2000; Ludwig and Kreitman, 1995; Ludwig et al.,
1998). The plasticity of this enhancer suggests that much variation in spatial placement
of individual transcription factors is possible, consistent with a model in which these
factors contact the basal machinery in a flexible framework, not necessarily as a rigid

complex.

We decided to test the functional significance of this flexibility in enhancer design
on the way in which the enhancer communicates with the basal promoter. Our studies
with composite enhancer elements (described in Chapter 2) containing binding sites for
the short-range repressors Giant or Knirps and different activators, led us to propose a
new model for enhancer function. We call this the’Infonnation display’ or ‘Billboard’
enhancer, where computation of regulatory inputs is not performed by the enhancer alone
but results from multiple iterative/simultaneous interactions between enhancer
subelements and the basal transcriptional machinery. Consistent with this model,
functional analysis of many cis-regulatory elements has demonstrated the presence of
redundant regulatory information which suggests that there are multiple configurations in
which enhancer-bound factors can interact with the basal transcriptional machinery (Han

etal., 1998; Hoch et al., 1992; Piano et al., 1999).

36

CIS-REGULATORY SEQUENCE AND THE EVOLUTION OF

MORPHOLOGICAL FEATURES

The morphological features of multicellular organisms are the products of
developmental processes, which are controlled by genomic regulatory programs. The
evolution of form has been suggested to largely reflect the evolution of these genomic
regulatory programs (Carroll, S. B, 2001). Given that many regulatory genes encode
highly conserved proteins, regulatory evolution is thought to be brought about primarily
by substitutions in cis-regulatory sequences rather than in the proteins themselves. The
creative potential of regulatory change and the comparatively greater constraints on
protein evolution have always been recognized. However, little is known about the tempo
and mode of cis-acting regulatory sequence evolution, in part a reflection of the technical

difficulty in dissecting regulatory sequence structure and function.

Evolutionary analysis of non-coding sequences is more difficult than that of
coding sequences given the paucity of structural information about cis-regulatory DNA
(Dermitzakis and Clark, 2002; Leung et al., 2000; Wasserman et al., 2000). There are no
straightforward properties in regulatory sequences analogous to the open reading frame
and codons in coding sequences, making it difficult to define the position, amount, and
strength of selective constraints on functional regulatory elements. In addition, the model
of transcriptional regulation is not a simple one of activation or suppression by
transcription factors, but rather, it includes competitive binding of proteins (Small et al.,
1991), cooperative binding (Burz et al., 1998; Zhao et al., 2000), DNA bending,
chromatin modifications and other molecular interactions that are not always reflected in

the nucleotide sequence.

37

Understanding the evolutionary processes that regulatory sequences undergo will
substantially improve our understanding of their functional constraints. Numerous
comparative methods (termed "phylogenetic footprinting") have been developed that use
sequence conservation to infer function (Hardison et al., 1997 ; Miller et al., 2001). It may
be problematic, however, to assume that all functional sequences are conserved and all
nonfunctional sequences have diverged. For instance, sequence comparisons of closely
related species produce many false positive results, and comparison of distantly related
species has low power to detect functional elements in the species compared (Dermitzakis
and Clark, 2002). Moreover, it has been shown that regulatory sequences can maintain
regulatory function despite structural reorganization as a result of species-specific loss
and gain of transcription factor binding sites (Cuadrado et al., 2001; Ludwig et al., 2000;
Piano et al., 1999). Given that transcription factor binding sites are the fundamental units
of regulatory structure, they are also likely to be the fundamental units of regulatory
evolution. Therefore, methods that can predict the binding site composition of a sequence
should provide a powerful means for quantifying regulatory sequence divergence in
comparative analyses. Application of predictive tools such as probability weight matrices
(PWMs) has been useful in the annotation of regulatory regions (Berman et al., 2002;
Stormo, 2000a; Stormo, 2000b; Stormo, 2000c), and such tools have been shown to be
useful for predicting the evolution of regulatory regions as well (Chuzhanova et al., 2000;
Liu et al., 2000; Ludwig et al., 2000). However, due to the low sequence specificity of
transcription factors for their binding sites, only a fiaction of the predicted binding sites

are thought to be functionally significant.

38

Thus, meaningful evolutionary analysis of a regulatory sequence will require
detailed information about the locations of transcription factor binding sites within a
sequence, the fimctional specificity of the binding sequences, and the spatial
requirements for their interaction. Therefore, if we knew in functional terms the
components of the specific genomic cis-elements that result in different morphological
outcomes in two animals of common ancestry, we could determine exactly what are the
essential causal differences in the DNA of these animals and provide a mechanistic

explanation of how the diverse forms actually arose during evolution.

TOWARDS DEVELOPING A ‘GRAMMAR’ THAT DESCRIBES GENE

REGULATORY INSTRUCTIONS CONTAINED IN GENOMIC DNA

Cis-regulatory function derives from cis-regulatory structure, therefore
understanding why a given developmental process occurs as it does requires an
understanding of the structure/function relations that exist within cis-regulatory enhancer
elements, which can be thought of as a “genetic code” for development. Despite the
importance of the genomic cis-regulatory enhancer sequences, our ability to identify and
predict functions for this category of DNA is extremely limited. Deciphering how these
elements interact with each other and the core promoter could in theory enable us to
predict the spatial and temporal activity of genes and to simulate their expression pattern
in silico. Thus, it is evident that from the structure and functional organization of cis-
regulatory DNA alone emerges an unprecedented explanatory and predictive power, in

respect to understanding and even controlling developmental processes.

39

In order to crack the ‘cis-regulatory code’ that is implicit in the sequence of
genomic regulatory DNA we need to 1) Identify cis-acting enhancer elements and
characterize their internal architectural organization; 2) Determine the biological
significance of their architectural design empirically and 3) Build and validate predictive

models for differential gene expression as a function of time and space.

Recent advances in computational methodologies for the identification and analysis of

cis-regulatory enhancer elements

C is-regulatory elements involved in gene regulation have been classically studied
over the last forty years using traditional genetic and biochemical approaches on a
promoter-by-promoter basis. Although the detailed inventory of single genes has offered
spectacular successes, a more global understanding of transcriptional regulatory switch
design is needed to comprehend the staggering complexity, versatility, and robustness of
living systems. Accompanying the expansion of large data sets that have resulted from
genomic sequencing and genome-wide expression profiling, new computational
strategies have been developed to contribute to the creation of a vocabulary that describes
gene regulatory instructions contained in genomic DNA on a global scale. High-
throughput sequencing technology, which has allowed the sequencing of complete
genomes of a large variety of species (including human, mouse, and rat), makes genome-
wide comparisons or phylogenetic footprinting feasible endeavors. High-throughput
expression profiling technology like microarrays (gene chip) and serial analysis of gene
expression (SAGE) have allowed rapid parallel analysis of expression levels of hundreds

of thousands of genes in a single assay. An underlying assumption of using high-

40

throughput expression profiling technology in regulatory region analysis is that co-

regulated genes ofien share a similar set of regulatory motifs.

Computational sequence analysis provides three broadly different approaches for
scanning genomic sequence to identify those regions predicted to participate in gene
regulation. First, inter-species sequence comparisons have been used to identify non-
coding sequences that have a reasonable likelihood of having gene regulatory properties
(Duret and Bucher, 1997; Gottgens et al., 2000; Hardison et al., 1997; Loots et al., 2000).
This is possible because sequences that mediate gene expression tend to be conserved
between species. It may be problematic, however, to assume that all functional sequences
are conserved and all nonfunctional sequences have diverged. For instance, sequence
comparisons of closely related species produce many false positive results, and
comparison of distantly related species has low power to detect functional elements in the
species compared (Dermitzakis and Clark, 2002). Moreover, it has been shown that
regulatory sequences can maintain function despite structural reorganization as a result of
species-specific loss and gain of transcription factor binding sites (Cuadrado et al., 2001;

Ludwig et al., 2000; Piano etal., 1999).

Sequence analysis of co-regulated genes within a species is a second approach for
predicting regulatory elements. This strategy is based on the fact that few transcription
factors exert their activity exclusively on single genes; rather, most bind to conserved
sites in several genes to coordinate their expression. Accordingly, genes are thought to be
co-regulated because they respond to similar regulatory pathways owing to shared non-
coding sequence motifs that direct the binding of specific sets of shared transcription

factors (F ickett and Wasserman, 2000; Wasserman et al., 2000). The correlation between

41

gene cluster and regulatory motifs is imprecise for at least three reasons. First, not all co-
regulated gene promoters share a common motif, because some of the identified genes in
a given cluster might in fact be secondary response genes. Second, because of the
combinatorial nature of transcription factors, the same motif can be found in the promoter
regions of genes that are not co-regulated. Third, some ‘motifs’ are likely to represent

random noise and in fact do not bind the transcription factor at all.

The third approach for the identification of gene regulatory sequences involves
generating and analyzing databases of known transcription-factor-binding sites and
screening genomic sequences for the presence of clusters of the known transcription-
factor-binding sites (Berman et al., 2002; Halfon et al., 2002; Halfon and Michelson,

2002; Markstein and Levine, 2002; Markstein et al., 2002).

One main difficulty with the output from computational searches for transcription
factor binding sites is the large number of false-positive and false negative results. For
example, Markstein et al. recently developed a computational method for identifying
clusters of Dorsal-binding sites in the Drosophila genome. A Dorsal-responsive silencer
from the target gene zerknullt was used to develop a specific model for a genome wide
computational scan with the aim of identifying other novel cis-acting enhancers that are
regulated by the Dorsal transcription factor. In this case clusters of at least three high
affinity Dorsal binding sites were sought within a 400 bp window. The search yielded 15
matches of which three were associated with known Dorsal-responsive genes. However,
many genes (for example, the rhomboid NEE) that are known to be regulated by the
Dorsal transcription factor were not identified on the basis of optimal binding clusters

(Markstein and Levine, 2002; Markstein et al., 2002). The short length and degenerate

42

nature of transcription-factor-binding sites account for most of these misleading
predictions. Also genes are rarely controlled by a single transcription factor, and in fact,
accumulating evidence suggests that specific combinations of transcription factors are
required to achieve proper biological function of cis-elements involved in directing
complex patterns of gene expression. One way to improve the computational
identification of cis-regulatory DNAs is to search for clustering of two or more different
classes of recognition sequences (Berman et al., 2002; Halfon et al., 2002; Halfon and
Michelson, 2002; Rajewsky et al., 2002). Despite these various ways to minimize the
number of false-positive binding sites and improve the hit rate, even those sequences that
meet the most stringent criteria might still be non-fiinctional in a genomic context. For
instance, many cases are known where factors must interact with other factors or
cofactors to be functional, and spatial correlations between their binding sites are

observed.

A Functional step towards understanding cis-regulatory architecture involved in

development

Comparative sequence analysis, coupled with the development of algorithms to
search genomic databases, have provided important tools for the identification of gene
regulatory elements at a scale not previously possible, but they have only been partially
successful at finding cis-regulatory motifs (as described above). In many cases, where
cis-regulatory elements predicted by computational and statistical methods appear

suitable, something in the arrangement of sites or a wider context renders the element

43

non-functional. Many times the techniques do not take into consideration the
combinatorial logic underlying enhancer architecture. Besides clustering of binding sites,
it is necessary to factor in binding site type, number, affinities, spacing, orientations and
order, in order to achieve better predictions and in the interpretation of their biological
function. These parameters are difficult to predict accurately a priori. Thus,
computational approaches would benefit from the availability of at least one well-defined

representative of a particular regulatory network to serve as a starting paradigm.

Our study (described in Chapter 3 of this thesis) takes a first step towards
providing just such a paradigm for early Drosophila developmental enhancers that are
regulated by the short-range transcriptional repressors. The construction and functional
assessment of synthetic enhancers to test particular combinatorial models of binding site
type, number, affinities, spacing, and relative positioning that need to be incorporated to
ensure the appropriate output aims to deduce the ‘grammar’ rules that define a functional
module. Incorporating the architectural parameters defined in our study into formal
clustering models might facilitate computational recognition of similar cis-regulatory

elements and aid in the interpretation of their biological function.

Regulatory Frontiers

The prospect of obtaining a truly global picture of the regulatory control system
of a complete eukaryotic organism with many thousands of genes seems daunting.
Despite the challenges, an initial framework offering a rough roadmap appears to have
been established. Current successes in large-scale sequencing and gene identification

have provided the identity and physical location of individual genes. Increasingly more

44

powerful genomic technologies will enable us to identify many of the changes in cis-
regulatory DNAs, the corresponding changes in gene expression patterns between
organisms, and the role of these differences in speciation. The systematic integration of
diverse data sets such as high throughput in situ hybridization screens that provide spatial
and temporal information about transcription factor expression profiles, genome wide
transcription factor binding analysis (Lee et al., 2002; Weinmann and Famham, 2002;
Weinmann et al., 2002), protein—protein interaction networks together with improved
regulatory sequence identification strategies will provide an integrated platform for
deciphering the transcriptional regulatory network. As we enter the post- genome era, it is
possible to envision the elucidation of the transcriptional cis-regulatory code, whereby
the information content of cis-regulatory DNAs can be predicted by simple sequence

analysis.

45

REFERENCES

Adams, M. D. Celniker, S. E. Holt, R. A. Evans, C. A. Gocayne, J. D. Amanatides, P.
G. Scherer, S. E. Li, P. W. Hoskins, R. A. Galle, R. F. et al. (2000). The genome
sequence of Drosophila melanogaster. Science 287, 2185-95.

Andrioli, L. P., Vasisht, V., Theodosopoulou, E., Oberstein, A. and Small, S. (2002).
Anterior repression of a Drosophila stripe enhancer requires three position-specific
mechanisms. Development 129, 4931-40.

Arnosti, D. N. (2002). Design and function of transcriptional switches in Drosophila.
Insect Biochem Mol Biol 32, 1257-73.

Arnosti, D. N. (2003). Analysis and function of transcriptional regulatory elements:
Insights from Drosophila. Annu Rev Entomol 48, 579-602.

Arnosti, D. N., Barolo, S., Levine, M. and Small, S. (1996a). The eve stripe 2 enhancer
employs multiple modes of transcriptional synergy. Development 122, 205-14.

Arnosti, D. N., Gray, 8., Barolo, 8., Zhou, J. and Levine, M. (1996b). The gap protein
knirps mediates both quenching and direct repression in the Drosophila embryo. Embo J
15, 3659-66.

Asoh, S., Lee-Kwon, W., Mouradian, M. M. and Nirenberg, M. (1994). Selection of
DNA clones with enhancer sequences. Proc Natl Acad Sci U S A 91, 6982-6.

Balasubramanian, P., Zhao, L. J. and Chinnadurai, G. (2003). Nicotinamide adenine
dinucleotide stimulates oligomerization, interaction with adenovirus EIA and an intrinsic
dehydrogenase activity of CtBP. FEBS Lett 537, 157-60.

Banerji, J., Rusconi, S. and Schaffner, W. (1981). Expression of a beta-globin gene is
enhanced by remote SV40 DNA sequences. Cell 27, 299-308.

Barolo, S. and Levine, M. (1997). hairy mediates dominant repression in the Drosophila
embryo. Embo J 16, 2883-91.

Barolo, S. and Posakony, J. W. (2002). Three habits of highly effective signaling
pathways: principles of transcriptional control by developmental cell signaling. Genes
Dev 16, 1167-81.

Belting, H. G., Shashikant, C. S. and Ruddle, F. H. (1998). Modification of expression
and cis-regulation of Hoxc8 in the evolution of diverged axial morphology. Proc Natl
Acad Sci U S A 95, 2355-60.

Berman, B. P., Nibu, Y., Pfeiffer, B. D., Tomancak, P., Celniker, S. E., Levine, M.,
Rubin, G. M. and Eisen, M. B. (2002). Exploiting transcription factor binding site
clustering to identify cis-regulatory modules involved in pattern formation in the
Drosophila genome. Proc Natl Acad Sci U S A 99, 757-62.

46

Biggar, S. R. and Crabtree, G. R. (2001). Cell signaling can direct either binary or
graded transcriptional responses. Embo J 20, 3167-76.

Blackwood, E. M. and Kadonaga, J. T. (1998). Going the distance: a current view of
enhancer action. Science 281, 61-3.

Burz, D. S., Rivera-Pomar, R., Jackle, H. and Hanes, S. D. (1998). Cooperative DNA-
binding by Bicoid provides a mechanism for threshold-dependent gene activation in the
Drosophila embryo. Embo J 17, 5998-6009.

Capovilla, M., Eldon, E. D. and Pirrotta, V. (1992). The giant gene of Drosophila
encodes a b-ZIP DNA-binding protein that regulates the expression of other segmentation
gap genes. Development 114, 99-112.

Carey, M., Lin, Y. S., Green, M. R. and Ptashne, M. (1990). A mechanism for
synergistic activation of a mammalian gene by GAL4 derivatives. Nature 345, 361-4.

Chi, T., Lieberman, P., Ellwood, K. and Carey, M. (1995). A general mechanism for
transcriptional synergy by eukaryotic activators. Nature 377, 254-7.

Chinnadurai, G. (2002). CtBP, an unconventional transcriptional corepressor in
development and oncogenesis. Mol Cell 9, 213-24.

Chuzhanova, N. A., Krawczak, M., Nemytikova, L. A., Gusev, V. D. and Cooper, D.
N. (2000). Promoter shuffling has occurred during the evolution of the vertebrate grth
hormone gene. Gene 254, 9-18.

Cuadrado, M., Sacristan, M. and Antequera, F. (2001). Species-specific organization
of CpG island promoters at mammalian homologous genes. EMBO Rep 2, 586-92.

Davidson, E. H. (2001). Genomic Regulatory Systems: Development and Evolution:
Academic Press: A Harcourt Science and Technology Company.

Dermitzakis, E. T. and Clark, A. G. (2002). Evolution of transcription factor binding
sites in Mammalian gene regulatory regions: conservation and turnover. Mol Biol Evol
19, 1114-21.

Duret, L. and Bucher, P. (1997). Searching for regulatory elements in human noncoding
sequences. Curr Opin Struct Biol 7, 399-406.

Durick, K., Mendlein, J. and Xanthopoulos, K. G. (1999). Hunting with traps:
genome-wide strategies for gene discovery and functional analysis. Genome Res 9, 1019-
25.

Eldon, E. D. and Pirrotta, V. (1991). Interactions of the Drosophila gap gene giant with
maternal and zygotic pattem-forming genes. Development 111, 367 -7 8.

47

Fickett, J. W. and Wasserman, W. W. (2000). Discovery and modeling of
transcriptional regulatory regions. Curr Opin Biotechnol 11, 19-24.

Fraseh, M., Hoey, T., Rushlow, C., Doyle, H. and Levine, M. (1987). Characterization
and localization of the even-skipped protein of Drosophila. Embo J 6, 749-59.

Frasch, M. and Levine, M. (1987). Complementary patterns of even-skipped and fushi
tarazu expression involve their differential regulation by a common set of segmentation
genes in Drosophila. Genes Dev 1, 981-95.

Fujioka, M., Emi-Sarker, Y., Yusibova, G. L., Goto, T. and Jaynes, J. B. (1999).
Analysis of an even-skipped rescue transgene reveals both composite and discrete
neuronal and early blastoderm enhancers, and multi-stripe positioning by gap gene
repressor gradients. Development 126, 2527-38.

Fukushige, S. and Ikeda, J. E. (1996). Trapping of mammalian promoters by Cre-lox
site-specific recombination. DNA Res 3, 73-80.

Gaudreau, L., Schmid, A., Blaschke, D., Ptashne, M. and Horz, W. (1997). RNA
polymerase 11 holoenzyme recruitment is sufficient to remodel chromatin at the yeast
PHOS promoter. Cell 89, 55-62.

Gaul, U., Seifert, E., Schuh, R. and Jackle, H. (1987). Analysis of Kruppel protein
distribution during early Drosophila development reveals posttranscriptional regulation.
Cell 50, 639-47.

Gergen, J. P. and Wieschaus, E. (1986). Dosage requirements for runt in the
segmentation of Drosophila embryos. Cell 45, 289-99.

Goto, T., Macdonald, P. and Maniatis, T. (1989). Early and late periodic patterns of
even skipped expression are controlled by distinct regulatory elements that respond to
different spatial cues. Cell 57, 413-22.

Gottgens, 8., Barton, L. M., Gilbert, J. G., Bench, A. J., Sanchez, M. J., Bahn, S.,
Mistry, S., Grafham, D., McMurray, A., Vaudin, M. et al. (2000). Analysis of
vertebrate SCL loci identifies conserved enhancers. Nat Biotechnol 18, 181-6.

Gray, S. and Levine, M. (1996). Transcriptional repression in development. Curr Opin
Cell Biol 8, 358-64.

Gray, 8., Szymanski, P. and Levine, M. (1994). Short-range repression permits
multiple enhancers to function autonomously within a complex promoter. Genes Dev 8,
1829-38.

Halfon, M. 8., Grad, Y., Church, G. M. and Michelson, A. M. (2002). Computation-
based discovery of related transcriptional regulatory modules and motifs using an
experimentally validated combinatorial model. Genome Res 12, 1019-28.

48

Halfon, M. S. and Michelson, A. M. (2002). Exploring genetic regulatory networks in
metazoan development: methods and models. Physiol Genomics 10, 131-43.

Han, W., Yu, Y., Su, K., Kohanski, R. A. and Pick, L. (1998). A binding site for
multiple transcriptional activators in the fushi tarazu proximal enhancer is essential for
gene expression in vivo. Mol Cell Biol 18, 3384-94.

Harding, K., Hoey, T., Warrior, R. and Levine, M. (1989). Autoregulatory and gap
gene response elements of the even—skipped promoter of Drosophila. Embo J 8, 1205-12.

Harding, K., Rushlow, C., Doyle, H. J., Hoey, T. and Levine, M. (1986). Cross-
regulatory interactions among pair-rule genes in Drosophila. Science 233, 953-9.

Hardison, R., Slightom, J. L., Gumucio, D. L., Goodman, M., Stojanovic, N. and
Miller, W. (1997). Locus control regions of mammalian beta-globin gene clusters:
combining phylogenetic analyses and experimental results to gain functional insights.
Gene 205, 73-94.

Hewitt, G. F., Strunk, B. S., Margulies, C., Priputin, T., Wang, X. D., Amey, R.,
Pabst, B. A., Kosman, D., Reinitz, J. and Arnosti, D. N. (1999). Transcriptional
repression by the Drosophila giant protein: cis element positioning provides an
alternative means of interpreting an effector gradient. Development 126, 1201-10.

Hoch, M., Gerwin, N., Taubert, H. and Jackle, H. (1992). Competition for overlapping
sites in the regulatory region of the Drosophila gene Kruppel. Science 256, 94-7.

Keller, S. A., Mao, Y., Struffi, P., Margulies, C., Yurk, C. E., Anderson, A. R.,
Amey, R. L., Moore, 8., Ebels, J. M., Foley, K. et al. (2000). dCtBP-dependent and -
independent repression activities of the Drosophila Knirps protein. Mol Cell Biol 20,

7247-585

Kim, T. K. and Maniatis, T. (1997). The mechanism of transcriptional synergy of an in
vitro assembled interferon-beta enhanceosome. Mol Cell 1, 119-29.

Kosman, D. and Small, S. (1997). Concentration-dependent patterning by an ectopic
expression domain of the Drosophila gap gene knirps. Development 124, 1343-54.

Kraut, R. and Levine, M. (1991a). Mutually repressive interactions between the gap
genes giant and Kruppel define middle body regions of the Drosophila embryo.
DeveIOpment 111, 611-21.

Kraut, R. and Levine, M. (1991b). Spatial regulation of the gap gene giant during
Drosophila development. Development 111, 601-9.

Kumar, V., Carlson, J. E., Ohgi, K. A., Edwards, T. A., Rose, D. W., Escalante, C.
R., Rosenfeld, M. G. and Aggarwal, A. K. (2002). Transcription corepressor CtBP is an
NAD(+)-regulated dehydrogenase. Mol Cell 10, 857-69.

49

La Rosee, A., Hader, T., Taubert, H., Rivera-Pomar, R. and Jackle, H. (1997).
Mechanism and Bicoid-dependent control of hairy stripe 7 expression in the posterior
region of the Drosophila embryo. Embo J 16, 4403-11.

La Rosee-Borggreve, A., Hader, T., Wainwright, D., Sauer, F. and J ackle, H. (1999).
hairy stripe 7 element mediates activation and repression in response to different domains
and levels of Kruppel in the Drosophila embryo. Mech Dev 89, 133-40.

Lee, T. 1., Rinaldi, N. J., Robert, F., Odom, D. T., Bar-Joseph, Z., Gerber, G. K.,
Hannett, N. M., Harhison, C. T., Thompson, C. M., Simon, I. et al. (2002).
Transcriptional regulatory networks in Saccharomyces cerevisiae. Science 298, 799-804.

Leung, J. Y., McKenzie, F. E., Uglialoro, A. M., Flores-Villanueva, P. O., Sorkin, B.
C., Yunis, E. J., Hartl, D. L. and Goldfeld, A. E. (2000). Identification of phylogenetic
footprints in primate tumor necrosis factor-alpha promoters. Proc Natl Acad Sci U S A
97, 6614-8.

Lewis, E. B. (1998). The bithorax complex: the first fifty years. Int J Dev Biol 42, 403-
15.

Licht, J. D., Ro, M., English, M. A., Grossel, M. and Hansen, U. (1993). Selective
repression of transcriptional activators at a distance by the Drosophila Kruppel protein.
Proc Natl Acad Sci U S A 90, 11361-5.

Liu, T., Wu, J. and He, F. (2000). Evolution of cis-acting elements in 5' flanking
regions of vertebrate actin genes. J Mol Evol 50, 22-30.

Loots, G. G., Locksley, R. M., Blankespoor, C. M., Wang, Z. E., Miller, W., Rubin,
E. M. and Frazer, K. A. (2000). Identification of a coordinate regulator of interleukins
4, 13, and 5 by cross-species sequence comparisons. Science 288, 136-40.

Ludwig, M. Z., Bergman, C., Patel, N. H. and Kreitman, M. (2000). Evidence for
stabilizing selection in a eukaryotic enhancer element. Nature 403, 564-7.

Ludwig, M. Z. and Kreitman, M. (1995). Evolutionary dynamics of the enhancer
region of even-skipped in Drosophila. Mol Biol Evol 12, 1002-11.

Ludwig, M. Z., Patel, N. H. and Kreitman, M. (1998). Functional analysis of eve stripe
2 enhancer evolution in Drosophila: rules governing conservation and change.
Development 125, 949-58.

Lunyak, V. V., Burgess, R, Prefontaine, G. G., Nelson, C., Sze, S. H., Chenoweth, J.,
Schwartz, P., Pevzner, P. A., Glass, C., Mandel, G. et al. (2002). Corepressor-
dependent silencing of chromosomal regions encoding neuronal genes. Science 298,
1747-52.

50

Macdonald, P. M., Ingham, P. and Struhl, G. (1986). Isolation, structure, and
expression of even-skipped: a second pair-rule gene of Drosophila containing a homeo
box. Cell 47, 721-34.

Macdonald, P. M. and Struhl, G. (1986). A molecular gradient in early Drosophila
embryos and its role in specifying the body pattern. Nature 324, 537-45.

Magis, W., Fiering, S., Groudine, M. and Martin, D. I. (1996). An upstream activator
of transcription coordinately increases the level and epigenetic stability of gene
expression. Proc Natl Acad Sci U S A 93, 13914-8.

Markstein, M. and Levine, M. (2002). Decoding cis-regulatory DNAs in the
Drosophila genome. Curr Opin Genet Dev 12, 601-6.

Markstein, M., Markstein, P., Markstein, V. and Levine, M. S. (2002). Genome-wide
analysis of clustered Dorsal binding sites identifies putative target genes in the
Drosophila embryo. Proc Natl Acad Sci U S A 99, 763-8.

Marmorstein, R. (2002). Dehydrogenases, NAD, and transcription--what's the
connection? Structure (Camb) 10, 1465-6.

Miller, R. D., Taillon-Miller, P. and Kwok, P. Y. (2001). Regions of low single-
nucleotide polymorphism incidence in human and orangutan xq: deserts and recent
coalescences. Genomics 71, 78-88.

Munshi, N., Agalioti, T., Lomvardas, S., Merika, M., Chen, G. and Thanos, D.
(2001). Coordination of a transcriptional switch by HMGI(Y) acetylation. Science 293,
1133-6.

Nardini, M., Spano, S., Cericola, C., Pesce, A., Massaro, A., Millo, E., Luini, A.,
Corda, D. and Bolognesi, M. (2003). CtBP/BARS: a dual-function protein involved in
transcription co-repression and Golgi membrane fission. Embo J 22, 3122-30.

Nibu, Y., Senger, K. and Levine, M. (2003). CtBP-independent repression in the
Drosophila embryo. Mol Cell Biol 23, 3990-9.

Nusslein-Volhard, C., Frohnhofer, H. G. and Lehmann, R. (1987). Determination of
anteroposterior polarity in Drosophila. Science 238, 1675-81.

Ogbourne, S. and Antalis, T. M. (1998). Transcriptional control and the role of
silencers in transcriptional regulation in eukaryotes. Biochem J 331 (Pt 1), 1-14.

Pankratz, M. J., Hoch, M., Seifert, E. and Jackle, H. (1989). Kruppel requirement for
knirps enhancement reflects overlapping gap gene activities in the Drosophila embryo.
Nature 341, 337-40.

51

Papatsenko, D. A., Makeev, V. J., Lifanov, A. P., Regnier, M., Nazina, A. G. and
Desplan, C. (2002). Extraction of functional binding sites from unique regulatory
regions: the Drosophila early developmental enhancers. Genome Res 12, 470-81.

Petschek, J. P., Perrimon, N. and Mahowald, A. P. (1987). Region-specific defects in
l(1)giant embryos of Drosophila melanogaster. Dev Biol 119, 175-89.

Piano, F., Parisi, M. J., Karess, R. and Kamhysellis, M. P. (1999). Evidence for
redundancy but not trans factor-cis element coevolution in the regulation of Drosophila
Yp genes. Genetics 152, 605-16.

Postigo, A. A. and Dean, D. C. (1999). Independent repressor domains in ZEB regulate
muscle and T-cell differentiation. Mol Cell Biol 19, 7961-71.

Rajewsky, N., Vergassola, M., Gaul, U. and Siggia, E. D. (2002). Computational
detection of genomic cis-regulatory modules applied to body patterning in the early
Drosophila embryo. BMC Bioinformatics 3, 30.

Rivera-Pomar, R. and Jackle, H. (1996). From gradients to stripes in Drosophila
embryogenesis: filling in the gaps. Trends Genet 12, 478-83.

Rossi, F. M., Kringstein, A. M., Spicher, A., Guicherit, O. M. and Blau, H. M.
(2000). Transcriptional control: rheostat converted to on/off switch. Mol Cell 6, 723-8.

Rothe, M., Nauber, U. and Jackle, H. (1989). Three hormone receptor-like Drosophila
genes encode an identical DNA-binding finger. Embo J 8, 3087-94.

Ryu, J. R. and Arnosti, D. N. (2003). Functional similarity of Knirps CtBP-dependent
and CtBP-independent transcriptional repressor activities. Nucleic Acids Res 31, 4654-62.

Schaeper, U., Boyd, J. M., Verma, S., Uhlmann, E., Subramanian, T. and
Chinnadurai, G. (1995). Molecular cloning and characterization of a cellular
phosphoprotein that interacts with a conserved C-tenninal domain of adenovirus ElA
involved in negative modulation of oncogenic transformation. Proc Natl Acad Sci U S A
92, 10467-71.

Carroll, S. B., Grenier, J. K., and Weatherbee, S. D. (2001). From DNA to Diversity:
Molecular Genetics and the Evolution of Animal Design: Blackwell Science.

Shashikant, C. S., Kim, C. B., Borbely, M. A., Wang, W. C. and Ruddle, F. H.
(1998). Comparative studies on mammalian Hoxc8 early enhancer sequence reveal a
baleen whale-specific deletion of a cis-acting element. Proc Natl Acad Sci U S A 95,
15446-51.

Shi, Y., Sawada, J., Sui, G., Affar el, B., Whetstine, J. R., Lan, F., Ogawa, H., Luke,
M. P. and Nakatani, Y. (2003). Coordinated histone modifications mediated by a CtBP
co—repressor complex. Nature 422, 735-8.

52

Simpson-Brose, M., Treisman, J. and Desplan, C. (1994). Synergy between the
hunchback and bicoid morphogens is required for anterior patterning in Drosophila. Cell
78, 855-65.

Small, S., Arnosti, D. N. and Levine, M. (1993). Spacing ensures autonomous
expression of different stripe enhancers in the even-skipped promoter. Development 119,
762-72.

Small, S., Blair, A. and Levine, M. (1992). Regulation of even-skipped stripe 2 in the
Drosophila embryo. Embo J 11, 4047-57.

Small, S., Blair, A. and Levine, M. (1996). Regulation of two pair-rule stripes by a
single enhancer in the Drosophila embryo. Dev Biol 175, 314-24.

Small, S., Kraut, R., Hoey, T., Warrior, R. and Levine, M. (1991). Transcriptional
regulation of a pair-rule stripe in Drosophila. Genes Dev 5, 827-39.

Small, S. and Levine, M. (1991). The initiation of pair-rule stripes in the Drosophila
blastoderm. Curr Opin Genet Dev 1, 255-60.

Stanojevic, D., Hoey, T. and Levine, M. (1989). Sequence-specific DNA-binding
activities of the gap proteins encoded by hunchback and Kruppel in Drosophila. Nature
341, 331-5.

Stanojevic, D., Small, S. and Levine, M. (1991). Regulation of a segmentation stripe by
overlapping activators and repressors in the Drosophila embryo. Science 254, 1385-7.

Stormo, G. D. (2000a). DNA binding sites: representation and discovery. Bioinformatics
16, 16-23.

Stormo, G. D. (2000b). Gene-finding approaches for eukaryotes. Genome Res 10, 394-7.

Stormo, G. D. (2000c). Identification of coordinated gene expression and regulatory
sequences. Pac Symp Biocomput, 416-7.

Struhl, K. (2001). Gene regulation. A paradigm for precision. Science 293, 1054-5.

Strunk, B., Struffi, P., Wright, K., Pabst, B., Thomas, J., Qin, L. and Arnosti, D. N.
(2001). Role of CtBP in transcriptional repression by the Drosophila giant protein. Dev
Biol 239, 229-40.

Subramanian, T. and Chinnadurai, G. (2003). Association of class I histone
deacetylases with transcriptional corepressor CtBP. FEBS Lett 540, 255-8.

Sundqvist, A., Sollerbrant, K. and Svensson, C. (1998). The carboxy-terminal region
of adenovirus ElA activates transcription through targeting of a C-terrninal binding
protein-histone deacetylase complex. FEBS Lett 429, 183-8.

53

Thanos, D. and Maniatis, T. (1995). Virus induction of human IFN beta gene
expression requires the assembly of an enhanceosome. Cell 83, 1091-100.

Turner, J. and Crossley, M. (2001). The CtBP family: enigmatic and enzymatic
transcriptional co-repressors. Bioessays 23, 683-90.

Vinson, C. R., Conover, S. and Adler, P. N. (1989). A Drosophila tissue polarity locus
encodes a protein containing seven potential transmembrane domains. Nature 338, 263-4.

Walters, M. C., Magis, W., Fiering, S., Eidemiller, J., Scalzo, D., Groudine, M. and
Martin, D. I. (1996). Transcriptional enhancers act in cis to suppress position-effect
variegation. Genes Dev 10, 185-95.

Wasserman, J. D., Urban, S. and Freeman, M. (2000). A family of rhomboid-like
genes: Drosophila rhomboid-l and roughoid/rhomboid-3 cooperate to activate EGF
receptor signaling. Genes Dev 14, 1651-63.

Weinmann, A. S. and Farnham, P. J. (2002). Identification of unknown target genes of
human transcription factors using chromatin immunoprecipitation. Methods 26, 37-47.

Weinmann, A. S., Yan, P. S., Oberley, M. J., Huang, T. H. and Farnham, P. J.
(2002). Isolating human transcription factor targets by coupling chromatin
immunoprecipitation and CpG island microarray analysis. Genes Dev 16, 235-44.

Weintraub, H. (1988). Formation of stable transcription complexes as assayed by
analysis of individual templates. Proc Natl Acad Sci U S A 85, 5819-23.

Wu, X., Vakani, R. and Small, S. (1998). Two distinct mechanisms for differential
positioning of gene expression borders involving the Drosophila gap protein giant.
Development 125, 3765-74.

Yan, R., Small, S., Desplan, C., Dearolf, C. R. and Darnell, J. E., Jr. (1996).
Identification of a Stat gene that functions in Drosophila deve10pment. Cell 84, 421-30.

Yuh, C. H., Bolouri, H. and Davidson, E. H. (1998). Genomic cis-regulatory logic:
experimental and computational analysis of a sea urchin gene. Science 279, 1896-902.

Zhao, C., Dave, V., Yang, F., Scarborough, T. and Ma, J. (2000). Target selectivity of
bicoid is dependent on nonconsensus site recognition and protein-protein interaction. Mol
Cell Biol 20, 8112-23.

Zuo, P., Stanojevic, D., Colgan, J., Han, K., Levine, M. and Manley, J. L. (1991).
Activation and repression of transcription by the gap proteins hunchback and Kruppel in
cultured Drosophila cells. Genes Dev 5, 254-64.

54

Chapter 11

Information Display by Transcriptional Enhancers

Kulkarni, M. M. and Arnosti, D. N. (2003). Information Display by Transcriptional

enhancer. Development 130, 6569-6575.

55

ABSTRACT

Transcriptional enhancers integrate positional and temporal information to
regulate the complex expression of developmentally controlled genes. Current models
suggest that enhancers act as computational devices, receiving multiple inputs from
activators and repressors and resolving them into a single positive or a negative signal
that is transmitted to the basal transcriptional machinery. We show here that a simple,
compact enhancer is capable of representing both repressed and activated states at the
same time and in the same nucleus. This finding suggests that closely apposed factor
binding sites, situated within compact cis-elements, can be independently interpreted by
the transcriptional machinery, possibly through successive enhancer-promoter
interactions. These results provide clear evidence that the computational fimctions
usually ascribed to the enhancer itself are actually shared with the basal machinery. In
contrast to the autonomous computer model of enhancer function, an infonnation-display
or “billboard” model of enhancer activity may better describe many developmentally

regulated transcriptional enhancers.

INTRODUCTION

Developmental programs of gene expression are controlled by “hard-wired”
transcriptional circuits comprised of modular enhancers that communicate with basal
promoter regions (Davidson, 2001). Prior studies in many systems have supported the

general notion that an enhancer acts as an information-processing device, or computer,

56

receiving multiple inputs in the form of distinct transcription factors, both activators and
repressors, that bind to it (Davidson, 2001; Ghazi and VijayRaghavan, 2000). The
analogy of an enhancer as a computer is usually simply that of an element that sorts out
inputs (processing) and resolves them into a single output that is instructive to the basal
machinery, either turning the gene on or off. An important point is that computational
functions - the decision to fire a promoter and at what level - have been ascribed to the
enhancer. This is not to suggest that a given enhancer has only a single possible output;
depending on signals or cell type, the same enhancer element might activate or repress,
and the magnitude of activation signals can be variable, in the manner of a rheostat
(Barolo and Posakony, 2002; Biggar and Crabtree, 2001; Rossi et al., 2000). However, it
has been thought that enhancers do perform an integrative function and that in a
particular nucleus, an enhancer represents a single information state at any given

moment.

Such integrative functions have been ascribed to the human interferon-B (IFN-B)
enhancer, which drives transcription of the IFN-B gene in response to viral infection
(Struhl, 2001). The presence of each transcription factor binding site and its precise
arrangement within the regulatory element are critical for the various regulatory proteins
(sequence-specific activators and architectural proteins) to assemble through cooperative
interactions into a well-defined nucleoprotein complex called the “enhanceosome”.
Assembly of the enhanceosome is essential for the transcription of the IFN-B gene in
response to viral infection in cells. In this structured element, the exact arrangement of
factor binding sites is critical to dictating the output of the element, so the enhanceosome

acts as a molecular computer, leading to a single output directed to the general machinery

57

(Thanos and Maniatis, 1995; Kim and Maniatis, 1997; Munshi et al., 2001). Such a
complex might provide a stereospecific interface for interaction with the basal
transcriptional machinery, possibly engaging several components of the basal machinery
simultaneously to effect synergistic activation (Carey et al., 1990; Chi et al., 1995). With
such an enhancer, the target gene would be activated only upon the assembly of a
“complete” complex, providing a precise on/off binary transcriptional switch in response

to the appropriate stimulus.

Studies of developmentally regulated genes have also provided examples of
enhancers as molecular computers. The developmentally regulated Drosophila even-
skipped (eve) gene is regulated by developmental enhancers that are thought to act in a
computational fashion. The reiterated stripe pattern of eve expression in the blastoderm
embryo is generated by modular enhancers bound by broadly expressed transcriptional
activators and regionally distributed repressors (Fujioka et al., 1999; Small et al., 1992;
Small et al., 1996). These enhancers interpret gradients of regulatory factors and are
active or inactive, depending on the particular set of regulatory proteins present in a given
nucleus. The eve stripe 2 enhancer is active only in a narrow band of cells where
activators Bicoid and Hunchback are present, but repressors Kriippel, Giant, and Sloppy-
paired are scarce or absent (Andrioli et al., 2002; Small et al., 1992). Key to the
functional autonomy of the modular eve enhancers is the short-range of the repressors
that regulate individual enhancers; for example, the short-range transcriptional repressor
Kruppel bound to the stripe 2 enhancer in central regions of the embryo does not interfere
with the activity of the adjacent eve stripe 3 enhancer (Small et al., 1993). An

assumption is that each enhancer works as a single computational unit, not a redundant

58

set of independently acting elements. Consistent with this view is the finding that
enhancer function is disrupted upon loss of a single activator or repressor site (Amosti et
al., 1996a; Small et al., 1992). However, these experiments have relied on minimal
elements that may already represent a subset of the actual regulatory region (see

Discussion).

A more detailed picture emerges from the functional dissection of the endo 16 cis-
regulatory region of Strongylocentrotus purpuratus. The endo 16 gene is regulated
during development by 2.3 kbp region containing binding sites for factors that contribute
to distinct functions such as early widespread activation, late activation, repression of the
early element, and potentiation of the repressor sites. Separate portions of the regulatory
region can be combined to recreate some or all of the expression pattern, and models
based on Boolean logical operators successfully simulate the output of these regulatory
regions (Yuh et al., 1998). These studies emphasize the integrative, computer-like
processing suggested to be a characteristic of developmental enhancers, and suggest that

basal elements respond to signals generated by these molecular logic circuits.

In contrast to this view of the enhancer as an information-processing unit, we find
that a single, compact enhancer can serve as an information display, representing on and
off states, at the same time and in the same nucleus. This finding suggests that rather than
acting as a computer that integrates various inputs, enhancers can simultaneously display
both the active and repressed states, which maybe interpreted by successive or multiple,
simultaneous interactions with the basal transcriptional machinery. In this case, the

enhancer does not act in a concerted, computational fashion, and the basal transcriptional

59

machinery plays an active, rather than a passive, role in interpreting signals from the

enhancer.

MATERIALS AND METHODS

l. Plasmid construction

GAL4 (aa1-93) - GAL4 AD (aa753-881)

A KpnI-Xbal fragment from pSCTEV GAL4 (1-93)- GAL4 (Seipel et al., 1992)
containing the reading frame for the yeast GAL4 activation domain (Gal4 AD) fiom
amino acid residues 753-881, was cloned into KpnI-Xbal cut pTwiggy (Amosti et al.,
1996b) vector, which contains the twist enhancer (2xPEe-Et) element, twist basal

promoter and the GAL4 DNA-binding domain from residues 1-93.

Reporter genes.

The plasmid UAS-lacZ (Brand and Pen'imon, 1993) was modified to contain two
Giant sites (5’ GGC CGC TAT GAC GCA AGA AGA CCC AGA TCT TTT TAT
GAC GCA AGA GA 3’) or two Knirps sites (S’GGC CGC ATC TGA TCT AGT TTG
TAC TAG ACA TCT GAT CTA GTT TCA 3’) twenty nucleotides upstream of the five
GAL4 binding sites. The resulting vectors named M2g5u-lacZ or M2k5u-lacZ (Fig. 1C,
D) respectively, consist of two Giant or Knirps binding sites, five tandemly arrayed
GAL4 binding sites, followed by the hsp 70 TATA box and transcriptional start driving

lacZ expression. These reporters were further modified by introducing oligos containing

60

two Twist (twi) and two Dorsal (dl) binding sites (Szymanski and Levine, 1995) at the
Not I site upstream of the Giant or Knirps sites resulting in the 2twi.dl-M2g5u-lacZ and

2twi.dl-M2k5u-lacZ reporters (Fig. 1A, B, E and F).

The regulatory element from 2twi.dl-M2g5u-lacZ, containing two Twist sites, two
Dorsal sites, two Giant sites and five GAL4 binding sites was introduced into the EcoRI
site of the C4PLZ vector in both orientations. The C4PLZ vector lies between two
divergently transcribed genes, the TATA-less white (w) gene and the lacZ gene. The lacZ
gene is driven by the TATA containing P element transposase basal promoter (Fig. 2).
Two additional Giant binding sites were introduced at the SphI site in the M2g5u-lacZ
vector between the five GAL4 binding sites and the hsp 70 TATA box. The resulting
vector was further modified by introducing oligos containing two Twist (twi) and two
Dorsal (d1) binding sites (Szymanski and Levine, 1995) at the Not I site upstream of the

Giant sites resulting in the 2twi.dl-M2g5u2g-lacZ (Fig. 3).

2. P-element transformation, crosses to reporter genes, and whole-mount in situ

hybridization of embryos.

P-element transformation vectors were introduced into the Drosophila gerrnline
by injection of yw"7 embryos as described (Small et al., 1992). Embryos were collected
either directly from each transgenic reporter line or from a cross between a reporter line
and a line expressing the GAL4 activator in the ventral regions of the embryo. The
embryos were fixed and stained using digoxigenin-UTP labeled antisense RNA probes to

either lacZ or w as described (Small et al., 1992).

61

RESULTS
Limited ability of short-range repressors to block activators

The activity of deve10pmental cis-regulatory elements has been studied mostly in
the context of complex endogenous enhancers (Amosti et al., 1996a; Gray et al., 1994;
Kosman and Small, 1997; Small et al., 1993; Small et al., 1992). This approach is
complicated by the functional complexity of many cis-regulatory elements where the
identity and/or the stoichiometry of transacting factors is not always well defined. To
analyze enhancer function in a setting in which activator-repressor stoichiometry and
spacing can be exactly defined, we constructed chromosomally integrated, compact
regulatory elements containing binding sites for endogenous short-range repressors Giant
or Knirps, endogenous activators Twist and Dorsal, and chimeric Gal4 activators. The
space between repressor and activator sites on these elements is less than 100 bp, a
distance over which short-range repressors have been previously shown to be effective
(Amosti et al., 1996b; Gray et al., 1994; Hewitt et al., 1999; Keller et al., 2000). Twist
and Dorsal drive gene expression in a ventral swathe approximately 22-24 cell in width,
while the Gal4 activator protein, expressed under the control of the twist enhancer, drives
reporter gene expression in a narrower 18-20 cell wide pattern. The protein product of the
gap gene giant is present in broad anterior and posterior stripes, while the Knirps protein
is present in a broad posterior stripe and more anterior regions in the early embryo. As
anticipated, Giant and Knirps mediate repression of adjacent Dorsal and Twist activators,
eliminating expression of the lacZ reporter gene in portions of the embryo where these

repressor proteins are localized (Figure II-lA, B). Strikingly however, Giant and Knirps

62

are unable to repress an element containing five Gal4 activator sites, although these
proteins also bind within 100 bp of the repressors, revealing a hitherto unknown
limitation of short-range repressors (Figure II-lC, D). This lack of repression is not due
to an inherent resistance of the Gal4 activation domain to repression, for Knirps and
Giant can effectively repress an element containing only three Gal4 binding sites (M.

Kulkami, unpublished).

Simultaneous repression and activation

When Gal4 activators are combined with Dorsal and Twist activators on a
composite element, strongly enhanced staining is noted in the central regions of the
embryo, indicative of additive or synergistic activation. In the regions of the embryo
containing the repressors Giant or Knirps, the width of the nuclei stained (a 18-20 swathe
of nuclei) is the same as the pattern of staining driven by the Gal4 protein alone. We
conclude that in nuclei containing Giant and Knirps protein, the pattern of staining
directed by Dorsal and Twist is being selectively repressed by the short-range repressors,
while transcription driven by Gal4 (a narrower 18-20 nuclei swathe) is unimpeded
(Figure II-lE, F). The pattern of gene expression indicates that, in nuclei where the
activators and repressors are co-expressed, transcription is driven by one cluster of
activators within the compact regulatory element, while at the same time other activators
within the same element are being actively repressed by Giant or Knirps. This compact

regulatory element therefore, has subelements that represent both “active” and “inactive”

63

Figure II-l: Simultaneous repression and activation from a compact regulatory
element.

(A) Knirps repression of adjacent Dorsal (dl) and Twist (twi) activators. Dorsal and
Twist proteins, normally active in a broad (22-24 nuclei) ventral swathe of the blastoderm
embryo, fail to activate a linked hsp 70 lacZ transgene in regions containing Knirps (kni)
protein (arrow).

(B) Giant repression of Dorsal and Twist. Repression is seen in anterior and posterior
regions where the Giant (gt) repressor is expressed (arrows). (C, D) Gal4 activators,
expressed in a narrower (18-20 nuclei) ventral swathe, are not inhibited by Knirps and
Giant.

(E) Composite element containing Dorsal, Twist, and Gal 4 activators exhibits repression
of Dorsal and Twist by Knirps, while narrower Gal4 driven expression pattern is
unaffected.

(F) Composite element with Dorsal, Twist, and Gal4 activators, and Giant repressor,
exhibits similar complex expression pattern.

(G) A similar pattern of selective repression of the Dorsal and Twist activators within the
composite element used in 1F is seen when the activator Gal4 is driven throughout the
embryo under the control of the nanos promoter (NGT40, Bloomington Stock #4442). In
the central regions of the embryo more intense staining is visible, indicative of additive or
synergistic gene activation by Dorsal, Twist and Gal4. In the regions of the embryo
where the repressor Giant is expressed (arrows), the intensity of lacZ staining is the same
as in the dorsal regions of the embryo where activation is driven by Gal4 alone.

The difference in lacZ staining intensity between cells containing or lacking Giant or
Knirps is due to a difference in intensity in each cell, not the number of cells stained.
Patterns of gene expression were visualized in 2-4 hour embryos by in situ hybridization
with digU labeled antisense lacZ probes. Embryos are oriented anterior to lefi;
ventrolateral views (A, B, C, D, E and, G) and ventral view (F) are shown.

64

Figure II-l: Simultaneous repression and activation from a compact regulatory
element

 

 

twl or knl twl an gt
C D
~~ .. “we” 4, 1. ,5“; w"
||£ttn I'mz IIE???) l""’z
knl Gal4 gt Gal4
E F

    

 

, ‘ff .. --- lacZ
twl dl knl Gal4 twl dl gt Gal4
G wu-
)5 .. .... \

65

states simultaneously, unlike the binary switch activity observed for many enhancers,

where it appears that a single signal to activate or repress is present.

We make this conclusion based on the activity of the elements when only one set
of activators is present (Figure II-lA-D), and on the characteristic narrower pattern
driven by the Gal4 activators. Consistent with this conclusion, a similar pattern of
exclusive repression of the Dorsal and Twist activators is seen when expression of Gal4
is driven in a ubiquitous pattern using the nanos promoter (Tracey et al., 2000). Here, we
can compare promoter activity with Gal4 alone or in combination with Dorsal/Twist
(Figure II-lG). In dorsal regions of the embryo, the only activator on the element is Gal4,
and no repression by Giant is visible. In the ventral regions where Dorsal and Twist are
present, but the repressor is absent, more intense staining is seen, consistent with
synergistic or additive activation. Irnportantly, in the ventral regions also containing the
Giant repressor (Figure II-IG, arrows), lacZ expression is similar to that observed in the
dorsal regions of the embryo. This result indicates that Dorsal and Twist are not working
together with Gal4, but are functionally independent and selectively repressed in the

regulatory element.

Compact element functions in a distance, orientation, promoter-independent manner

To further evaluate the properties of this element, we tested whether it possessed
classical characteristics of a transcriptional enhancer, namely, acting in a distance- and

orientation-independent manner (Banerji et al., 1981). The element containing Giant

66

binding sites was placed in either orientation between the divergently transcribed white
(at —265 bp) and lacZ genes (at —130 bp). In both orientations tested, this element
directed white expression from —265 bp in a manner closely resembling that seen for the
hsp 70 lacZ reporter; Giant efficiently repressed Dorsal and Twist, while Gal4 activated
transcription in a continuous ventral swathe (Figure II-lB, F; Figure II-2A, B, D, and E).
A similar pattern of repression and activation is seen with the transposase lacZ gene
(Figure II-2C, F). The identical results observed in Figure II-lF and Figure II-ZB, C, E,
and F indicate that the specific patterns of activation and repression are not dependent on

the particular promoter context or orientation of activators and repressors.

See Appendix A

Conversion of enhancer output to a binary on/ofl' state

The compact regulatory element assayed in Figure II-l and Figure II-2 fits the
classical definition of an enhancer, functioning in a distance- and orientation-independent
manner. In addition, the size of this element resembles that of naturally occurring
enhancers ~ 200-800 bp in length. However, the element does not function in the
biphasic either “on or off’ mode normally thought to be a characteristic feature of
enhancers. We are unaware of documented cases where a single enhancer displays two
different states at the same time and in the same nucleus, thus this dual activity appears to
be unusual. It is possible that rather than being an inherent functional property of
enhancers, the uniform output of enhancers might reflect evolutionary pressure to arrange

repressor and activator binding sites to optimize a consistent output. To simulate this

67

Figure II-2: Compact regulatory element displays enhancer-like properties of
distance and orientation independence.

(A, D) The regulatory element shown in Figure II-lF was inserted in either orientation
into a vector containing divergently transcribed white and transposase lacZ reporter
genes. When situated at ——265 bp, Dorsal/Twist activators within the element drive
expression of the white reporter gene. Repression by Giant is evident in anterior and
posterior regions (arrows).

(B, E) In the presence of Dorsal, Twist, and Gal4 activators, a composite pattern of gene
regulation is seen as in Fig. 1F with inhibition of Dorsal/Twist and activation by Gal4.

(C, F) A similar expression pattern is observed with the divergently transcribed
transposase lacZ promoter, with repression by Giant of Dorsal/Twist and activation by
Gal4.

Embryos are oriented anterior to left; lateral views (A, D) and ventrolateral views (B, C,
E, and F) are shown.

68

Figure II-2: Compact regulatory element displays enhancer-like properties of
distance and orientation independence.

 

 

 

A B C

M ..

221 N “1°" ----- 5|: :2! .... film”; W;
D E F

4...... «7...:

MT“: Mtg; gflfiiuhwt

69

situation, two additional Giant repressor binding sites were introduced into this element
3’ of the Gal4 binding sites. Now, complete loss of staining is evident in nuclei
containing the Giant protein (arrows) yielding a classical biphasic “on or off” state

(Figure II-3A, B, C, and D).

DISCUSSION

Redundancy in enhancer function

If an enhancer were an indivisible unit of transcriptional regulation, the functional
independence of adjacent binding sites within the composite element (Figure II-lE, F)
would suggest that this compact element is in fact two separate enhancers. However, this
element is of similar size to natural enhancers and does conform to the classical definition
of an enhancer, namely a compact element that functions to regulate transcription in a

position and orientation independent manner (Banerji et al., 1981).

Functional analyses of cis-regulatory regions provide evidence for redundancy
and hence divisibility, of natural enhancers, suggesting that they can also contain
multiple, independently acting subelements. In the viral setting, the well-studied SV40
enhancer comprises two independently acting subelements that can be separately assessed
(Herr and Clarke, 1986). In Drosophila, recent evidence suggests that eve enhancers
possess redundant activities. Deletion of the entire 480 bp eve stripe 2 element within the

eve locus fails to completely abrogate stripe 2 expression (M. Kreitman, personal

70

Figure II-3: Conversion of a multiple state element to a binary on/off switch.

Two additional Giant binding sites were introduced at the 3’ end of the Gal4 activator
cluster.

(A, B) As observed previously, Dorsal/Twist activators are repressed in anterior and
posterior regions of Giant expression (arrows).

(C, D) In the presence of Dorsal/Twist and Gal4 activators, complete repression of
transcription is observed in areas of Giant expression (arrows).

Embryos are oriented anterior to the left. Lateral (A, C) and ventral (B, D) views are
shown.

71

Figure lI-3: Conversion of a multiple state element to a binary on/ofl’ switch.

 

A B
_____ lacZ
twl dl gt gt
C D
/" '\
. e - 4- lacZ

twldl gt Gan gt

72

communication) indicating the presence of redundant regulatory sequences in the locus.
Furthermore, tissue-specific expression of the yolk protein genes ypI and yp2, is
supported by flanking sequences after deletion of the 125 bp yolk protein enhancer
(Piano et al., 1999). The resilience of natural enhancers to loss of single binding sites
further supports the notion that these elements are built of redundantly acting sequences

(Amosti, 2003).

Selection for uniformity of enhancer output

A scenario of an enhancer with simultaneously displayed activation and
repression states is reminiscent of the modular, autonomous pair rule stripe enhancers,
such as even-skipped stripe elements, where separate enhancers represent different
“states” of repression and activation in the same nucleus (Gray and Levine, 1996). An
important distinction is that our findings suggest that a similar discrimination is taking
place within the tight confines of a single enhancer, and that in order to establish a
uniform signal output, enhancers require a proper stoichiometry or distribution of
repressor and activator binding sites to ensure that all possible enhancer subelements
provide the same information (Figure II-4). Indeed a distributed pattern of short-range
transcriptional repressor binding sites is typical of many developmental enhancers that
function in the early Drosophila embryo; this configuration would allow repressors to
block multiple modes of enhancer-promoter interactions (La Rosee et al., 1997; Small et
al., 1992; Small et al., 1996). In this study we actually measure the simultaneous
independent activity of sub-elements (Figure II-l, Figure II-2) and show that they can be

deployed to give a unitary response (Figure II-3) as is seen with natural enhancers. Thus,

73

the carefully designed internal organization of cis-regulatory modules can provide
uniform information that closely simulates an integrative information processing

capacity.

In contrast to the precision of the enhanceosome, a more flexible arrangement of
regulatory proteins has been suggested to be the predominant pattern for elements that
provide diverse patterns in developing systems (Struhl, 2001). Evolutionary and
experimental studies of the eve stripe 2 enhancer suggest that this element can tolerate
and has undergone considerable rearrangement, with great flexibility in the ntunber and
arrangement of individual sites (Amosti et al., 1996a; Ludwig et al., 2000; Ludwig and
Kreitman, 1995; Ludwig et al., 1998). For example, the recent acquisition of a strong
Bicoid activator site appears to have been counterbalanced by the closer apposition of a
nearby Giant binding site (Hewitt et al., 1999; Ludwig et al., 1998). The plasticity of this
enhancer suggests that much variation in spatial placement of individual transcription
factors is possible, consistent with a model in which these factors contact the basal

machinery in a flexible framework, not necessarily as a rigid complex.

With such flexibility, the transcription factors of an enhancer might still engage
the transcriptional machinery in simultaneous cooperative interactions, as is suggested
with enhanceosomes. However, our studies suggest that an individual enhancer is
capable of representing both the state of activation and repression, suggesting that the
basal machinery may “sample” discrete regions, comprising a small number of
transcription factor binding sites, within the enhancer (Figure II-4B, C). Successive

interactions with the basal machinery, and the biochemical consequence of these multiple

74

Figure II-4: Enhanceosome versus Information Display enhancer models.

(A) In the enhanceosome model, the enhancer serves as an information processing center,
receiving inputs from multiple transcription factors that bind it. A highly structured
complex or enhanceosome, creates a stereospecific interface for docking with and
recruiting the basal transcription machinery. Here the enhancer serves as a molecular
computer, resolves multiple inputs and provides a single output to the basal transcription
machinery. With such an enhancer, the target gene would be activated only upon the
assembly of a complex, providing a precise on/off binary transcriptional switch in
response to the appropriate stimulus. Graded responses from such an element could be
achieved by varying the stability of the entire complex, possibly in response to activator
concentrations.

(B, C) Information Display or “Billboard” enhancer. Rather than acting as a central
processing unit, subelements can display contrasting information, which is then
interpreted by basal transcription machinery. In this model, the basal machinery
“samples” discrete regions of the enhancer each comprising of a small number of
transcription factor binding sites, either iteratively (B) or simultaneously (C).
Successive/multiple interactions with the basal machinery, and the biochemical
consequence of these interactions, would dictate the overall output of the enhancer.

75

Figure II-4: Enhanceosome versus Information Display enhancer models.

 

  

0N r OFF 136,

 

NET OUTPUT

 

 

 

 

PARTIALLY
ON

 

 

 

 

ON PARTIALLY ON

 

 

 

76

interactions would dictate the overall output of the enhancer (Figure II-4B).
Alternatively the enhancer may engage in multiple, simultaneous contacts with some or
all of the enhancer bound proteins, with repressors such as Giant and Knirps preventing
some of these interactions (Figure II-4C). In either case, multiple iterative sampling of
the enhancer, or simultaneous readout, the enhancer would function as an information

display element with computation at the level of enhancer-promoter interactions.

Our results suggest that a closer examination of enhancer classifications is
warranted. The terms enhancer and enhanceosome are frequently used interchangeably to
denote a complex of DNA bound regulatory proteins, yet there appear to be important
functional distinctions between enhanceosomes, as typified by the IFN-B enhancer, and
other regulatory elements. In the light of the functional differences outlined above, a
distinction should be made between the terms enhanceosome, which requires the
cooperative assembly of a higher order structure within an enhancer, and other cis-
regulatory elements that may or may not function in this manner. We propose a model,
the information display or “billboard” model for enhancer action, in which an enhancer,
rather than acting as a central processing unit, can display contrasting information, which
is then interpreted by basal transcriptional machinery (Figure II-4B, C). The binary “on
or off’ decisions that appear to be transmitted by the enhancer to the basal machinery
actually result fiom the basal machinery reading a series of redundant signals encoded
within the enhancer. The model does not explicitly describe the molecular mechanisms
of repression and activation, but direct contacts between the Drosophila activators used
here and components of the basal machinery are supported by biochemical studies (Koh

et al., 1998; Pham et al., 1999; Yuh et al., 1998; Zhou et al., 1998)

77

The billboard enhancer model appears to more accurately describe many
developmentally regulated enhancers, whose internal architecture is subject to rapid
evolutionary change, even as the overall output remains constant (Ludwig et al., 2000;
Ludwig et al., 1998). Although studies such as those on the [FN-B gene indicate that cells
may commonly use enhanceosomes to achieve regulatory precision in gene expression, it
is likely that eukaryotic organisms use the “billboard” type of enhancers to achieve

diversity in gene expression patterns and evolutionary flexibility.

ACKNOWLEDGEMENTS

We thank R. W. Henry, 8.]. Triezenberg, L. Kroos, and three anonymous reviewers for
helpful suggestions on the manuscript, and E. Femandez-Villatoro for technical
assistance. This work was supported by grant GM56976 from the National Institutes of

Health to DNA.

78

REFERENCES

Andrioli, L. P., Vasisht, V., Theodosopoulou, E., Oberstein, A. and Small, S. (2002).
Anterior repression of a Drosophila stripe enhancer requires three position-specific
mechanisms. Development 129, 4931-40.

Arnosti, D. N., Barolo, S., Levine, M. and Small, S. (1996a). The eve stripe 2 enhancer
employs multiple modes of transcriptional synergy. Development 122, 205-14.

Arnosti, D. N., Gray, S., Barolo, S., Zhou, J. and Levine, M. (1996b). The gap protein
knirps mediates both quenching and direct repression in the Drosophila embryo. Embo J
15, 3659-66.

Arnosti, D. N. (2003). Analysis and function of transcriptional regulatory elements:
Insights from Drosophila. Annu Rev Entomol 48, 579-602.

Banerji, J., Rusconi, S. and Schaffner, W. (1981). Expression of a beta-globin gene is
enhanced by remote SV40 DNA sequences. Cell 27, 299-308.

Barolo, S. and Posakony, J. W. (2002). Three habits of highly effective signaling
pathways: principles of transcriptional control by developmental cell signaling. Genes
Devl6, 1167-81.

Biggar, S. R. and Crabtree, G. R. (2001). Cell signaling can direct either binary or
graded transcriptional responses. Embo J 20, 3167-76.

Brand, A. H. and Perrimon, N. (1993). Targeted gene expression as a means of altering
cell fates and generating dominant phenotypes. Development 118, 401-15.

Carey, M., Lin, Y. S., Green, M. R. and Ptashne, M. (1990). A mechanism for
synergistic activation of a mammalian gene by GAL4 derivatives. Nature 345, 361-4.

Chi, T., Lieberman, P., Ellwood, K. and Carey, M. (1995). A general mechanism for
transcriptional synergy by eukaryotic activators. Nature 377 , 254-7.

Davidson, E. H. (2001). Genomic Regulatory Systems: Development and Evolution:
Academic Press: A Harcourt Science and Technology Company.

Fujioka, M., Emi-Sarker, Y., Yusibova, G. L., Goto, T. and Jaynes, J. B. (1999).
Analysis of an even-skipped rescue transgene reveals both composite and discrete
neuronal and early blastoderm enhancers, and multi-stripe positioning by gap gene
repressor gradients. Development 126, 2527-38.

79

Ghazi, A. and VijayRaghavan, K. V. (2000). Developmental biology. Control by
combinatorial codes. Nature 408, 419-20.

Gray, S. and Levine, M. (1996). Transcriptional repression in development. Curr Opin
Cell Biol 8, 358-64.

Gray, S., Szymanski, P. and Levine, M. (1994). Short-range repression permits
multiple enhancers to function autonomously within a complex promoter. Genes Dev 8,
1829-38.

Herr, W. and Clarke, J. (1986). The SV40 enhancer is composed of multiple functional
elements that can compensate for one another. Cell 45, 461-70.

Hewitt, G. F., Strunk, B. S., Margulies, C., Priputin, T., Wang, X. D., Amey, R.,
Pabst, B. A., Kosman, D., Reinitz, J. and Arnosti, D. N. (1999). Transcriptional
repression by the Drosophila giant protein: cis element positioning provides an
alternative means of interpreting an effector gradient. Development 126, 1201-10.

Keller, S. A., Mao, Y., Struffi, P., Margulies, C., Yurk, C. E., Anderson, A. R.,
Amey, R. L., Moore, S., Ebels, J. M., Foley, K. et al. (2000). dCtBP-dependent and -
independent repression activities of the Drosophila Knirps protein. Mol Cell Biol 20,
7247-58.

Kim, T. K. and Maniatis, T. (1997). The mechanism of transcriptional synergy of an in
vitro assembled interferon-beta enhanceosome. Mol Cell 1, 119-29.

Koh, S. S., Ansari, A. Z., Ptashne, M. and Young, R. A. (1998). An activator target in
the RNA polymerase 11 holoenzyme. Mol Cell 1, 895-904.

Kosman, D. and Small, S. (1997). Concentration-dependent patterning by an ectopic
expression domain of the Drosophila gap gene knirps. Development 124, 1343-54.

La Rosee, A., Hader, T., Taubert, H., Rivera-Pomar, R. and Jackle, H. (1997).
Mechanism and Bicoid-dependent control of hairy stripe 7 expression in the posterior
region of the Drosophila embryo. Embo J 16, 4403-11.

Ludwig, M. Z. and Kreitman, M. (1995). Evolutionary dynamics of the enhancer
region of even-skipped in Drosophila. Mol Biol Evol 12, 1002-11.

Ludwig, M. Z., Patel, N. H. and Kreitman, M. (1998). Functional analysis of eve stripe
2 enhancer evolution in Drosophila: rules governing conservation and change.

Development 125, 949-58.

Ludwig, M. Z., Bergman, C., Patel, N. H. and Kreitman, M. (2000). Evidence for
stabilizing selection in a eukaryotic enhancer element. Nature 403, 564-7.

80

Munshi, N., Agalioti, T., Lomvardas, S., Merika, M., Chen, G. and Thanos, D.
(2001). Coordination of a transcriptional switch by HMGI(Y) acetylation. Science 293,
1133-6.

Pham, A. D., Muller, S. and Sauer, F. (1999). Mesodenn-determining transcription in
Drosophila is alleviated by mutations in TAF(II)60 and TAF(II)110. Mech Dev 84, 3-16.

Piano, F., Parisi, M. J., Karess, R. and Kamhysellis, M. P. (1999). Evidence for
redundancy but not trans factor-cis element coevolution in the regulation of Drosophila
Yp genes. Genetics 152, 605-16.

Rossi, F. M., Kringstein, A. M., Spicher, A., Guicherit, O. M. and Blau, H. M.
(2000). Transcriptional control: rheostat converted to on/off switch. Mol Cell 6, 723-8.

Seipel, K., Georgiev, O. and Schaffner, W. (1992). Different activation domains
stimulate transcription fi'om remote ('enhancer') and proximal ('promoter') positions.
EmboJll, 4961-8.

Small, S., Blair, A. and Levine, M. (1992). Regulation of even-skipped stripe 2 in the
Drosophila embryo. Embo J 11, 4047-57.

Small, S., Arnosti, D. N. and Levine, M. (1993). Spacing ensures autonomous
expression of different stripe enhancers in the even-skipped promoter. Development 119,
762-72.

Small, S., Blair, A. and Levine, M. (1996). Regulation of two pair-rule stripes by a
single enhancer in the Drosophila embryo. Dev Biol 175, 314-24.

Struhl, K. (2001). Gene regulation. A paradigm for precision. Science 293, 1054-5.

Szymanski, P. and Levine, M. (1995). Multiple modes of dorsal-bHLH transcriptional
synergy in the Drosophila embryo. Embo J 14, 2229-38.

Thanos, D. and Maniatis, T. (1995). Virus induction of human IFN beta gene
expression requires the assembly of an enhanceosome. Cell 83, 1091-100.

Tracey, W.D., Ning, X., Klingler, M., Kramer, S.G., and Gergen, J.P. (2000)
Quantitative analysis of gene function in the Drosophila embryo. Genetics 2000 154,
273--284.

Yuh, C. H., Bolouri, H. and Davidson, E. H. (1998). Genomic cis-regulatory logic:
experimental and computational analysis of a sea urchin gene. Science 279, 1896-902.

81

Zhou, J., Zwicker, J., Szymanski, P., Levine, M. and Tjian, R. (1998). TAFII
mutations disrupt Dorsal activation in the Drosophila embryo. Proc Natl Acad Sci U S A
95, 13483-8.

82

Chapter III1

Operating Principles of Short-range Transcriptional Repressors

INTRODUCTION

The identification and characterization of functionally significant noncoding
regions, especially those that control transcription, is one of the major challenges in
understanding the regulatory language encoded in genome sequences. Although we now
have genome sequences for many metazoans, our understanding of how this regulatory

information is encoded is extremely limited.

The cis-regulatory information that orchestrates complex spatial and
temporal patterns of gene expression in development of higher eukaryotes is
typically organized into modular units termed enhancers or cis-regulatory modules,
of a few hundred base pairs in size. A common feature of these modules is the
presence of multiple binding sites for several distinct transcription factors including
sequence-specific activators and repressors (Davidson, 2001; Guss et al., 2001;
Carroll, SB, 2001). The transcription factors in turn interact with each other, with
cofactors, chromatin and with the basal transcriptional machinery to either activate
or repress the corresponding target gene. Because a cis element can contain multiple
copies of the same or different binding motifs, a limited number of transcription
factors can be arranged in numerous possible combinations that can result in distinct
transcriptional outputs. Such combinatorial action by transcription factors can

confer temporal and spatial specificity (Struhl, 1991).

 

' The data in Chapter 3 is presented in the form of a manuscript to be submitted soon.

83

Identification of a particular regulatory module within a promoter region ofien
reveals little about its role in the expression of a given gene. Therefore empirical tests
have been extensively employed to decipher how various regulatory elements within a
promoter work together to modulate transcription. In such assays a normal or modified
region of a promoter is fused to a reporter gene and introduced into a cell or whole
organism, where it can be exposed to the shifting array of transcription factors that
normally modulate the expression of the endogenous gene. The resulting pattern of
reporter gene expression can reveal, for instance, whether a particular regulatory element
can activate or repress transcription at a specific time and place. Because of the
complexity of most regulatory regions, multiple experiments of this kind are needed to
gain even a rough overview of how an expression pattern is generated (Amosti et al.,
1996a; Goto et al., 1989; Harding et al., 1989; Small et al., 1993; Small et al., 1996;
Small et al., 1991; Yuh et al., 1998). As a result relatively few enhancers in all animal
systems combined have been well characterized, and our understanding of general
principles dictating interactions of cis-regulatory elements with each other and with the

basal promoter remain elusive.

Recently, whole genome sequence assemblies have become available providing a
powerful foundation to identify and analyze cis-regulatory modules function and
organization on a global scale. Two current approaches to identify candidate regulatory
regions from genomic data are computational methods that look for clusters of
transcription factor binding sites (Berman et al., 2002; Markstein and Levine, 2002;

Markstein et al., 2002; Rebeiz et al., 2002) and phylogenetic comparisons that identify

84

evolutionarily conserved sequences (Bergman and Kreitman, 2001; Maier et al., 1990).
One main difficulty with the output from computational searches for clusters of
transcription factor sites is the large number of false-positive results. The short length and
degenerate nature of transcription-factor-binding sites account for most of these
misleading predictions. Furthermore, genes are rarely controlled by a single transcription
factor and accumulating evidence suggests that specific combinations of transcription
factors are required to achieve the complex differential expression of genes in higher
organisms (Davidson, 2001; Gray et al., 1994; Hewitt et al., 1999; Keller et al., 2000;
Struhl, 2001; Yuh et al., 1998). The structural basis for combinatorial regulation exists in
the specific organization of multiple transcription factor binding sites (Struhl, 2001). In
cases where transcription factors bind cooperatively to form a higher order nucleoprotein
complex that activates gene expression, spatial correlations or constraints between the
binding sites have been observed (Kim and Maniatis, 1997; Merika and Thanos, 2001;
Munshi et al., 2001; Thanos and Maniatis, 1995). Even in cases where a higher order
structure is not required to regulate gene expression, spacing between binding sites with
cis-elements can be of utmost important. For example, with short-range transcriptional
repressors, spacing between the repressor and activator sites within the cis-regulatory
element is critical in dictating repression effectiveness (Amosti et al., 1996b; Gray et al.,
1994; Hewitt et al., 1999; Keller et al., 2000). Thus, in many cases where cis-regulatory
modules are predicted by computational methods to exist, something in the arrangement
of sites (or a wider context) renders the cis-element non-functional. In order to achieve
better predictions and eliminate false-positive and false—negative results, computational

methods should include, in addition to binding site density and relative affinities, other

85

parameters for spacing and position of binding motifs within cis-elements in the search

algorithms.

Alternatively, evolutionary conservation can often be used to indicate the
existence of regulatory regions. Interspecific sequence comparisons (phylogenetic
footprinting) of noncoding regions reveal conserved features, many of which are cis-
regulatory elements. However, despite obvious indications of selective constraints
(Hardison et al., 1997; Loots et al., 2000) the structure and sequences of cis-elements
change over time, sometimes dramatically even in cases where expression patterns are
conserved (Ludwig et al., 2000; Ludwig and Kreitman, 1995; Piano et al., 1999). Thus,
phylogenetic comparisons that identify cis-regulatory elements would be greatly
facilitated by empirical determination of spatial constraints between binding sites of cis-

regulatory elements or between different elements themselves.

The early Drosophila embryo has provided a paradigmatic model for studying
transcriptional control of development. Most of the important factors have been identified
by exhaustive genetic analysis (Nusslein-Volhard et al., 1985; Nusslein-Volhard and
Wieschaus, 1980), and there are sophisticated tools for characterizing the design and
function of complex cis-regulatory DNA elements. Typical of complex regulatory
systems in higher eukaryotes, the cis-regulatory elements of key patterning genes employ
extensive sequences to translate broad patterns of maternal and early embryonic factors
into precisely defined segmental distributions of transcription factors (Amosti, 2002;

Driever and Nusslein-Volhard, 1988; Driever et al., 1989; Nusslein-Volhard et al., 1987;

86

Pankratz and Jackle, 1990; Rivera-Pomar and Jackle, 1996). One of the best-studied
complex loci in Drosophila is the pair-rule gene even-skipped, which encodes a
homeodomain transcription factor that is expressed in a series of seven transverse stripes
across the anterior-posterior length of the blastoderm embryo. This gene has several cis-
regulatory elements scattered within a 16 Kb region around the eve coding sequence. Five
of these regions, the stripe enhancer elements, are responsible for early expression of eve
in the form of seven regularly spaced stripes in the blastoderm embryo (Fujioka et al.,
1999; Goto et al., 1989; Harding et al., 1989; Sackerson et al., 1999; Small et al., 1992;
Small et al., 1996; Small et al., 1991; Small and Levine, 1991; Stanojevic et al., 1991).
One key feature of these enhancers is that their action is autonomous; the repression of
one element does not lead to the general repression of the entire locus (Amosti et al.,
1996b; Gray et al., 1994; Small et al., 1993). This autonomy is based on the properties of
the short-range transcriptional repressors that regulate these enhancers and include the
products of the gap genes, giant, knirps and kriippel. These transcriptional repressors
block the activity of enhancer elements when bound within ~ 100 bp of key activator
sites or of basal promoter elements when cognate sites are introduced close to the start of
transcription. The short-range of the repression activity provides a highly flexible, yet
precisely tunable mechanism for specific gene regulation (Amosti et al., 1996a; Amosti
et al., 1996b; Gray and Levine, 1996; Gray et al., 1994; Hewitt et al., 1999; Keller et al.,
2000; Strunk et al., 2001). This flexibility contrasts with long-range repressors, typified
by the Drosophila Hairy protein, which can block multiple enhancers over distances of
several kilobases, regardless of location within a gene complex. The different activities of

short- and long-range repressors probably reflect distinct mechanisms employed by

87

transcriptional repressors. The mechanisms by which short-range and long-range
Drosophila repressors inhibit transcription are poorly understood, although one model of
repression in the embryo suggests that the short-range/long-range distinction results from
the recruitment of distinct cofactors (Nibu et al., 1998a; Nibu et al., 1998b; Zhang and
Levine, 1999). Short-range repressors recruit the corepressor CtBP to mediate repression,
whereas long-range repressors have been shown to interact with the Groucho corepressor
(Barolo and Levine, 1997; Chen and Courey, 2000; Fisher and Candy, 1998; Mannervik
and Levine, 1999; Mannervik et al., 1999; Poortinga et al., 1998). One model for
Groucho mediated long-range repression is through the recruitment of HDACs by
Groucho, resulting in the production of a large transcriptionally silent chromosomal
domain. Just as the Sir repressosome generates a transcriptionally silent chromatin
structure that is able to spread along the chromatin fiber, it has been proposed that the
Groucho protein nucleates a silenced chromosomal state that spreads to mediate long-
range repression (Brantjes et al., 2001; Chen et al., 1999; Flores-Saaib and Courey,

2000)

Another striking feature of the eve stripe enhancers is that they contain multiple
binding sites for both activators and transcriptional repressors. The best characterized eve
enhancer controls the expression of eve stripe 2 and contains a total of eighteen known
factor-binding sites, including eight activator (three for Hunchback, five for Bicoid)
binding sites and ten repressor (three for Giant, six for Kruppel, and at least one for
Sloppy-paired) binding sites (Andrioli et al., 2002; Ludwig et al., 1998; Small et al.,

1991; Small and Levine, 1991; Stanojevic et al., 1991). It has been proposed that a high

88

local density of transcription factor binding sites may be sufficient for the proper function
of these cis-elements and can be used as a convenient signpost for computational
identification novel cis-elements. However, several attempts to construct artificial stripe 2
enhancers by multimerizing Bicoid, Hunchback activator with Kruppel and Giant
repressor sites have been unsuccessful (S. Small and M. Levine, unpublished). These
results suggest that cis-regulatory sequences other than the known sites are important for
activation. Such sequences may contain sites for other activator proteins. Alternatively,
they may simply function to provide the correct spacing between known activator or
repressor sites. Thus, the grammar of the cis-regulatory code is clearly more complex
than simply the density of transcription factor binding sites. The relative affinity, spacing,
and positioning of transcriptional activator and repressor binding sites within cis-
regulatory modules has been demonstrated to be significant in many cases (Amosti et al.,
1996a; Courey and Huang, 1995; Hanes et al., 1994; Hewitt et al., 1999; Lifanov et al.,
2003; Szymanski and Levine, 1995). Previous analyses of short-range repressors on
native enhancers had demonstrated that these proteins can block gene expression when
within ~ 100 bp of the target —either activators within an enhancer element or the core
promoter (Amosti et al., 1996b; Hewitt et al., 1999; Keller et al., 2000). However, since
such studies have focused on the activity of short-range repressors in the context of
complex, endogenous regulatory elements, the distance requirements have not been

exactly defined as the positions of the nearest activators within the element is not known.

Using synthetic enhancer elements where the identity, stoichiometry and the exact

arrangement of activator and repressor binding site are well-defined, we demonstrate that

89

the previously held simple notion that short-range repressors block the activity of all
protein complexes within a 100 bp is incorrect. The manipulation of these composite
elements in terms of the number of activator and repressor binding sites, relative
affinities, spacing and distribution of these binding sites further allowed us to define the
contextual parameters that dictate repression effectiveness and include architectural
features such as stoichiometry of activators and repressors, relative affinity, spacing and
position of binding sites. These elements constitute the ‘grammar’ of short-range
repressor function, and the empirical identification of such rules for different classes of
transcriptional regulators will facilitate computational searches for cis-elements regulated

by them.

MATERIALS AND METHODS

l. Gal4- activator chimeric constructs

GAL4 (aal-93) - GAL4 AD (aa753-881)

A KpnI-Xbal fragment from pSCTEV GAL4 (1-93)- GAL4 (Seipel et al., 1992)
containing the reading frame for the yeast GAL4 activation domain (Gal4 AD) from
amino acid residues 753-881, was cloned into KpnI-Xbal cut pTwiggy (Amosti et al.,
1996a) vector, which contains the twist enhancer (2xPEe-Et) element, twist basal

promoter and the GAL4 DNA-binding domain from residues 1-93.

90

Insulated GAL4 (aa1-93) - GAL4 AD (aa753-881)

The following primers with EcoRI ends were used to PCR amplify a 420 bp
fragment of DNA containing the gypsy insulator from the Green Pelican GFP
transformation vector (Barolo et al., 2000).

DA639: 5’ CGG AAT TCC GAA TTG TAA GCG TTA ATG ACT 3’
DA640: 5’ CGG AAT TCC GAT ACA TAC TAG AAT TGA TCG 3’
The fragment containing the gypsy insulator was cloned into the pTwiggy transformation
vector at the EcoRI site between the twist regulatory elements and the white gene. The
vector was further modified to contain the reading fiame for the yeast GAL4 activation

domain (Gal4 AD) from amino acid residues 753-881 as a KpnI-Xbal fragment.

GAL4 (aal-93) — VP16 (aa4l2-490)

The C-terminal transcriptional activation domain of the herpes simplex virus
VP16 protein from residues 412-490 was amplified from pRevTet ofir (Ryu et al., 2001)
which contains the bacterial Tet repressor DNA binding domain fused to the VP16
activation domain using the following primers-
DA410: 5’ GGG TCG GTA CCG CAA CGG CCC CCC CGA CCG ATG TC 3’
DA411: 5’ GGG GAA TCT AGA CTA ACT AAT TAC TAC CCA CCG TAC TCG
TCA AT 3’
The PCR product was digested with KpnI and XbaI enzymes and cloned into KpnI- XbaI

cut pTwiggy vector (Amosti et al., 1996a).

91

GAL4 (aal-93) - Spl (aa132-243)

A KpnI — Xbal fragment from pSCTEV Gal4 (1-93)-Sp1: Q1 (Seipel et al., 1992)
containing the open reading frame for the activation domain (residues 132-243) of the
transcription factor Spl was cloned into KpnI — Xbal cut pTwiggy vector (Amosti et al.,

1996a)

GAL4 (aa1-93) - hTBP (aal-339)

The following oligos with KpnI — Xbal ends were used to amplify full-length
hTBP (aa1-339) a gift from J. Geiger.
DA162: 5’ GGG TCG GTA CCG CAG CCG CAA TGG ATC AGA ACA ACA GCC
TG 3’
DA164: S’GGG GAA TCT AGA CTA ACT AAT TAC TAC GTC GTC TTC CTG AAT
CCC TT 3’
The PCR product was digested with KpnI and Xbal enzymes and cloned into Kpnl- Xbal

cut pTwiggy vector (Amosti et al., 1996a).

2. Fly stocks

Flies carrying a mutation in the giant gene on the X chromosome gtAa/FM7c
(Stock # 1004.1) and gt’m/FM6 (Stock # 1529) were obtained from the Bloomington
Stock center. To analyze reporter gene expression in a giant mutant background, males
carrying the reporter and the Gal4-activator transgenes were crossed to females carrying

the giant mutation.

92

Flies expressing the full-length yeast transcriptional activator Gal4 ubiquitously
throughout the embryo under the control of the actin5C enhancer, act5cGAL4/Cy0
(Stock # 4414) were also obtained from Bloomington. In order to obtain ubiquitous
activation of the lacZ reporter gene in the early (2-4 hour old) embryo, act5cGAL4/Cy0

females were crossed to males carrying the reporter transgene.

3. Reporter genes

The stripe 2/2x UAS/eve-lacZ vector (Amosti et al., 1996a) containing two GAL4
binding sites and the minimal eve basal promoter driving lacZ expression was modified
to include two Giant (Capovilla et al., 1992) binding sites (DA127/ 128: 5’AAT TCG
CAT GCT ATG ACG CAA GAA GAC CCA GAT CTT TTT ATG ACG CAA GAG
CAT GCG 3’) using EcoRI-BssH2 enzymes, upstream of the GAL4 sites. The vector was
further modified to incorporate three additional GAL4 binding sites (DA139/140: 5’
TCG GAT TAG AAG CCG CCG TCG CTA GAG GAA GAC TCT CCT CCG ACG
TGA ACG CAG GAC ACT CCT GC GCT GCA 3’) at the PstI site downstream of the
existing GAL4 sites. Oligos containing a 50 bp spacer (DA125: 5’TCG CTA GAC GTG
AAT CTC GTA GCT TCC GTA CCA AAT GCG TAT CAG CTG CA 3’; DA126: 5’
GCT GAT ACG CAT TTG GTA CGG AAG CTA CGA GAT TCA CGT CTA GCG
ATG CA 3’) were introduced at the PstI site downstream resulting in the vector H2g5u-
50 (Figure III-1E, F, G, and H) which contains two Giant binding sites, five tandemly
arrayed GAL4 binding sites, a 50 bp spacer and minimal eve basal promoter driving lacZ

expression.

93

The plasmid UAS-lacZ (Brand and Perrimon, 1993) was modified to contain two
Giant sites (DA321/322: 5’ GGC CGC TAT GAC GCA AGA AGA CCC AGA TCT
TTT TAT GAC GCA AGA GA 3’) or two Knirps sites (DA319/320: 5’GGC CGC
ATC TGA TCT AGT TTG TAC TAG ACA TCT GAT CTA GTT TCA 3’) or two
Kruppel (DA694/695: 5’GGC CGC AAA ACG GGT TAA GCG ACC CAA AAC
GGG TTA AGC A 3’) sites or two Hairy (DA604/605: 5’GGC CGC GCG GCA CGC
GAC ATG ACC CGC GGC ACG CGA CAT A 3’) sites twenty nucleotides upstream
of the five GAL4 binding sites (Amosti et al., 1996c; Capovilla et al., 1992; Gray et al.,
1994; Nibu et al., 2001). The resulting vectors named M2g5u-lacZJM2k5u-
lacZ/M2kr5u-lacZ/M2h5u-lacZ respectively, consist of two Giant or Knirps or Krilppel
or Hairy Binding sites, five tandemly arrayed GAL4 binding sites, followed by the hsp 70
TATA box and transcriptional start driving lacZ expression (FigureIII-lA, B; Figure 111-
2A, B; 7B, D, F).

The vector M2g5u-lacZ was modified by introducing oligos containing a 55 bp
neutral spacer (DA65/66: 5’TCC ATG ATA AAC GCG TGC TAG ACT ATT GCA
GGT ACT GAT CGA ATG CCT CTG CAT G 3’) at the Sphl site downstream of the
Gal4 binding sites. The vector was further modified by introducing a 340 bp fragment of
the Knirps open reading frame, which was amplified using oligos DA572/573 (DA 572:
5’ACA TGC ATG CAA CCG CTT TAG TCC CGC CAG 3’; DA 573: 5’ACA TGC
ATG CTG TGC ACG GAG CTC CGC GAG 3’) from the vector Gal4-km F1 resulting in
the spaced construct M2g5u-55-340 bp kni ORF-lacZ (Figure III-1C, D).

The vector M2g5u-lacZ was modified to replace the five tandemly arrayed

GAL4 sites with HindIII-Sphl oligos containing three high affinity GAL4 (Brand and

94

Perrimon, 1993) binding sites (DA469/470: 5’AGC TTG CCT GCA GGT CGG AGT
ACT GTC CTC CGA GCG GAG TAC TGT CCT CCG AGC GGA GTA CTG
TCC TCC GAG GCA TG 3’) to give M2g3u-lacZ (Fig. 2C, D). This was further
modified by introducing SphI spacer oligos (DA471: 5’ TCA TAC AAC TGG TCA GTG
AGC ATA CAA CTG GTC AGT GAG CAT G 3’; DA472: 5’ CTC ACT GAC CAG
TTG TAT GCT CAC TGA CCA GTT GTA TGA CAT G 3’) equal to the length of two
GAL4 sites downstream resulting in M2g3u2x-lacZ (Figure III-2E, F). The two Giant
binding sites in M2g3u2x-lacZ were replaced by two Knirps sites (DA319/320: S’GGC
CGC ATC TGA TCT AGT TTG TAC TAG ACA TCT GAT CTA GTT TCA 3’) or
two Kriippel (DA694/695: S’GGC CGC AAA ACG GGT TAA GCG ACC CAA AAC
GGG TTA AGC A 3’) sites or two Hairy (DA604/605: 5’GGC CGC GCG GCA CGC
GAC ATG ACC CGC GGC ACG CGA CAT A 3’) sites twenty nucleotides upstream
of the three GAL4 binding sites. The resulting vectors named M2k3u2x-
IacZ/M2kr3u2x-lacZ/M2h3u2x-lacZ (Figure III-7A, C, and E) consist of two
Knirps/Kriippel/Hairy binding sites respectively, three tandemly arrayed GAL4 binding
sites, a spacer, followed by the hsp 70 TATA box and transcriptional start driving lacZ
expression.

The vector M2g3u-lacZ was cut with HindIII to introduce spacer oligos (DA473:
5’ AGC TTC ATA CAA CTG GTC AGT GAG CAT ACA ACT GGT CAG TG 3’;
DA474: 5’ AGC TCA CTG ACC AGT TGT ATG CTC ACT GAC CAG TTG TAT GA
3’) equal to the length of two GAL4 sites in between the Giant and three GAL4 sites to
result in M2g2x3u-lacZ (Figure III-6B, D, F). The five high affinity GAL4 binding sites

in M2g5u-lacZ were replaced with five low gffinity (LA) GAL4 sites (Burns and

95

Peterson, 1997; Johnston and Davis, 1984) (site no. 1 from the Gall,10 UAS region) by
sequentially cloning in HindIII-Sphl oligos containing three low affinity sites
(DA600/601: 5’AGC TTG CCT GCA GGT CGG ATT AGA AGC CGC CGA GCG
GAT TAG AAG CCG CCG AGC GGA TTA GAA GCC GCC GCA TG 3’) followed
by SphI oligos containing two low affinity Gal4 sites (DA602/603: 5’ TCG GAT TAG
AAG CCG CCG AGC GGA TTA GAA GCC GCC GCA TC 3’) resulting in the
vector M2g5u (LA)-lacZ (Figure III-3).

Two additional Giant binding sites were introduced either at the SphI site
(DA50/51) in the M2g5u-lacZ vector between the five GAL4 binding sites and the hsp 70
TATA box resulting in M2g5u2g-lacZ (Figure III-5A, B), or at the Nail site
(DA637/638) in the M2g5u—lacZ vector upstream of the two Giant sites resulting in
M4g5u-lacZ (Figure III-5 E, F). Two additional binding sites for Giant (DA50/51) as
well as two high affinity Gal4 sites (DA598/599) were introduced sequentially in the

M2g3u-lacZ vector at the SphI site resulting in M2g3u2g2u-lacZ (Figure III-5 C, D).

4. P-element transformation, crosses to reporter genes, and whole-mount in situ
hybridization of embryos.

P-element transformation vectors were introduced into the Dr030phila gerrnline
by injection of yw67 embryos as described (Small et al., 1992). Embryos were collected
either directly from each transgenic reporter line or from a cross between a reporter line
and a line expressing the GAL4-activator chimeric proteins in the ventral regions or

ubiquitously throughout the embryo. The embryos were fixed and stained using

96

digoxigenin-UTP labeled antisense RNA probes to either lacZ or w as described (Small et

al., 1992).

RESULTS
Context dependence of short-range repression

The activity of short—range transcriptional repressors has been studied mostly in
the context of complex endogenous enhancers (Amosti et al., 1996a; Amosti et al.,
1996b; Gray et al., 1994; Kosman and Small, 1997; Small et al., 1993; Small et al.,
1992). This approach is complicated by the complexity of many cis-acting regulatory
modules, where stoichiometry, spacing and even the identity of transacting factors are not
always well defined. To analyze cis-acting element function in a setting in which
activator-repressor composition, stoichiometry and spacing can be exactly defined, we
constructed chromosomally integrated, compact regulatory modules containing binding
sites for the endogenous short-range repressor Giant and chimeric Gal4 activators. The
space between repressor and activator sites on these elements is less than 100 bp, a
distance over which short-range repressors have been previously shown to be effective
(Amosti et al., 1996b; Gray et al., 1994; Hewitt et al., 1999; Keller et al., 2000).
Strikingly, Giant was unable to repress the hsp70 lacZ reporter containing five high
affinity Gal4 activator sites from an upstream position, although these proteins bind
within 100 bp of the repressors, revealing a hitherto unknown limitation of short-range
repressors (Figure III-1A, B). The close proximity of the G314 activators to the hsp 70
basal promoter may prevent Giant from mediating repression on this reporter. However,

Giant is still unable to repress even after introduction of a 400 bp spacer that moves the

97

 

activators away from the transcriptional start site (Figure III-1C, D). The inability of
Giant to repress is not due to an inherent resistance of the Gal4 activation domain, for
Giant repressed a similar cluster of five Gal4 binding sites 5’ of the eve basal promoter
(Figure III-1E, F). The repression in anterior and posterior regions is relieved when this
transgene is assayed in giant mutant embryos (Figure III-1G, H) confirming that

repression is mediated by Giant.

These results (Figure III-1) overturn the simple notion that short-range repressors
block the activity of all protein complexes within 100 bp. Clearly, mere proximity is not
the only determinant affecting repression by Giant. We set out to systematically define
other factors that dictate repression effectiveness to uncover a potential ‘grammar’ of

repression.

Differences in activator site affinity or spacing, basal promoters, or repressor
positioning with respect to the transcriptional start site may be the basis for the difference
in repression effectiveness in the transgenes shown in Figure III-1. Activator site affinity
or spacing appears to be the most likely cause of the difference in repression
effectiveness, because both types of basal promoters have been shown to be directly
repressed previously, and the relative spacing of the repressors to +1 should in fact favor
repression in Figure III-1A and B, where the Giant repressor is closer to the start site of
transcription. Weaker or suboptimal spacing of activator sites might decrease the average
number of activators on the promoter, which may in turn favor repression (Figure III-1E,

F).

98

Figure III-1: Context dependence of short-range repression.

A schematic structure of the reporter transgene is shown below the corresponding
embryos. Patterns of gene expression were visualized in 2-4 hour embryos by in situ
hybridization with digU labeled antisense lacZ probes. Embryos are oriented anterior to
lefi; lateral views (A, C, D, E, and G) and ventrolateral views (B, F, and H) are shown.

(A, B) LacZ expression driven by the Gal4- Gal4 AD activator fusion from a cluster of
five high affinity Gal4 binding sites upstream of the hsp 70 basal promoter elements is not
repressed by the short-range repressor Giant.

(C, D) LacZ expression driven by the Gal4- Gal4 AD activator fusion from a cluster of
five high affinity Gal4 binding sites upstream ~ 450 bp upstream of the hsp70 basal
promoter elements is also not repressed by the short-range repressor Giant.

(E, F) Lack of repression by Giant is not due to the inherent resistance of the Gal4
activation domain. The activity of the same Gal4 -activation domain from a similar
cluster of five Gal4 binding sites 5’ of the minimal eve basal promoter is repressed by the
short-range repressor Giant in both the anterior and posterior domains of the embryo
where the repressor is expressed (arrows).

(G, H) The repression of lacZ in the anterior and posterior regions is relieved when the
eve lacZ transgene is assayed in an embryo carrying the gtA8 mutation.

99

ling
Sim

1r [0

i314
ilar
the
ryo

he

Figure III-1: Context dependence of short-range repression.

A B

 

mun“ 1"“

 

 

 

 

91 Gal4 -60
C D
3"
E F
7' ' 1 i " M r K wt
G H
.._. 3i“ . s, a t:
I gr“ 91M
.1 L M42 I’acz
gt Gal4 -92

100

Figure III-7: Repressors can be distinguished by their differential ability to repress
in different enhancer configurations.

A schematic of the reporters used is shown below. The reporters contain two binding
sites for either one of the short-range repressors Giant, Knirps or Kruppel; or the long-
range repressor Hairy and five/three high affinity Gal4 binding sites.

All three of the short-range repressors Giant (A), Knirps (C), and Kruppel (E) and the
long-range repressor Hairy (G) are able to repress the activity of the Gal4-activator in the
context of three Gal4 binding sites.

While the short-range repressors Giant (B), Knirps (D), and Kruppel (F) are unable to
repress in the context of five high affinity Gal4 sites the long-range repressor Hairy (H) is
able to effectively repress.

Embryos are oriented with anterior to left; lateral views (A, B, C, D, and E) and
ventrolateral views (F, G, and H) are shown.

118

Specificity of regulatory ‘grammar ’

The contextual dependencies of repression described above were developed for
the Giant repressor. To determine if similar rules applied to other types of repressors, we
carried out parallel evaluations of the short-range repressors Giant, Knirps, and Kruppel.
To test quantitative similarities or differences between these factors, we created reporters
that would compare repressor activity on genes that represented permissive or non-
pennissive contexts for the Giant protein. All three of these short-range repressors were
unable to inhibit lacZ expression driven by the Gal4 activator from five high affinity Gal4
sites indicating a similar limitation of repression on even proximally bound activators
(Figure III-7B, D, F). The Giant and Kruppel factors exhibited repression activity in the
corresponding regions of the embryo when tested against three Gal4 sites (Figure HI-7A,
C). The Knirps repressor was also active in this context, although in general the levels of
repression appeared to be lower (Figure III-7E). In contrast, the long-range repressor
Hairy was able to mediate repression of transgenes containing three or five high affinity
Gal4 sites (Figure III-7G, H). Interestingly as the embryo aged, repression of Gal4 by
Hairy was attenuated and was completely abolished by the time germ band elongation is
completed (data not shown), indicating that this type of repression, though potent, is also
transient. The similarity in the activity of the short-range repressors Giant, Knirps and
Kruppel, in contrast to that of Hairy, suggests that the contextual rules for repression are

governed by the functional class of repressor, and likely reflect mechanistic differences.

117

Figure III-6: The ability to repress depends on enhancer configuration rather than
the nature of the activation domain.

 

 

 

A B
Gal4-gal4AD 4 ‘ ‘
f uni.- K 7‘ ’ijzrw
Full-length C ' “‘1‘ 7“ D
Gal4 .
. g
f ’ ' K
E F
Gal4-VP16 ' _.;,
.94.; a 7 _ — i; ":4 a: :1“: g" 1’.»
f 3 K '
III-2900- '“2 Imam IE?
9‘ Gal4 '00 Cl Gal4-60

116

Figure III-6: The ability to repress depends on enhancer configuration rather than
the nature of the activation domain.

A schematic representation of the reporter genes are shown above and the chimeric Gal4-
activators used to drive expression from the reporter gene are indicated to the left of the

embryos.

(A, C, E) Giant is able to repress all three activators when the number of Gal4 sites is
reduced from five to three.

(B, D, F) However, when the three Gal4 sites are moved ~40 nucleotides away from the
Giant sites, the ability to repress all three types of activators is abolished. In this context
the three Gal4 sites are still within a 100 bp from the Giant binding sites.

Embryos are oriented with anterior to left; lateral views are shown.

115

Giant binding sites endowed a promoter with high or low sensitivity to repression (Hewitt
et al., 1999). We tested whether Giant’s ability to repress a smaller cluster of Gal4 sites
could be attenuated by small changes in spacing between the activator and repressor
binding sites. Moving the smaller cluster of three Gal4 sites 37 bp away from the Giant
binding sites results in the loss of repression (Figure III-6A compare to Figure III-6B)
suggesting that reducing the amount of activation potential does not guarantee repression
by Giant in all cases even when the activators are located within 100bp of the repressor
sites. In order to ascertain whether the spacing effects we see are specific to the activator
or a general property of the repression activity of Giant, we tested the ability of Giant to
block transcription mediated by the full length Gal4 protein expressed ubiquitously
throughout the embryo (Figure III-6C, D), and another activation domain, namely Gal4-
VP16 (Figure III-6E, F). As seen with the Gal4 activation domain, Giant is able to
repress lacZ expression mediated by the full length Gal4 protein (Figure III-6C) and
Gal4-VP16 (Figure III-6E) from three sites that are adjacent to the Giant binding sites.
However, moving the three sites 37 bp away again results in the loss of repression by
Giant of both Gal4 (Figure III-6D) and Gal4-VP16 (Figure III-6F) mediated activation.
These results overturn the previously held simple notion that short-range repressors block
all protein complexes within a 100 bp. The loss of repression caused by moving the
activator binding sites 37 bp away from Giant sites is not specific to a particular activator,
suggesting that the range over which short-range repressors function can be influenced by

the strength of the DNA binding domain of the activator.

114

, u u.

N‘A w .

it. we.” .xqi 4.3. u... w . N u t, I. .1; ..lv . .I. a

..\_ 3M». .fi .7. J... . .tw “fin . r u t . 1...“..— - .Na . 4n.» «aw Inn «1:... . . .r thuar

lb. .1. L Flu. “A.“ of. .l elm “Is a ,c u . UH“ . .r .. at» n... thunk rt

11.. al.,:- J v 7 IA Flt RV“ iin . 1. il . mil \. x :to r 71'
5.. I '1» .' I ‘l r ’ n ”r“

its

 

Figure III-5: The In a ‘ ’il-:-ibutiuu of short-range repressor binding sites is
critical in dictating repression effectiveness.

 

A B
x . ‘ y __ ’7
f W K lacZ f . R w
' I802 // '7
gt Gal4 gt 4;”,
C D

x, i .

faker-z f "" K

 

w
I. ',”>Ilf;j lIacZ ,, Iw
gt Gal4 gt Gal4 ,4"; kb
E F
I lacZ ‘ w

gt Gal4 ~4.5 kb

113

Figure III-5: The arrangement/distribution of short-range repressor binding sites is
critical in dictating repression effectiveness.

A schematic of the reporter gene is shown below the corresponding embryos.

The short-range repressor Giant is able to repress (arrows) both the proximal hsp 70 lacZ
gene (A) and the distal w (B) 4.5 kb away, when the repressor flanks the Gal4-activator.

Giant is also able to repress (arrows) in a reporter context where its binding sites are
interspersed between the Gal4 sites and repression is seen both at the proximal hsp70
lacZ gene (C) and the distal w (D) 4.5 kb away.

However, Giant is unable to repress in a context in which all the repressor sites are
clustered together on one side (E, F) of the five Gal4 sites although the number of
repressor and activator sites is the same as in A, B, C and D.

Embryos are oriented anterior to left; lateral views are shown.

112

side resulted in repression of the proximal lisp 7O lacZ reporter gene (Figure III-5A).
Interspersing the Giant repressor binding sites in between the Gal4 activator sites also
resulted in inhibition of lacZ expression (Figure III-5C). However, placing all four Giant
binding sites 5’ of the five Gal4 sites prevented Giant from repressing the hsp 70 lacZ
expression (Figure III-5E) suggesting that promoter response cannot be calculated simply
from overall activator to repressor stoichiometries. We noted that the Giant binding sites
in the reporter genes showing repression (Figure III-5A, C) are in close proximity to the
basal promoter and therefore, it is possible that in these contexts Giant directly represses
the basal promoter adding a confounding factor (Amosti et al., 1996b; Gray and Levine,
1996; Hewitt et al., 1999). To distinguish between repressor-basal promoter and
repressor-activator effects, we assayed for the activity of the w gene, which is ~4.5 kb
downstream of these cis-elements in each of the reporter gene constructs (Figure III-SB,
D, and F). Here again we observed that Giant mediated repression only when flanking or
interspersed with activators. The similar patterns of repression of proximal lacZ promoter
and distal w promoter suggest that Giant is acting on the activator cluster rather than only
on the basal promoter element (Figure III-SB, D). The preferential arrangement of
transcription factor binding sites within regulatory regions might be considered as a
specific type of functional information encoded in regulatory DNA that is critical in
dictating the transcriptional output of the cis-element. Indeed a distributed pattern of
short-range transcriptional repressor binding sites is typical of many developmental
enhancers that function in the early Drosophila embryo.

Previous analysis of the short-range repressor Giant demonstrated that due to the

extreme distance-dependent activity of this protein, subtle changes in the spacing of

111

Figure III-4: Repression not dependent on the nature of the activation domain.

 

 

A
Gal4-gal4AD

B
Gal4-VP16
Gal4-Sp1

D
Gal4-hTBP

110

Figure III-4: Repression not dependent on the nature of the activation domain.

A schematic representation of the reporter gene is shown below and the chimeric Gal4-
activators used to drive expression from the reporter gene are indicated to the left of the
embryos. Giant is unable to repress the activity of strong activators Gal4-Gal4 AD (A),
and Gal4-VP16 (B), or weaker activators Gal4-Spl (C), and Gal4-hTBP (D) on the hsp 70
reporter containing five high affinity Gal4 binding sites.

Embryos are oriented with anterior to left and lateral views are shown.

109

activator chimeric proteins were used to drive expression of the hsp 701acZ reporter from
a cluster of five high affinity Gal4 sites (Figure III-4A, B, C, and D). Giant could inhibit
neither the strong Gal4 (Figure III-4A) and VP16 (Figure III-4B) activators, nor the weak
activation domains of Spl (Figure III-4C) and hTBP (Figure III-4D) (Seipel et al., 1992).
These results indicate that the ability to repress does not depend on the strength of the
activation domain or the activation pathway. Only those genes in which the number or
affinity of Gal4 sites was reduced showed a response to Giant, suggesting that the Gal4
DNA binding domain provides a stable platform that can resist the activity of Giant and,
and suggests a mechanism for short-range repression that involves blocking activator

access to its cognate site.

The arrangement/distribution of short-range repressor binding sites is critical in

dictating repression effectiveness

Statistical models, based on motif clustering, are only partially successful at
finding novel cis-regulatory elements in the genome, perhaps because they consider only
site density and relative site affinity (Berman et al., 2002; Makeev et al., 2003; Markstein
and Levine, 2002; Markstein et al., 2002). However, it is probable that specific
arrangements of binding motifs are critical for proper biological function. We tested the
effect of alternative arrangements of Giant repressor and Gal4 activator sites to determine
if different arrangements or combinations resulted in distinct transcriptional outputs. In
all reporter arrangements tested, we used four Giant binding sites and five high affinity

Gal4 binding sites. Flanking the five Gal4 activator sites with two Giant sites on either

108

progressively refined into a two-stripe pattern, in regions where giant (Figure III-3A) is
not expressed (arrows; Figure III-3D-E)). Analysis of the transgene in a giant mutant
background, in the absence of Gal4, confirms that refinement of reporter gene expression
is due to repression by Giant (Figure III-3G). These results suggest that five Bicoid
binding sites are more susceptible to repression than are five high affinity Gal4 sites,
indicating that stoichiometric relationships of repressors to activators may depend also on

distinct DNA binding domains or activation domains.

Repression not dependent on the nature of activation domain

The differential effectiveness of Giant against five Gal4 or five Bicoid sites
suggests that the nature of the activation domain itself or the DNA binding domain, may
play a role in dictating the response to repressors. To distinguish between those two
possibilities, we compared the activities of a variety of activation domains fused to the
DNA binding domain of Gal4 (a 1-93). In addition to the well characterized activation
domain of the yeast transcriptional activator Gal4 (from aa 753-881), we tested the acidic
transcriptional activation domain of the herpes simplex virus activator VP16 (from a
412-490), the glutamine rich activation domain of the mammalian transcription factor
Spl (from aa 132-243), and the full-length human TATA binding protein (hTBP) which
has been shown to function as an activator when fused to the Gal4 DNA binding domain
(Seipel et al., 1992). We also sought to test the activity of Gal4-Bicoid activators (J anody
et al., 2001) but unfortunately these chimeras exhibit strong promoter specificity and are

not active on the hsp 70 promoter, precluding a side-by-side comparison. The Gal4-

107

Figure III-3: Effectiveness of repression correlates with the afl'mity of activator
binding sites.

 

 

 

 

Consensus UAS sites: 8 D _
5'CGGAGGAXXXTCC‘I’CCGS‘ 3“ a?"
Sequence of the Gal4 sites in the ‘9" ' K ‘7'"
hsp70 reporter (not repressed):
5'CGGAGTACTGTCCTCCGa' C E
. s
4'9 , . j,
The Low Affinity (LA) Gal4 sites: f K r f
5'CGGATTAGAAGCCGCCGts'
—_ IL 5 [a I. ---- 4 fizz.
High affinity Bicoid binding site °' °"‘ 4'" 9‘ 9'“ *°
" : __ . 0e ' f
is)" t ' P
I Q m
“' u-
..L $ [:65 u TTTTT 1 IE.2
9' Gal4 -oo at Gal4 450

106

Figure III-3: Effectiveness of repression correlates with the affinity of activator
binding sites.

Sequence of consensus Gal4 binding sites, high affinity Gal4 sites in the hsp70 reporter
(Figures III-1 and 2; not repressed), low affinity Gal4 binding sites, and high affinity
bicoid (Eisen et al., 2001) binding sites (underlined with a red line) created while creating
low affinity Gal4 sites are indicated on the left.

(A) giant expression in the early blastoderm embryo is dynamic and refines into two
stripe anteriorly and one stripe posteriorly. giant expression is visualized by in situ
hybridization using digoxigenin labeled antisense giant mRNA probe.

(B, C) Giant repression of the Gal4 activator. LacZ expression driven by the Gal4- Gal4
AD activator fusion from a cluster of five low affinity Gal4 binding sites upstream of the
hsp70 basal promoter elements is now repressed by the short-range repressor Giant
(arrows) in both the anterior and posterior regions of the embryo where the repressor is
present.

(D, E) Giant represses bicoid mediated activation of the hsp70 lacZ reporter (arrows).
Even in the absence of the Gal4 activator, LacZ expression is activated by the
transcription factor bicoid in the anterior region of the embryo from five high affinity
sites created in the reporter. Bicoid mediated activation gets progressively refined (D-E)
into two stripes of expression as the embryo develops, in regions where giant is not
expressed. A similar bicoid mediated activation of lacZ and subsequent refinement by the
repressor Giant is also evident in the anterior regions of the embryo in the presence of the
Gal4 activator (B, D).

(F, G) Analyses of lacZ expression from the hsp70 reporter in a giant mutant
background.

(F) Giant repression of Gal4- Gal4 AD mediated activation of hsp 70 lacZ is relieved in a
giant mutant background and lacZ staining is evident in the posterior region where Giant
is usually expressed (arrow).

(G) Analysis of the hsp70 lacZ reporter in the absence of the Gal4 activator. Bicoid
mediated activation is no longer refined into a two stripe pattern in a giant mutant
background.

Embryos are oriented anterior to left; ventral (F) and lateral (A, B, D, E, and G) views
are shown.

105

Binding site affinity influences threshold responses to activator gradients in the
embryo, (J iang and Levine, 1993; Struhl, 1989; Szymanski and Levine, 1995) and indeed
transcription factor binding sites of varying affinities are typically found in many
developmental enhancers that function in the early Drosophila development. Such
differences in activator site affinity might similarly influence responses to short-range
repressors. We tested whether maintaining the number of activator sites but weakening
their affinity would in fact change the reponse to repressors. We replaced the five high
affinity Gal4 binding sites in the hsp701acZ reporter, with five low affinity sites from the
endogenous yeast S. cerevisiae Gall-GallO promoters that have been characterized to
bind Gal4 poorly (Burns and Peterson, 1997; Johnston and Davis, 1984). The lower
affinity Gal4 sites drive gene expression in a weaker striped pattern, but anterior and
posterior repression by Giant is clearly evident (Figure III-3B, C; arrows). Repression is
relieved in a giant mutant embryo (Figure III-3F), confirming that the gaps in expression
seen in Figure III-3B, and C, are mediated by the Giant repressor.

In the process of weakening the Gal4 binding sites, we inadvertently created five
high affinity binding sites for the Bicoid activator (Berman et al., 2002) providing an
additional opportunity to assay Giant repression activity. Bicoid is maternally deposited
in the anterior regions of the embryo, forming an anterior to posterior gradient (Driever
and Nusslein-Volhard, 1988; Eldon and Pirrotta, 1991; Kraut and Levine, 1991). LacZ
expression from the hsp 70 reporter is activated even in the absence of the Gal4 activator
by Bicoid transcription factor in anterior regions (Figure III-3D and E). As the embryo

develops (Figure III-3D-E), Giant inhibits Bicoid mediated lacZ activation, which is

104

Figure 111- 2: Stoichiometry between the number of activators and repressors
influences repression effectiveness.

 

 

 

103

Figure III- 2: Stoichiometry between the number of activators and repressors
influences repression effectiveness.

(A, B) LacZ expression driven by the Gal4- Gal4 AD activator fusion fi'om a cluster of
five high affinity Gal4 binding sites upstream of the hsp 70 basal promoter elements is not
repressed by the short-range repressor Giant.

(C-F) Reducing the amount of activator by reducing the number of Gal4 sites from five
to three restores the ability of Giant to block lacZ expression (arrows).

Embryos are oriented anterior to left; lateral views (A, C, E, and F) and ventrolateral
views (B,andD) shown.

102

Repression sensitivity correlated to the strength of the activating signal

Studies of the cis-regulatory elements of the hairy gene in Drosophila led to the
suggestion that the overall stoichiometry, rather than the absolute number, of activators
and repressors maybe critical in dictating enhancer output (La Rosee et al., 1997). To test
whether the stoichiometry of enhancer- bound activators to repressors is a critical factor
in determining short-range repression levels by Giant, we reduced from five to three the
number of Gal4 activator binding sites on the hsp70 - lacZ reporter. As anticipated, the
levels of transcriptional activation were lower in the transgene containing three Gal4
sites, leading to a less robust ventral staining pattern (Figure III-2A, B). In this context,
Giant was able to block transcription of the lacZ gene (Figure III-2C, D). However, the
removal of two Gal4 sites also positions the repressors closer to the start of transcription,
which may facilitate direct basal promoter repression. Therefore, to maintain the distance
between Giant binding sites and the start of transcription we tested a reporter gene where
a neutral spacer was placed downstream of the three Gal4 sites (Figure III-2E, F). Again,
Giant was also able to repress the Gal4 activators. These results demonstrate that
repression is critically dependent on the number of activator binding sites but do not
explicitly differentiate between overall level of transcriptional activation and binding site
number, an issue explored below. These results are also consistent with previous analysis
of the eve stripe 2 element, where the insertion of additional Bicoid binding sites in an
otherwise normal stripe 2 enhancer causes a slight anterior expansion of its expression
pattern, suggesting that an excess of Bicoid activators can ‘overwhelm’ the Giant

repressor (Amosti et al., 1996a).

101

Figure III-7: Repressors can be distinguished by their differential ability to repress
in different enhancer configurations.

 

 

A B
Glant .
f 0v ‘2 My: K V
C D
Knirps h
f a 2“. ~- ~ n. ‘ .. 1r
\. ”D... .
K
E F
Kriippel ‘
i I 3!; Ed- ‘- I' ' f?” 3 .-
\- sew ”gti'm‘?_g y
f
G H
Hairy .F I ‘- = -
a u . ‘0 f3. "3 ., 1 1
Mai if. a, if. . 2 i g ’3: it
mm It»(:2 1%,?“
rep Gal4 -so rep Gel4 -60

119

DISCUSSION

Using defined synthetic enhancer elements to study short-range repressor action,
we demonstrate that there is a rich contextual grammar that influences repressor activity,
extending beyond the generalization that short-range repressors block the activity of all
protein complexes within a 100 bp. Although the distance between short-range repressors
and their targets is a critical factor in dictating repression effectiveness, it is not the only
one, and in some cases, proximity is not sufficient. Activators can escape repression even
when they are within the previously defined effective range of short-range repression.
The manipulation of these composite elements in terms of the number of activator and
repressor binding sites, relative affinities, spacing and distribution of binding sites firrther
allowed us to define other contextual parameters that dictate repression effectiveness.
First, we find that a balanced ratio of enhancer-bound activators and repressors is an
important factor dictating repressibility. In the context of five high affinity Gal4 sites,
two adjacent repressor sites are insufficient to permit the repression by Giant. Reducing
the number of Gal4 sites from five to three allowed Giant to block Gal4 mediated
activation of the lacZ reporter gene. Second, although the effectiveness of repression
depends on stoichiometry between the number of activators and repressors, Giant
repression of a smaller cluster of activators can be attenuated by subtle changes (< 40 bp)
in the spacing between the repressor and activator binding sites even when binding sites
are within the previously defined 100 bp “repression zone”. A third parameter is that
repression effectiveness correlates with the affinity of activator binding sites, and

although binding affinity influences the strength of the activating signal, repression does

120

not depend on the chemical nature of the activation domain. Fourth, the short-range
repressors need to be judiciously placed, either flanking activator sites or interspersed
among them, possibly to block multiple modes of enhancer-promoter interactions. These
contextual parameters constitute the ‘grammar’ of short-range repressor function.
Empirical determination of such elements contributes to our understanding of enhancer
design and should find application in bioinfonnatic analysis of novel gene regulatory
sequences, as well as providing insights into the evolution and biochemical activity of

short-range repressors as we describe below.

Computational analysis of cis-regulatory elements

Enhancers serve as binding platforms to localize multiple distinct transcription
factors, including sequence-specific activators and repressors (Amosti, 2002; Davidson,
2001). These factors act combinatorially to confer context-specific transcriptional
activity. In effect, changing the combinations and arrangements of transcription factor
binding sites within the enhancer can vary the nature of the transcriptional output
generated by cis—regulatory elements. Thus, any computational algorithm designed to
identify cis-regulatory modules must take into account the combinatorial logic underlying

cis-regulatory module architecture.

Extensive analysis of the Drosophila eve stripe enhancers and their trans-acting

factors has provided insights into the complexities inherent in decoding the combinatorial

nature of cis-regulatory logic. The eve stripe 2 element is one of the best characterized

121

enhancer and contains a total of eighteen known factor-binding sites, including eight
activator (three for Hunchback, five for Bicoid) binding sites and ten repressor (three for
Giant, six for Kruppel and at least one for Sloppy-paired) binding sites (Andrioli et al.,
2002; Ludwig et al., 1998; Stanojevic et al., 1991). Detailed site mutageneses studies
have demonstrated that every (most) transcription factor binding site within the stripe 2
element contributes to the overall transcriptional output (Small et al., 1992) suggesting
that stripe 2 expression, to a large extent, depends on a balanced ratio of repressors and
transcriptional activators. However, studies have also indicated that this notion is an
oversimplification and that simple multimerization of the known activators and repressor
sites is not sufficient to recapitulate the expression of the native element. Other
contextual features, such as spacing/orientation constraints between binding sites within

the enhancer, may play an important role.

Different concentrations of the Giant, Knirps and Kruppel repressor proteins have
also been shown to be important for proper regulation of many zygotic genes during early
embryogenesis. For example, low concentrations of Giant are sufficient to repress the
Krilppel promoter, while greater amounts appear to be required to repress the eve stripe 2
element (Wu et al., 1998). In a similar vein, higher levels of the Knirps protein are
required to regulate the eve stripe 3 + 7 enhancer while lower levels are sufficient for the
proper regulation of the eve stripe 4 + 6 element (Clyde et al., in press; Struffi et al.,
submitted). Sequence analysis of the eve gene indicates that there are more high affinity
Knirps binding sites within the eve stripe 3 + 7 element than in the 4 + 6 enhancer,

consistent with the relative sensitivities of these elements determined experimentally

122

(Berman et al., 2002; Papatsenko et al., 2002). Removal of some of the Knirps binding
sites from the eve 3 + 7 enhancer reduces the sensitivity of this element to the Knirps
gradient (Clyde et al., in press). Thus, these studies indicate that in addition to the number
of binding sites, binding site affinities within target genes define the molecular basis of
differential responses to these transcriptional repressors and are crucial for the precise

positioning of gene expression borders.

Transcription factors that function as morphogens generate unique threshold
responses at different concentrations. This point is relevant to the clustering of cognate
DNA binding sites, because the number and affinities of these sites provide a readout of
the local transcription factor concentration. For example, a cis-regulatory module with a
large number of sites or with high affinity sites would be activated in response to low
levels of the corresponding factor. Additionally, it has been demonstrated that the same
cis-regulatory module can generate two different transcriptional outputs depending on its
position within the morphogenetic gradient of the transcription factor that regulates it

(Hewitt et al., 1999).

Transcription factors bound to regulatory DNA are often involved in specific
protein-protein interactions. Thus, the exact arrangement of binding motifs may control
the formation of three dimensional protein complexes that are essential for the
appropriate biological output of that element (Cai et al., 1996; Merika and Thanos, 2001).
Therefore, the binding motifs are distributed in a non-random fashion within cis-elements

leading to specific, functionally relevant arrangements. However, some plasticity in the

123

positioning of binding sites is tolerated in some situations and might reflect the way in
which enhancer-boundfactorsinteract with the basal transcriptional machinery (Kulkami,
2003). For example, in the case of the information display enhancer, discrete portions of
the element can be ‘read’ off by the basal transcriptional machinery. Individual
subelements within the enhancer can signal independently of each other to the basal
transcriptional machinery and therefore, the exact spatial arrangement between the
autonomously acting subelements may not be critical. The independent information units
might consist of one or a few factor binding sites, the enhancer substructure (Kulkami,
2003) and thus, it is conceivable that within an independently acting information unit
functional constraints between factor binding sites exist, which define the ‘grammar’

rules governing enhancer function.

Thus, the combinatorial logic underlying cis-regulatory module architecture
requires not only consideration of binding site numbers and affinities, but also
combinations of binding site type, spacing, and positioning within the cis-element. These
parameters are difficult if not impossible to predict accurately a priori. So far structure-
function analyses of cis-regulatory elements have focused on complex endogenous
elements and therefore are complicated by the complexity of these elements where the
exact identity, stoichiometry, relative spacing and arrangement of binding sites is not
known. Although these studies have shed some light on the contextual features that might
play a critical role in defining a function cis-regulatory module, the parameters have not
been accurately determined. Computational approaches will benefit from the availability

of at least one well-defined representative of a particular network as a starting paradigm.

124

Using synthetic well-defined genetic elements we have defined such a paradigm for the
short-range transcriptional repressors. Incorporating the ‘grammar’ rules for short-range
repressors into search algorithms will be better able to predict the possibility of Giant or
Knirps regulating a given element given the relative the relative number, affinities,
spacing and distribution of repressor and activator sites within the element. Empirical
testing of identified candidates should enable further refinement of the model, which in
turn can be utilized for another round of screening. This underscores the importance of

combining bioinforrnatics with experimental biology.

Evolution of transcriptional regulation

It has been shown that regulatory sequences can maintain regulatory function
despite structural reorganization as a result of species-specific loss and gain of
transcription factor binding sites (Cuadrado et al., 2001; Ludwig et al., 2000; Piano et al.,
1999). Thus, to decipher what kind of changes in cis-element structure cause a change in
function, an underlying cause for morphological evolution, detailed information about the
locations of transcription factor binding sites within the element, the functional
specificity of the binding sites and the spatial requirements for their interaction is
required. The empirical determination of the structural parameters that dictate enhancer
function, such as those described here for elements regulated by short-range repressors,
can facilitate the modeling of the evolutionary dynamics of cis-regulatory regions in a
stepwise fashion, starting with sequence comparisons between closely related species,
and then extending to more distantly related organisms. Thus, if we knew in functional

terms the components of the specific genomic cis-elements that result in different

125

morphological outcomes in two animals of common ancestry, we could see exactly what
are the essential causal differences in the DNA of these animals and provide an
explanation, in a mechanistically relevant way, of how the diverse forms of animal life

actually arose during evolution.

Mechanism of short-range repression

A repressor that functions in all, or most, contexts is more likely to function via a
‘general’ mechanism that is inhibitory at all promoters, such as preventing activator
binding by establishing a repressive chromatin domain. In contrast, a repressor that can
inhibit transcription only in particular contexts is more likely to target a specific activator
or proteins that mediate the action of that specific activator. Selective, context related
repression affords an added layer of combinatorial control of gene expression by
sequence specific transcription factors. Previous analyses of short-range repression have
demonstrated that Giant, Knirps, Kriippel and Snail can block the activity of a number of
distinct activators such as Bicoid, Hunchback, Dorsal, Twist, and D-Stat (Amosti et al.,
1996a; Gray et al., 1994; Small et al., 1996). Many biochemical and genetic analyses
suggest that at least some of these activators activate transcription via distinct pathways
(Koh et al., 1998; Pham et al., 1999; Yuh et al., 1998; Zhou et al., 1998). The apparent
lack of activator specificity demonstrated by short-range repressors suggests that these
proteins function via a general mechanism. However, the interpretation of the results
from previous studies is complicated by the fact that the activators tested represent

distinct classes of transcriptional activators not only in terms of the nature of their

126

activation domains and activation pathways, but also because they possess different DNA
binding specificities. Using synthetic enhancer elements where the identity, stoichiometry
and exact spacing of activator and repressor binding sites is well-defined, we have
demonstrated that repression effectiveness does not depend on the nature of the activation
domain but correlates instead with DNA binding site affinity. Thus, we propose that the
short-range repressors inhibit transcription by blocking access to DNA by transcriptional
activators through chromatin remodeling at a very local scale. Subtle (< 40 bp) changes
in the spacing of Giant from the nearest activators sharply attenuates repression
effectiveness suggesting that the action of short-range repressors is more localized than

the previously defined range of a 100 bp.

A common property of short-range transcriptional repressors is their interaction
with the evolutionarily conserved corepressor CtBP (C-terminal Binding Protein) through
short PXDLS peptide motifs found in the transcriptional repressors. CtBP has been
shown to interact with chromatin modifying factors, including histone deacetylases
(HDACl and HDAC2) and histone methyltransferases (Chinnadurai, 2002; Chinnadurai,
2003; Shi et al., 2003; Subramanian and Chinnadurai, 2003; Sundqvist et al., 1998).
Thus, the major activity of CtBP might be to serve as a bridging molecule, recruiting
chromatin or transcription factor modifying enzymes to the promoter. In addition, studies
in our laboratory have demonstrated a genetic and physical interaction between de3, the
Drosophila homolog of HDACl, and the short-range repressor Knirps (Struffi,

unpublished data). The de3 protein in yeast is known to deacetylate histones at an

127

extremely local level, consistent with its role in short-range repression in Drosophila
(Beckett and Struhl, 2002).
Although the contextual elements defined above allow us to hypothesize about the
molecular mechanisms of short-range repression, we do not yet understand the physical
and biochemical changes in promoter complexes that accompany short-range repression.
Two proposed mechanisms are consistent with their activity seen in vivo: Short-range
repressors might “quench” activator proteins locally, within a short distance of their
binding site by displacing them from the DNA through chromatin structure modification,
or by preventing them from contacting their target in the basal transcriptional machinery;
or the repressors might directly contact some component of the basal transcriptional
machinery, but only when brought to the promoter by a closely linked activator
(hitchhiking) or when bound near the start site of transcription. The simple transcriptional
switch elements defined in this study will facilitate further molecular and biochemical
characterization of short-range repressors. These elements can be used in whole
Drosophila embryos chromatin immunoprecipitation assays to examine the nature of the

promoter complexes and the chromatin state before and after repression.

Similar to the studies that have categorized activators according to transcriptional
activation pathways, analyses of transcriptional repression in the fruit fly has revealed
that repressors can be classified as short-range or long-range repressors on the basis of
their range of action. The mechanism by which short-range and long-range Drosophila
repressors inhibit transcription are poorly understood, although one model of repression

in the embryo suggests that the short-range/long-range distinction results from the

128

I.ecfuitment of distinct cofactors (Nibu et al., 1998a; Nibu et al., 1998b; Zhang and
Levine. 1 999). Short-range repressors recruit the corepressor CtBP to mediate repression
whereas long-range repressors have been shown to interact with the Groucho corepressor
(Barolo and Levine, 1997; Chen and Courey, 2000; Fisher and Candy, 1998; Mannervik

and Levine, 1999; Mannervik et al., 1999; Poortinga et al., 1998).

We show that the long-range repressor Hairy can block transcription in a context
where the short-range repressors cannot function and appears not to require a
stoichiometric relationship of activators to repressors. Thus, the Hairy long-range
repressor seems to be quantitatively stronger (as a repressor) than a short-range repressor,
not only due to its longer range of action but also by the virtue of its ability to block the
Gal4 activators, in the context of five high affinity sites, that were uninhibited by Giant,
Knirps and Kriippel. However, we do not exclude the possibility that Hairy may be able
to repress in a context where we do not see repression by the short-range repressors

simply due to higher levels of Hairy protein.

Long-range repressors have been likened to silencers, as they function in a
dominant fashion to shut down multiple enhancers in a complex modular promoter
(Barolo and Levine, 1997; Cal et al., 1996). However, we demonstrate that Hairy
mediated repression is transient and easily reversible as the embryo ages, much like the
action of the short-range repressors Giant, Knirps and Kruppel. Another striking paradox
is that like Groucho, a number of studies have suggested that CtBP may function, at least

in part, by recruiting histone deacetylases (Criqui-Filipe et al., 1999; Sundqvist et al.,

129

1998). Thus, if both long- and short-range corepressors function through histone
deacetylation, why is Hairy is more potent with a longer range of action than Giant? The
synthetic enhancer elements defined in this study can serve as useful tools to compare the
biochemical activities of these two classes of transcriptional regulators in order to resolve
this issue and distinguish between the molecular mechanisms of short-range and long-

range I‘Cpl’CSSOI‘S.

130

REFERENCES

Andrioli, L. P., Vasisht, V., Theodosopoulou, E., Oberstein, A. and Small, S. (2002).
Anterior repression of a Drosophila stripe enhancer requires three position-specific
mechanisms. Development 129, 4931-40.

Arnosti, D. N. (2002). Design and function of transcriptional switches in Drosophila.
Insect Biochem Mol Biol 32, 1257-73.

Arnosti, D. N., Barolo, S., Levine, M. and Small, S. (1996a). The eve stripe 2 enhancer
employs multiple modes of transcriptional synergy. Development 122, 205-14.

Arnosti, D. N., Gray, S., Barolo, S., Zhou, J. and Levine, M. (1996b). The gap protein
knirps mediates both quenching and direct repression in the Drosophila embryo. Embo J
15, 3659-66.

Barolo, S., Carver, L. A. and Posakony, J. W. (2000). GFP and beta-galactosidase
transformation vectors for promoter/enhancer analysis in Drosophila. Biotechniques 29,
726, 728, 730, 732.

Barolo, S. and Levine, M. (1997). hairy mediates dominant repression in the Drosophila
embryo. Embo J 16, 2883-91.

Bergman, C. M. and Kreitman, M. (2001). Analysis of conserved noncoding DNA in
Drosophila reveals similar constraints in intergenic and intronic sequences. Genome Res
11, 1335-45.

Berman, B. P., Nibu, Y., Pfeiffer, B. D., Tomancak, P., Celniker, S. E., Levine, M.,
Rubin, G. M. and Eisen, M. B. (2002). Exploiting transcription factor binding site
clustering to identify cis-regulatory modules involved in pattern formation in the
Drosophila genome. Proc Natl Acad Sci U S A 99, 757-62.

Brand, A. H. and Perrimon, N. (1993). Targeted gene expression as a means of altering
cell fates and generating dominant phenotypes. Development 118, 401-15.

Brantjes, H., Roose, J., van De Wetering, M. and Clevers, H. (2001). All ch HMG
box transcription factors interact with Groucho-related co-repressors. Nucleic Acids Res
29, 1410-9.

Burns, L. G. and Peterson, C. L. (1997). The yeast SWI—SNF complex facilitates
binding of a transcriptional activator to nucleosomal sites in vivo. Mol Cell Biol 17, 4811-
9.

Cal, H. N., Arnosti, D. N. and Levine, M. (1996). Long-range repression in the
Drosophila embryo. Proc Natl Acad Sci U S A 93, 9309-14.

131

Capovilla, M., Eldon, E. D. and Pirrotta, V. (1992). The giant gene of Drosophila
encodes a b-ZIP DNA-binding protein that regulates the expression of other segmentation
gap genes. Development 114, 99-112.

Chen, G. and Courey, A. J. (2000). Groucho/TLE family proteins and transcriptional
repression. Gene 249, 1-16.

Chen, G., Fernandez, J., Mische, S. and Courey, A. J. (1999). A functional interaction
between the histone deacetylase de3 and the corepressor groucho in Drosophila
development. Genes Dev 13, 2218-30.

Chinnadurai, G. (2002). CtBP, an unconventional transcriptional corepressor in
development and oncogenesis. Mol Cell 9, 213-24.

Chinnadurai, G. (2003). CtBP family proteins: more than transcriptional corepressors.
Bioessays 25, 9-12.

Courey, A. J. and Huang, J. D. (1995). The establishment and interpretation of
transcription factor gradients in the Drosophila embryo. Biochim Biophys Acta 1261, 1-
18.

Criqui-Filipe, P., Ducret, C., Maira, S. M. and Wasylyk, B. (1999). Net, a negative
Ras-switchable TCF, contains a second inhibition domain, the CID, that mediates
repression through interactions with CtBP and de-acetylation. Embo J 18, 3392-403.

Cuadrado, M., Sacristan, M. and Antequera, F. (2001). Species-specific organization
of CpG island promoters at mammalian homologous genes. EMBO Rep 2, 586-92.

Davidson, E. H. (2001). Genomic Regulatory Systems: Development and Evolution:
Academic Press: A Harcourt Science and Technology Company.

Driever, W. and Nusslein-Volhard, C. (1988). A gradient of bicoid protein in
Drosophila embryos. Cell 54, 83-93.

Driever, W., Thoma, G. and Nusslein-Volhard, C. (1989). Determination of spatial
domains of zygotic gene expression in the Drosophila embryo by the affinity of binding
sites for the bicoid morphogen. Nature 340, 363-7.

Eldon, E. D. and Pirrotta, V. (1991). Interactions of the Drosophila gap gene giant with
maternal and zygotic pattern-forming genes. Development 111, 367-78.

Fisher, A. L. and Candy, M. (1998). Groucho proteins: transcriptional corepressors for
specific subsets of DNA-binding transcription factors in vertebrates and invertebrates.
Genes Dev 12, 1931-40.

Flores-Saaib, R. D. and Courey, A. J. (2000). Analysis of Groucho-histone interactions
suggests mechanistic similarities between Groucho- and Tupl-mediated repression.
Nucleic Acids Res 28, 4189-96.

132

Fujioka, M., Emi-Sarker, Y., Yusibova, G. L., Goto, T. and Jaynes, J. B. (1999).
Analysis of an even-skipped rescue transgene reveals both composite and discrete
neuronal and early blastoderm enhancers, and multi-stripe positioning by gap gene
repressor gradients. Development 126, 2527-38.

Goto, T., Macdonald, P. and Maniatis, T. (1989). Early and late periodic patterns of
even skipped expression are controlled by distinct regulatory elements that respond to
different spatial cues. Cell 57, 413-22.

Gray, S. and Levine, M. (1996). Transcriptional repression in development. Curr Opin
Cell Biol 8, 358-64.

Gray, S., Szymanski, P. and Levine, M. (1994). Short-range repression permits
multiple enhancers to function autonomously within a complex promoter. Genes Dev 8,
1829-38.

Guss, K. A., Nelson, C. E., Hudson, A., Kraus, M. E. and Carroll, S. B. (2001).
Control of a genetic regulatory network by a selector gene. Science 292, 1164-7.

Hanes, S. D., Riddihough, G., Ish-Horowicz, D. and Brent, R. (1994). Specific DNA
recognition and intersite spacing are critical for action of the bicoid morphogen. Mol Cell
Biol 14, 3364-75.

Harding, K., Hoey, T., Warrior, R. and Levine, M. (1989). Autoregulatory and gap
gene response elements of the even-skipped promoter of Drosophila. Embo J 8, 1205-12.

Hardison, R., Slightom, J. L., Gumucio, D. L., Goodman, M., Stojanovic, N. and
Miller, W. (1997). Locus control regions of mammalian beta-globin gene clusters:
combining phylogenetic analyses and experimental results to gain functional insights.
Gene 205, 73-94.

Hewitt, G. F., Strunk, B. S., Margulies, C., Priputin, T., Wang, X. D., Amey, R.,
Pabst, B. A., Kosman, D., Reinitz, J. and Arnosti, D. N. (1999). Transcriptional
repression by the Drosophila giant protein: cis element positioning provides an
alternative means of interpreting an effector gradient. Development 126, 1201-10.

Janody, F., Sturny, R., Schaeffer, V., Azou, Y. and Dostatni, N. (2001). Two distinct
domains of Bicoid mediate its transcriptional downregulation by the Torso pathway.
Development 128, 2281-90.

Jiang, J. and Levine, M. (1993). Binding affinities and cooperative interactions with
bHLH activators delimit threshold responses to the dorsal gradient morphogen. Cell 72,
741-52.

Johnston, M. and Davis, R. W. (1984). Sequences that regulate the divergent GALl-
GALIO promoter in Saccharomyces cerevisiae. Mol Cell Biol 4, 1440-8.

133

Keller, S. A., Mao, Y., Struffi, P., Margulies, C., Yurk, C. E., Anderson, A. R.,
Amey, R. L., Moore, S., Ebels, J. M., Foley, K. et al. (2000). dCtBP-dependent and -

independent repression activities of the Drosophila Knirps protein. Mol Cell Biol 20,
7247-58.

Kim, T. K. and Maniatis, T. (1997). The mechanism of transcriptional synergy of an in
vitro assembled interferon-beta enhanceosome. Mol Cell 1, 119-29.

Koh, S. S., Ansari, A. Z., Ptashne, M. and Young, R. A. (1998). An activator target in
the RNA polymerase 11 holoenzyme. Mol Cell 1, 895-904.

Kosman, D. and Small, S. (1997a). Concentration-dependent patterning by an ectopic
expression domain of the Drosophila gap gene knirps. Development 124, 1343-54..

Kraut, R. and Levine, M. (1991a). Mutually repressive interactions between the gap
genes giant and Kruppel define middle body regions of the Drosophila embryo.
Development 111, 611-21.

Kraut, R. and Levine, M. (1991b). Spatial regulation of the gap gene giant during
Drosophila development. Development 111, 601-9.

Kulkami, M. M. and Arnosti, D. N. (2003). Information Display by Transcriptional
enhancer. In press.

La Rosee, A., Hader, T., Taubert, H., Rivera-Pomar, R. and Jackle, H. (1997).
Mechanism and Bicoid-dependent control of hairy stripe 7 expression in the posterior
region of the Drosophila embryo. Embo J 16, 4403-11.

Lifanov, A. P., Makeev, V. J., Nazina, A. G. and Papatsenko, D. A. (2003).
Homotypic regulatory clusters in Drosophila. Genome Res 13, 579-88.

Loots, G. G., Locksley, R. M., Blankespoor, C. M., Wang, Z. E., Miller, W., Rubin,
E. M. and Frazer, K. A. (2000). Identification of a coordinate regulator of interleukins
4, 13, and 5 by cross-species sequence comparisons. Science 288, 136-40.

Ludwig, M. Z., Bergman, C., Patel, N. H. and Kreitman, M. (2000). Evidence for
stabilizing selection in a eukaryotic enhancer element. Nature 403, 564-7.

Ludwig, M. Z. and Kreitman, M. (1995). Evolutionary dynamics of the enhancer
region of even-skipped in Drosophila. Mol Biol Evol 12, 1002-1 1.

Ludwig, M. Z., Patel, N. H. and Kreitman, M. (1998). Functional analysis of eve stripe
2 enhancer evolution in Drosophila: rules governing conservation and change.
Development 125, 949-58.

Maier, D., Preiss, A. and Powell, J. R. (1990). Regulation of the segmentation gene
fushi tarazu has been functionally conserved in Drosophila. Embo J 9, 3957-66.

134

Makeev, V. J., Lifanov, A. P., Nazina, A. G. and Papatsenko, D. A. (2003). Distance
preferences in the arrangement of binding motifs and hierarchical levels in organization
of transcription regulatory information. Nucleic Acids Res 31, 6016-26.

Mannervik, M. and Levine, M. (1999). The de3 histone deacetylase is required for
segmentation of the Drosophila embryo. Proc Natl Acad Sci U S A 96, 6797-801.

Mannervik, M., Nibu, Y., Zhang, H. and Levine, M. (1999). Transcriptional
coregulators in development. Science 284, 606-9.

Markstein, M. and Levine, M. (2002). Decoding cis-regulatory DNAs in the
Drosophila genome. Curr Opin Genet Dev 12, 601-6.

Markstein, M., Markstein, P., Markstein, V. and Levine, M. S. (2002). Genome-wide
analysis of clustered Dorsal binding sites identifies putative target genes in the
Drosophila embryo. Proc Natl Acad Sci U S A 99, 763-8.

Merika, M. and Thanos, D. (2001). Enhanceosomes. Curr Opin Genet Dev 11, 205-8.

Munshi, N., Agalioti, T., Lomvardas, S., Merika, M., Chen, G. and Thanos, D.
(2001). Coordination of a transcriptional switch by HMGI(Y) acetylation. Science 293,
1133-6. -

Nibu, Y., Zhang, H., Bajor, E., Barolo, S., Small, S. and Levine, M. (1998a). dCtBP
mediates transcriptional repression by Knirps, Kruppel and Snail in the Drosophila
embryo. Embo J 17, 7009-20.

Nibu, Y., Zhang, H. and Levine, M. (1998b). Interaction of short-range repressors with
Drosophila CtBP in the embryo. Science 280, 101-4.

Nibu, Y., Zhang, H. and Levine, M. (2001). Local action of long-range repressors in the
Drosophila embryo. Embo J 20, 2246-53.

Nusslein-Volhard, C., Frohnhofer, H. G. and Lehmann, R. (1987). Determination of
anteroposterior polarity in Drosophila. Science 238, 1675-81.

Nusslein-Volhard, C., Kluding, H. and Jurgens, G. (1985). Genes affecting the
segmental subdivision of the Drosophila embryo. Cold Spring Harb Symp Quant Biol 50,
145-54.

Nusslein-Volhard, C. and Wieschaus, E. (1980). Mutations affecting segment number
and polarity in Drosophila. Nature 287, 795-801.

Pankratz, M. J. and Jackle, H. (1990). Making stripes in the Drosophila embryo.
Trends Genet 6, 287-92.

135

Papatsenko, D. A., Makeev, V. J., Lifanov, A. P., Regnier, M., Nazina, A. G. and
Desplan, C. (2002). Extraction of functional binding sites from unique regulatory
regions: the Drosophila early developmental enhancers. Genome Res 12, 470-81.

Pham, A. D., Muller, S. and Sauer, F. (1999). Mesoderm-determining transcription in
Drosophila is alleviated by mutations in TAF(II)60 and TAF(II)110. Mech Dev 84, 3-16.

Piano, F., Parisi, M. J., Karess, R. and Kamhysellis, M. P. (1999). Evidence for
redundancy but not trans factor-cis element coevolution in the regulation of Drosophila
Yp genes. Genetics 152, 605-16.

Poortinga, G., Watanabe, M. and Parkhurst, S. M. (1998). Drosophila CtBP: a Hairy-
interacting protein required for embryonic segmentation and hairy-mediated
transcriptional repression. Embo J 17, 2067-78.

Rebeiz, M., Reeves, N. L. and Posakony, J. W. (2002). SCORE: a computational
approach to the identification of cis-regulatory modules and target genes in whole-
genome sequence data. Site clustering over random expectation. Proc Natl Acad Sci U S
A 99, 9888-93.

Rivera-Pomar, R. and Jackle, H. (1996). From gradients to stripes in Drosophila
embryogenesis: filling in the gaps. Trends Genet 12, 478-83.

Ryu, J. R., Olson, L. K. and Arnosti, D. N. (2001). Cell-type specificity of short-range
transcriptional repressors. Proc Natl Acad Sci U S A 98, 12960-5.

Sackerson, C., Fujioka, M. and Goto, T. (1999). The even-skipped locus is contained
in a l6-kb chromatin domain. Dev Biol 211, 39-52.

Sean B. Carroll, J. K. G., Scott D. Weatherbee. (2001). From DNA to Diversity:
Molecular Genetics and the Evolution of Animal Design: Blackwell Science.

Seipel, K., Georgiev, O. and Schaffner, W. (1992). Different activation domains
stimulate transcription from remote ('enhancer') and proximal ('promoter') positions.
EmboJll, 4961-8.

Shi, Y., Sawada, J., Sui, G., Affar el, B., Whetstine, J. R., Lan, F., Ogawa, H., Luke,
M. P. and Nakatani, Y. (2003). Coordinated histone modifications mediated by a CtBP
co-repressor complex. Nature 422, 735-8.

Small, S., Arnosti, D. N. and Levine, M. (1993). Spacing ensures autonomous
expression of different stripe enhancers in the even-skipped promoter. Development 119,
762-72.

Small, S., Blair, A. and Levine, M. (1992). Regulation of even-skipped stripe 2 in the
Drosophila embryo. Embo J 11, 4047-57.

136

Small, S., Blair, A. and Levine, M. (1996). Regulation of two pair-rule stripes by a
single enhancer in the Drosophila embryo. Dev Biol 175, 314-24.

Small, S., Kraut, R., Hoey, T., Warrior, R. and Levine, M. (1991). Transcriptional
regulation of a pair-rule stripe in Drosophila. Genes Dev 5, 827-39.

Small, S. and Levine, M. (1991). The initiation of pair-rule stripes in the Drosophila
blastoderm. Curr Opin Genet Dev 1, 255-60.

Stanojevic, D., Small, S. and Levine, M. (1991). Regulation of a segmentation stripe by
overlapping activators and repressors in the Drosophila embryo. Science 254, 1385-7.

Struhl, G. (1989). Morphogen gradients and the control of body pattern in insect
embryos. Ciba Found Symp 144, 65-86; discussion 86-91, 92-8.

Struhl, K. (1991). Mechanisms for diversity in gene expression patterns. Neuron 7, 177-
81.

Struhl, K. (2001). Gene regulation. A paradigm for precision. Science 293, 1054-5.

Strunk, B., Struffi, P., Wright, K., Pabst, B., Thomas, J., Qin, L. and Arnosti, D. N.
(2001). Role of CtBP in transcriptional repression by the Drosophila giant protein. Dev
Biol 239, 229-40.

Subramanian, T. and Chinnadurai, G. (2003). Association of class I histone
deacetylases with transcriptional corepressor CtBP. FEBS Lett 540, 255-8.

Sundqvist, A., Sollerbrant, K. and Svensson, C. (1998). The carboxy-tenninal region
of adenovirus ElA activates transcription through targeting of a C-terrninal binding
protein-histone deacetylase complex. FEBS Lett 429, 183-8.

Szymanski, P. and Levine, M. (1995). Multiple modes of dorsal-bHLH transcriptional
synergy in the Drosophila embryo. Embo J 14, 2229-38.

Thanos, D. and Maniatis, T. (1995). Virus induction of human IFN beta gene
expression requires the assembly of an enhanceosome. Cell 83, 1091-100.

Wu, X., Vakani, R. and Small, S. (1998). Two distinct mechanisms for differential
positioning of gene expression borders involving the Drosophila gap protein giant.
Development 125, 3765-74.

Yuh, C. H., Bolouri, H. and Davidson, E. H. (1998). Genomic cis-regulatory logic:
experimental and computational analysis of a sea urchin gene. Science 279, 1896-902.

Zhang, H. and Levine, M. (1999). Groucho and dCtBP mediate separate pathways of
transcriptional repression in the Drosophila embryo. Proc Natl Acad Sci U S A 96, 535-
40.

137

Zhou, J., Zwicker, J., Szymanski, P., Levine, M. and Tjian, R. (1998). TAFII
mutations disrupt Dorsal activation in the Drosophila embryo. Proc Natl Acad Sci U S A
95, 13483-8.

138

Chapter IV

Conclusions and Future Directions

Enhancers: Molecular Computers or Information displays- Biochemical implications

Our analysis of structure-function relationships within enhancers has
demonstrated that there are at least two models for enhancer function- the enhanceosome
model and the ‘Information Display’ or ‘Billboard’ model. The differences between the
enhanceosome and billboard enhancers reflect the differences in the mechanism by which
regulatory information incident on the enhancer element is converted into transcriptional

output.

In the enhanceosome model, the enhancer functions like a ‘molecular computer’
that is solely responsible for directly sensing and computing both the quality and quantity
of regulatory signals (transcription factor inputs), while the core promoter element simply
responds to the instructions generated by these molecular logic circuits. Although
demonstrated for only a handful of elements, this conceptual formulation and
terminology is rife in most contemporary literature (Flores-Saaib and Courey, 2000;
Halfon et al., 2002; Halfon and Michelson, 2002; Xu et al., 2000; Yuh et al., 1998). The
best studied example of computer-like integration of signaling inputs to provide highly
specific gene activation is the human interferon-l3 (IFN-B) enhancer, which drives
transcription of the IFN-l3 gene in response to viral infection (Struhl, 2001). The
regulatory proteins (activators and architectural proteins) assemble through cooperative

interactions into a well-defined nucleoprotein complex called the ‘enhanceosome’.

139

Assembly of the enhanceosome is absolutely essential for the transcription of the IFN-B
gene in response to viral infection in cells. Individual activators bound to their sites in the
enhancer do not by themselves stimulate transcription; enhancers that contain multiple
copies of any one activator site are less inducible by the virus and respond non-
specifically to other signals. These observations indicate that the higher order
nucleoprotein complex or enhanceosome provides a stereospecific interface for
interaction with the basal transcriptional machinery, possibly engaging several
components of the basal machinery simultaneously to effect synergistic and highly
specific gene activation (Cai et al., 1996; Carey et al., 1990; Chi et al., 1995). In this
structured element, the presence of each transcription factor binding site and its precise
arrangement within the regulatory element are critical in dictating the output of the
element (Kim and Maniatis, 1997; Munshi et al., 2001; Thanos and Maniatis, 1995). The
enhanceosome therefore imposes the restriction that the target gene would be activated
only under very precise conditions, when all the regulatory proteins are present and are
functionally active. Thus, the IFN-B enhanceosome functions as a precise on/off binary

transcriptional switch in response to the appropriate stimulus (Struhl, 2001).

In my study (described in Chapter 2) I find clear evidence for the first time that
the computational functions usually ascribed to the enhancer itself are actually shared
with the basal transcriptional machinery and suggest instead, that the basal machinery
plays an active, rather than passive role in interpreting transcriptional signals from
enhancers. Using synthetic compact enhancer elements I demonstrated that closely

spaced factor binding sites, situated within compact cis-elements, are actually able to be

140

independently "read off’ by the transcriptional machinery. In such an enhancer, exact
binding site locations are not critical and each bound factor contributes to the overall
transcriptional output. Thus, the removal or addition of individual binding sites might
attenuate the overall level of the transcriptional output, but will not completely abolish
enhancer function, as is the case with the enhanceosome element. The ‘on or of binary
switch decisions that appear to be transmitted from the enhancer to the basal machinery
may actually result from the basal transcriptional machinery reading a series of redundant
signals encoded within the enhancer. Thus, instead of acting in a concerted, all- or-
nothing computational fashion, the enhancer regulates target gene expression through
successive dialogues/interactions with the basal transcriptional machinery. This feature
provides a novel view of biochemical mechanisms of enhancer-promoter interactions,
suggesting that activators are not all making simultaneous contacts with the basal
machinery but that discrete portions of the enhancer can be successively “sampled” by

the basal transcriptional machinery from different directions.

What is the nature of the interaction(s) between the enhancer bound regulatory
proteins and the cognate promoter? One idea is that the ‘effect’ produced by the
transcription factors bound at the enhancer spreads along the DNA to the target promoter.
Just as the Sir repressosome generates a transcriptionally silent chromatin structure that is
able to spread along the chromatin fiber, it is possible that the enhancer bound factors
nucleate an open or active chromosomal state that spreads to the core promoter allowing
the binding of the basal transcriptional machinery. Alternatively, the enhancer-bound

proteins and their associated cofactors establish a productive interaction(s) with the

141

cognate promoter by looping out the DNA between the enhancer and promoter such that
the regulatory proteins at the enhancer make multiple direct contacts with the basal
promoter elements. The information display model clearly favors the latter, where
multiple successive or simultaneous interactions between promoter and enhancer

subelements lead to a transcriptional readout.

If the enhancer subelements are indeed making direct contacts with basal
promoter elements, then what are the timescales for these interactions? Again, the
billboard model would favor transient interactions versus stable enhancer-promoter
complexes. The dynamic nature of the interactions would allow different enhancer
subelements to communicate the regulatory information contained within them to the
basal transcriptional machinery. Such a situation is of critical importance in the case of
modular cis-regulatory regions controlling the expression of developmental genes where
multiple independently acting enhancer elements either turn the gene off or on at the

same time.

Multiple, temporal ‘snapshots’ of the precise physical architecture of enhancer-
promoter complexes would provide definitive information about the molecular aspects of
the interactions between cis-regulatory element and the core promoter. However, current
structural techniques have not yet succeeded in capturing the complexity of multiple
activators and repressors interacting with the basal machinery on a chromatinized gene.
An important consideration towards attempting to solve a ‘crystal structure’ for the entire

transcriptional complex, is whether regulatory proteins are stably bound to their cognate

142

sites with the enhancer element or are in constant flux between the bound state and
unbound state. It is conceivable that such a ‘snapshot’ of an enhancer-promoter complex
is possible in the case of the enhanceosome, which activates expression of the linked
gene only when all the regulatory proteins are present and bound in the correct

conformation to their sites within the element.

Modularity within the ‘Billboard’ enhancer substructure

I have demonstrated that the ‘billboard’ enhancer is capable of simultaneously
representing both the states of activation and repression at the same time and in the same
nucleus. In this case, the independent information units might consist of one or a few
factor binding sites, the enhancer substructure. On a larger scale, this ability to display
contrasting information at the same time is similar to the information content of the
multiple modular stripe enhancers that regulate even-skipped (Goto et al., 1989; Harding
et al., 1989; Small et al., 1992). In this case a separate enhancer module within a complex
promoter can be repressed while an adjacent enhancer module that regulates the same
promoter can be activated, at the same time and in the same nucleus. Modular
organization of cis—regulatory regions is typical of many developmentally regulated genes
in higher eukaryotes and allows the gene to be expressed in temporally and spatially
complex expression patterns. The modularity of cis-regulatory regions of genes also
facilitates evolution because individual elements can evolve to either acquire additional
fiinction(s) or lose function(s) independently without destroying the primary function of

the gene. In a similar vein, on a local level in the ‘billboard’ model a single enhancer can

143

be thought to contain a similar modular organization where separate modules signal

independently to the basal promoter.

Enhanceosome and ‘Billboard’ enhancer — two extremes of a continuum

The ‘enhanceosome’ and the ‘billboard’ enhancer are likely to represent two
extremes of a continuum that describes the range of function of cis-regulatory elements.
Most enhancers are likely to lie within this spectrum, exhibiting properties that are more
like one or the other. A number of experimental studies have revealed enhancers with
properties of both enhanceosomes and billboards. The Drosophila zen ventral repression
region (VRR) is approximately 600 bp in length and is responsible for both the repression
of zen expression in the ventral region of the embryo and for its expression in the dorsal
region. The VRR contains important binding sites for the Dorsal protein (Doyle et al.,
1989; Jiang et al., 1992), which normally activates transcription. On this element
however, adjacent AT-rich binding sites bind the Cut and Dead Ringer factors, which in
turn allow the recruitment of the Groucho corepressor that converts Dorsal fiom an
activator to a repressor. Exact spacing between the AT-rich sites and the Dorsal sites is
required for Dorsal to function as a repressor. A 180-bp region from the zen VRR
containing three AT-rich sites, ATl to AT3, and three Dorsal binding sites, (111 to dl3,
was altered by the insertion of a 5-bp spacer between the AT2 and d12 sites. This change
resulted in the elimination of the repression activity of the element, while the insertion of
a 10-bp spacer restored repression (Cai et al., 1996; Valentine et al., 1998). This result

implies that the correct stereospecific positioning of Dorsal relative to AT2 bound

144

proteins is critical for repression, a feature suggesting a high degree of cooperativity
characteristic of the enhanceosome model. At the same time, a separate portion of the 600
bp zen VRR binds transcriptional activators responsible for transcription of zen in the
dorsal regions of the embryo. This regulation is independent of the repression module and
can fimction autonomously, reminiscent of the ‘Billboard’ model. Thus, the zen VRR
exhibits properties that are enhanceosome-like and at the same time functions like an

information display.

Another example of an enhancer with properties of both the enhanceosome and
billboard models is the element that is involved in the transcriptional control of the
Drosophila terminal gap gene huckebein (hkb) that depends on Torso (Tor) receptor
tyrosine kinase (RTK) signaling and the Rel/NFkappaB homolog Dorsal. Analysis of the
interplay between Dorsal, Groucho and Dead ringer on the hkb enhancer demonstrates
that when the Dorsal and Dead Ringer binding sites are separated by a distance of ~ 90
bp, Groucho recruited by Dead ringer blocks rather than converts Dorsal activator
function, which is in contrast to what is observed in case of the repressosome described
above. In this case, Groucho quenches or blocks the Dorsal activator from signaling to
the basal machinery. Removing the binding site for Dead ringer allows Dorsal to activate
transcription, indicating that the element is essentially bipartite and constitutes two
independently acting subelements. Reducing the distance between Dorsal and Dri binding
sites, however, switches Dorsal into a Gro-dependent repressor resulting in a
repressosome type of activity that overrides activation of transcription (Hader et al.,

2000)

145

In the enhanceosome model transcription factor binding sites are highly
constrained, while the billboard enhancer affords more plasticity in the arrangement of
transcription factor binding sites. A defective eve stripe 2 enhancer lacking a crucial
Bicoid binding site can be complemented with a high affinity Bicoid binding site inserted
at a new location, therefore exhibiting structural flexibility that is a characteristic of the
billboard enhancer. However, there are some constraints on enhancer organization;
insertion of the same high affinity Bicoid sequence at a different location within the
element results in only partial restoration of the defective stripe (Amosti et al., 1996a). In
addition, several attempts to construct an artificial stripe 2 enhancer element by simply
clustering binding sites for Bicoid and Hunchback activators that regulate the native
element have been unsuccessfirl, indicating that there may be some spacing or orientation
constraints between the factor binding sites. Thus, in this case although there is some
flexibility in the arrangement of transcription factor binding sites, it is clearly not without
any functional constraints. As described before, independent information units within the
information display type of enhancer might consist of one or a few factor binding sites. It
is conceivable that within an independently acting information unit functional constraints
between factor binding sites exist, which are the ‘grammar’ rules for enhancer

substructure as described in Chapter 3.

At the other end of the spectrum is the transcriptional regulation of cut expression
in the wing imaginal disc of Drosophila melanogaster. The transcription factor Scalloped
potentiates the transcriptional activation of this gene throughout the disc. Specific Notch

signaling along the dorsal/ventral boundary of the disc activates the Suppressor of

146

Hairless [Su(H)] transcription factor, and Su(H) together with Scalloped bind to the cut
enhancer to activate Cut expression specifically along the dorsal/ventral boundary of the
disc. A simple cluster of Scalloped and Su(H) binding sites without regard to exact
distance and arrangement, is sufficient to recapitulate this pattern of expression (Guss et
al., 2001). In a similar manner, a simple cluster of Dorsal and Twist binding sites can
recapitulate the expression pattern driven by endogenous regulatory elements that bind
these transcription factors (Szymanski and Levine, 1995). These observations suggest
that some cis-regulatory elements do not require the formation of a higher order
nucleoprotein complex to regulate gene expression, and are free of detectable constraints
in their internal design. The lack of ‘grammar’ rules might reflect the nature and the
mechanism(s) of function of the transcription factors that bind to such elements. Clearly,
elements that bind short-range repressors will exhibit some constraints in internal design

that are dictated by the way in which these proteins function.

A number of studies (Bouallaga et al., 2000; Ellwood et al., 2000; Read et al.,
1994) suggest that enhancer elements designed for sudden inducibility, such as those
involved in immune response, might tend to function as enhanceosomes. Thus, it is likely
that eukaryotic organisms use ‘billboard’ enhancers to achieve diversity in gene
expression patterns and evolutionary flexibility, and enhanceosomes to achieve high

levels of specific gene activation (Struhl, 2001).

147

‘Grammar’ rules for repressor function: Insights into cis-regulatory evolutionary
dynamics

Cis-regulatory DNA function may evolve either by de novo evolution of cis-
regulatory elements via changes in nonfunctional DNA or through the evolution of cis-
regulatory elements from preexisting functional elements. These modifications include
duplications and DNA rearrangements of existing functional elements, and the loss or
gain of binding sites through changes in nucleotide sequence. The ability of regulatory
DNA to evolve is greatly facilitated by the modularity of cis-regulatory elements.
Individual elements can act, and therefore evolve, independently of others (Fink and
Scandalios, 2002). The typical organization of the cis-regulatory regions of
developmental regulatory genes, composed of many independent elements, is tacit
evidence for the expansion and diversification of cis-regulatory systems in evolution. In a
similar manner, the independently acting subelements of the billboard enhancer have a
greater capacity for evolutionary flexibility than the nonmodular element such as the
enhanceosome, where every component (transcription factor binding site) is optimally

linked to every other component and is thus effectively frozen.

To understand the role played by cis-regulatory DNA in evolution it is important
to appreciate two common features of cis-acting elements. First, cis-regulatory elements
are regulated by multiple distinct transcription factors that interact with each other, with
cofactors, with chromatin and with components of the basal transcriptional machinery.
Second, the spatial relationship of binding sites for transcription factors within cis-

elements can be critically important, as with enhanceosomes, but also for billboard

148

enhancers that still are subject to contextual grammar. Thus, a meaningful evolutionary
analysis of a cis-regulatory sequence requires detailed information about the locations of
transcription factor binding sites within a sequence, the firnctional specificity of the
binding sequences, and the spatial requirements for their interactions. Without a deep
understanding of the structure-function correlations that exist with cis-regulatory
elements it is impossible to discern functional consequences of evolutionary changes in
regulatory sequence. Using defined synthetic enhancer elements I have systematically
deduced the contextual parameters in terms of the number, affinity, relative spacing and
arrangements of binding sites that define the enhancer ‘grammar’ required for the
function of an important class of regulators, the short-range transcriptional repressors
(described in Chapter 3). Although previous studies had provided evidence that some
parameters might be important, my work with defined elements has provided a
quantitative look at these factors. This study has important ramifications on the field of
regulatory evolution as it provides quantitative measures to further assess the

evolutionary dynamics of enhancers regulated by short-range repressors.

From our limited perspective enhancer elements appeared to have haphazard
assemblages of regulatory factor binding sites that were adequate to get the job done with
little logic to their intemal organization. However, I have demonstrated (Chapter2 and 3),
that there are internal design principles that direct how different features of transcription
factor binding sites provide varied transcriptional outputs. Discrete regions of the
billboard enhancer are ‘read’ by the basal transcriptional machinery through multiple

successive or simultaneous interactions and the sum of these multiple interactions

149

dictates the transcriptional output of the element. Therefore, the ‘billboard’ enhancer can
display contrasting regulatory information, such that both the activated and repressed
states are represented within the same enhancer element at the same time and in the same
nucleus. This would appear to be in contrast to the action of natural enhancers, which are
seen to firnction as binary on or off transcriptional switches. Resolving this apparent
paradox, I demonstrated that the billboard enhancer can provide uniform information in
the manner of a binary transcriptional switch if the stoichiometry of repressor and
activator binding sites is adjusted to ensure that all possible enhancer subelements
provide the same information. Thus, using simple well defined synthetic enhancer
elements we were able to highlight that stoichiometry between activators and repressors
is one of design principles of cis-acting elements that allow them to function as binary

transcriptional switches.

Further manipulation of these composite elements to alter the number of activator
and repressor binding sites, relative affinities, spacing and distribution of these binding
sites allowed us to define other contextual parameters that dictate repression
effectiveness. First, we find that although a close proximity between the short-range
repressors and their targets is a critical factor in dictating repression effectiveness, it is
not sufficient. I found that the general notion that short-range repressors block the activity
of all protein complexes within a 100 bp is an oversimplification and that in some
contexts (greater number of activators or high affinity of activator binding sites)
activators can escape repression even when bound close to short-range repressors.

Second, the effectiveness of repression correlates with the number and affinity of

150

activator binding sites. Although stronger binding of activators correlates with greater
transcriptional output, I was able to demonstrate that repression is not inversely
proportional to activator strength, as even weak activators with good binding sites are not
repressed. A third general parameter is the importance of proper distribution of activator
and repressor binding sites for effective repression. The ‘Billboard’ model of enhancer
action, which suggests that subelements of a cis-regulatory enhancer can be
independently sampled by the basal machinery, can explain this finding. Effective
repression of an entire enhancer would thus require distributed pattern of repressors that
would interfere with the activity of each subelement. Fourth, short-range repressors do
not exhibit activator specificity, indicating that the repression mechanism may be targeted
to the binding potential of the activator rather than its activation domain. These
contextual parameters constitute a ‘grammar’ of repressor function. The quantitative
description of these elements will facilitate the identification of functional changes in cis—

regulatory sequences that are the raw material for morphological evolution.

Interspecific sequence comparisons of noncoding cis-regulatory regions have
shown that inspite of selective constraints, the structure and sequences of cis-elements
change over time, sometimes dramatically so, even in cases where expression patterns are
conserved (Hardison et al., 1997; Loots et al., 2000; Ludwig et al., 2000; Ludwig and
Kreitman, 1995; Ludwig et al., 1998). Homologous sequences to the D. melanogaster
eve stripe 2 element have been isolated from a host of other Drosophila species. These
elements all drive an accurately positioned stripe of reporter gene expression in the eve

stripe 2 domain in D. melanogaster, although considerable sequence divergence occurs

151

among the elements (Figure IV-l). Interestingly, some of the sequence changes in the
other Drosophila species abolish sites that are known to be essential in the D.
melanogaster element. For instance, the third Bicoid binding site (bcd-3) and the first
Hunchback binding site found in D. melanogaster lack counterparts in D. pseudobscura,
D. erecta, and D. yakuba. This analysis suggests that the bed-3 site may be a new site in
D. melanogaster (Kreitman, 1996; Ludwig et al., 2000; Ludwig and Kreitman, 1995;
Ludwig et al., 1998). In light of the contextual rules defined for the short-range
repressors, we predict that the D. melanogaster stripe 2 element requires higher levels of
repression to compensate for the additional activation potential. Indeed, sequence
analysis of the homologous elements demonstrates that a 30 bp deletion has moved a
nearby Giant site closer to the site of this Bicoid activator, presumably compensation for
the increase in activation potential. A theoretical framework for interpreting the
evolutionary changes in any cis-regulatory elements, although still in its infancy, will be
greatly facilitated by the empirical determination of such spatial rules for other classes of

transcriptional regulators.

‘Grammar ’ rules for repressor function: Improved computational identification of
functional cis-regulatory modules

Recently, whole genome sequence assemblies have become available providing a
powerful foundation to identify and analyze cis-regulatory module function and
organization on a global scale. Two current approaches to identify candidate regulatory
regions from genomic data are computational methods that look for clusters of

transcription factor binding sites (Berman et al., 2002; Markstein and Levine, 2002;

152

  

 

 

Km Kr5 9f? {‘9 Kr4 .0119 Kr3 l(r2 Krl
BcdS Bcd4 Bcd3 Bch H.233 13ch ‘19.. ......”

B

Conservation of Transcription Factor Binding Sites in Eve Stripe 2 Enhancer

 

 

Sites ,

l Kr Bcd Hb th

Species 123456 12345 123 123

D.simulans PPPPPP PPSSP SPP PPS

D.yakuba PPSPPP SPWSP SSP PSS

D.erecta PPSPPP SPSSP SSP PSS

D.pseudobscuru SSSSPP SSASP APS SWS

D.picticomis WSWSPP WSASP ASW WWW
Conservation
P=perfect

S = strong(l—2 changes)
W = weak (3 or more changes)
A = absent

Figure IV-l: Evolutionary dynamics of transcription factor binding sites in a
conserved cis-regulatory element.

(A) Binding sites for the Kruppel (Kr), Giant (Gt), Bicoid (Bed), and Hunchback (Hb)
proteins in the 670 bp D. melanogaster eve stripe 2 cis-regulatory element are shown. (B)
The conserved binding sites in five different Drosophila species are tabulated. The
degree of sequence conservation within each site is indicated (P, S, W, A). Note that
certain sites such as Kr5, Kr6, and BcdS are perfectly conserved, whereas other sites such

as Hbl and Bcd3 are missing from certain species. (Figure adapted from Carroll, SB,
2001)

153

Markstein et al., 2002; Rebeiz et al., 2002) and phylogenetic comparisons that identify
evolutionarily conserved sequences (Bergman and Kreitman, 2001; Maier et al., 1990).
One main difficulty with the output from computational searches for clusters of
transcription factor sites is the large number of false-positive and false-negative results.
The short length and degenerate nature of transcription-factor-binding sites account for
some of these misleading predictions. Furthermore, genes are rarely controlled by a
single transcription factor and accumulating evidence suggests that specific combinations
of transcription factors are required to achieve the complex differential expression of
genes in higher organisms. In addition, as I have demonstrated in Chapters 2 and 3, the
pr0per spatial organization of transcription factor binding sites within the element may be
critical for their proper biological function. The structural basis for combinatorial
regulation exists in the specific organization of multiple transcription factor binding sites.
Transcription factors interact with each other or with cofactors to mediate their function,
therefore spatial correlations/constraints between their binding sites need to be taken into
account. Thus, in many cases where cis-regulatory modules predicted by computational
methods appear suitable, something in the arrangement of sites (or a wider context)
renders the cis-element non-firnctional. In order to achieve better predictions and
eliminate false-positive and false—negative results, computational methods should
include, in addition to binding site density and relative affinities, parameters for spacing
and position of binding motifs within cis-elements in the search algorithms. These
parameters are difficult to predict accurately a priori. Thus, computational approaches

will benefit from the availability of at least one well-defined representative of a particular

154

regulatory network to serve as a starting paradigm. An example of how this will actually

work is described on p. 170.

At the other end of the spectrum, researchers are studying cis-regulatory element
function using extensive deletion and mutational analyses and experimentally
determining the consequence of these changes on expression pattern in vivo. While such
an approach is useful for discovering the firnctional interactions between the different
components of a cis-regulatory element, it is constrained in its focus to a specific cis
element. Furthermore, the analyses of complex endogenous regulatory elements is often
confounded by the lack of information about the internal organization of the element with
regard to the number, relative affinities, order, spacing and orientation of binding sites,
and even the identity of the transacting regulatory proteins. Because of this complexity of
cis-regulatory regions, the contributions of individual parameters to the overall
transcriptional output cannot be accurately quantified and multiple experiments are

required to gain even a rough overview of how an expression pattern is generated.

My study seeks to find a bridge between the two approaches by using simple
well-defined synthetic enhancer elements containing binding sites for transcriptional
activators and repressors. This approach allowed us to systematically identify the general
operating principles in terms of cis-element architecture for the short-range
transcriptional repressors. We defined a set of ‘grammar’ rules in terms of the number,
relative affinities, spacing and arrangement of binding sites that are required for proper

regulation by short-range repressors. Since this approach allowed us to systematically

155

alter one parameter at a time, we were able to dissect with greater accuracy the
contribution of each parameter to the transcriptional output of the element. The empirical
determination of these elements and the relative contribution of each element to repressor
function will facilitate bioinforrnatic analysis of novel gene regulatory sequences. For
example, suppose that a computational search for novel cis-elements regulated by Giant
based on clustering of binding sites for transcription factors that are known to act together
with, it identifies a putative regulatory module. Given the relative number, affinities,
spacing and distribution of repressor and activator sites within the module we would be

better able to predict the possibility of Giant regulating that element.

FUTURE DIRECTIONS

Molecular mechanism (s) of sh art-range transcriptional repression

Gene-specific repression of transcription plays a central role in gene regulation.
This is true not only for the spatial control of gene activity in development, during which
boundaries of gene expression are determined by the spatially restricted localization or
activity of transcriptional repressors (Mannervik and Levine, 1999; Mannervik et al.,
1999) but also true for the control of gene expression by extracellular signals, in which
genes are often maintained in an off state by repressor proteins until signal transduction

alleviates the repression (Barolo and Posakony, 2002; Roose and Clevers, 1999).

156

As introduced in Chapter 1, studies on transcriptional repression in the
Drosophila embryo suggest that there may be two basic forms of repression, long-range
and short-range repression (Amosti et al., 1996b; Barolo and Levine, 1997; Gray and
Levine, 1996; Gray et al., 1994; Hewitt et al., 1999; Keller et al., 2000). Short-range
repression represents a flexible form of gene regulation exhibiting either enhancer- or
promoter- specific effects depending on the position of the repressor binding sites. This
flexibility contrasts with long-range repressors, which can block multiple enhancers over
distances of several kilobases regardless of location within a gene complex. The different
activities of each class probably reflect distinct mechanisms employed, however, the

molecular events that differentiate one from the other are not well understood.

Two modes of short-range repression have been proposed that are consistent with
the activity of these proteins seen in vivo: Short-range repressors might “quench”
activator proteins locally, within a short distance of their binding site by displacing them
from the DNA through chromatin structure modification, or by preventing them from
contacting their target in the basal transcriptional machinery; or the repressors might
directly contact some component of the basal transcriptional machinery, but only when
brought to the promoter by a closely linked activator (hitchhiking) or when bound near
the start site of transcription. Studies on the Drosophila short-range repressor Snail
prompted the proposal that quenching might involve direct protein-protein interactions
between repressors and upstream activators (Gray et al., 1994). This type of mechanism
has been proposed for the Drosophila Kriippel protein where transient transfection assays

have suggested that Kriippel can selectively repress transcription activated by a

157

glutamine-rich activator but not by an acidic activator (Licht et al., 1993). Further studies
identified two evolutionarily conserved repression domains in the Kriippel protein that
differ in activator specificity (Hanna-Rose et al., 1997). Knirps and Giant have been
shown to repress heterologous activators (Amosti et al., 1996a; Amosti et al., 1996b;
Hewitt et al., 1999; Kulkami, 2003), and therefore do not appear to be “dedicated”

repressors.

As discussed in Chapter 1, a common property of short-range transcriptional
repressors is their interaction with the evolutionarily conserved corepressor CtBP (C-
terminal Binding Protein) through short PXDLS peptide motifs found in the
transcriptional repressors (and in other interacting proteins). CtBP has been shown to
interact with chromatin modifying factors, including Histone deacetylases (HDACI and
HDAC2) and Histone methyltransferases (Chinnadurai, 2002; Chinnadurai, 2003; Shi et
al., 2003; Subramanian and Chinnadurai, 2003; Sundqvist et al., 1998). Thus, the major
activity of CtBP might be to serve as a bridging molecule, recruiting chromatin or
transcription factor modifying enzymes to the promoter. CtBP has also been shown to
have a weak dehydrogenase activity in vitro, but the role of this activity if any in
repression is yet unknown (Balasubramanian et al., 2003; Kumar et al., 2002).
Alternatively, CtBP might possess an enzymatic activity similar to the NAD-dependent
Sir 2 repressor protein, which requires NAD to mediate deacetylation of histone proteins

(Marmorstein, 2002).

158

Drosophila short-range repressors also possess CtBP-independent repression
activities and the dCtBP requirement can vary on a gene-to-gene as well as on an
enhancer-to-enhancer basis (Keller et al., 2000; Nibu et al., 2003; Strunk et al., 2001).
Multiple repression activities may allow for quantitative or qualitative effects on gene
expression and may be context-dependent. Thus, qualitatively, a repressor may function
selectively in a tissue-specific manner or in a activator-specific manner (Postigo and
Dean, 1999) or in different promoter contexts (Lunyak et al., 2002). Quantitatively, dual

activities may increase the overall level of repression.

My analysis of transgenes with varying number and affinities of activator binding
sites indicate a mechanism of short-range repression that is highly dependent on a critical
ratio of enhancer bound activators and repressors. However, I also find that stoichiometry
by itself is insufficient for proper regulation by the Giant short-range repressor and subtle
(< 40 bp) changes in the spacing from the nearest activators sharply attenuates repression
effectiveness. This study clearly demonstrated for the first time that three parameters, the
number, affinity and relative distance of activator binding sites, contribute equally and
are critical in order for Giant to mediate repression. These contextual parameters
together with the observation that Giant’s ability to repress does not depend on the nature
of activation domains suggest that short-range repressors might firnction by preventing
access to DNA by activators perhaps through chromatin remodeling on a very local scale.
As indicated previously, biochemical purification of CtBP complexes from HeLA cells
has shown that such complexes contain chromatin modifying factors, including Histone

deacetylases (HDACI and HDAC2) (Shi et al., 2003). Preliminary biochemical studies in

159

our laboratory (Struffi, unpublished) also indicate that the Knirps protein interacts with
the Drosophila homolog of HDACl, de3, in embryos. Recruitment of de3 has been
shown to create a localized domain of deacetylated histones that extends one to two
nucleosomes to either side of the recruitment site (Kadosh and Struhl, 1998a; Kadosh and
Struhl, 1998b; Rundlett et al., 1998). In yeast, it has been demonstrated that de3-
dependent repression occurs only when the recruitment site is located at a distance of
<200 bp relative to the region containing the activator binding site and core promoter
elements, which is also in accord with the size of the domain of histone deacetylation.
Mapping experiments with target promoters indicated that histone deacetylation peaks at
the de3 recruitment site and extends 200 to 300 bp in both directions (Deckert and
Struhl, 2002). Our hypothesis is that Giant and other short-range repressors may prevent
activator binding by remodeling chromatin on a local scale to establish a repressive
chromatin domain. High affinity or greater number of activator binding sites may allow

the activator to bind in spite of such repressive chromatin remodeling.

Although we have defined the contextual elements of cis-regulatory enhancers
that allow short-range repressors to mediate repression effectively, we do not yet
understand the physical and biochemical changes in promoter complexes that accompany
short-range repression. Chromatin Irnmunoprecipitation (ChIP) assays in Drosophila
embryos using the simple genetic switch elements defined in this study will facilitate our
understanding of the molecular mechanisms of repression employed at developmentally
regulated genes. These experiments can be designed to directly test mechanistic questions

about the nature of short-range transcriptional repression, specifically whether repression

160

blocks activator access to the DNA, alters chromatin structure, or whether changes in the
composition or presence of the basal machinery is observed. The use of the synthetic
genetic elements defined in this study for ChIP analysis offers two advantages. First, the
endogenous targets of short-range repressors such as the pair-rule gene even-skipped, are
expressed in discrete domains in the early blastoderm embryo and therefore, are not
useful targets for ChIP experiments as such assays would result in high background
noise. Using transgenes as those described in Chapter 2 and 3 of this study, would allow
one to turn the expression of the reporter transgene on or off throughout the embryo using
inducible and ubiquitous expression of the activators and repressors. Second, the use of
enhancer elements where the exact identity of all the components and their spatial
positions within the element are known, willallow the determination of the molecular
and biochemical changes that accompany small changes in the internal design (in terms
of activator-repressor stoichiometry, relative affinities, spacing and positioning of
binding sites) of the enhancer that either result in repression or the lack thereof. Such
studies would provide quantitative data that could facilitate the predictive modeling

studies described in the last section.

The biochemical identification and analysis of protein complexes associated with
short-range transcriptional repressors or their cofactor, CtBP, from Drosophila extracts

will also provide insights on the molecular aspects of short-range repression.

The availability of both short- and long-range repressors adds another layer of

flexibility to gene regulation and may serve specific gene regulatory needs. Long-range

161

repression provides the possibility of shutting down an entire locus regardless of the
number of separate regulatory modules that control the activity of that locus. This kind of
repression has often been referred to as silencing because an entire chromosomal locus is
inactivated. On the other hand, short-range repression provides a way to control the
activity of one enhancer without interfering with the activity of others and is useful for
creating precise, tunable repression required to establish intricate patterns of expression.
In addition to having a longer range of action, I find that the long-range repressor Hairy is
more potent than Giant even on a local scale and can block the activity of a greater

number of activators as compared to Giant.

The mechanisms by which short-range and long-range Drosophila repressors
inhibit transcription are poorly understood, although one model of repression in the
embryo suggests that the short-range/long-range distinction results from the recruitment
of distinct cofactors (Nibu et al., 1998a; Nibu et al., 1998b; Zhang and Levine, 1999).
Short-range repressors recruit the corepressor CtBP to mediate repression whereas long-
range repressors have been shown to interact with the Groucho corepressor (Barolo and
Levine, 1997; Chen and Courey, 2000; Fisher and Caudy, 1998; Mannervik and Levine,
1999; Mannervik et al., 1999; Poortinga et al., 1998). One model for Groucho mediated
long-range repression is through the recruitment of HDACs by Groucho resulting in the
production of a large transcriptionally silent chromosomal domain. Just as the Sir
repressosome generates a transcriptionally silent chromatin structure that is able to spread
along the chromatin fiber, it has been proposed that the Groucho repressosome nucleates a

silent chromosomal state and spreads to mediate long-range repression (Brantjes et al.,

162

2001; Chen et al., 1999; Flores-Saaib and Courey, 2000). However, I find that the
repression mediated by Hairy is transient, similar to the activity of the short-range
repressors and is easily reversed as the embryo develops. Thus, if Hairy mediates
transcriptional repression by creating a repressive heterochromatin domain that can
spread over large distances, then the nature of this heterochromatin must be such that it

can be easily modified back to a more open conformation.

Another similarity between short-range and long-range repressors is that like
Groucho, a number of studies have suggested that CtBP may function, at least in part, by
recruiting histone deacetylases (Criqui-Filipe et al., 1999; Sundqvist et al., 1998). Thus, if
both long- and short-range corepressors function through histone deacetylation, why is
Hairy is more potent with a longer range of action than Giant? There are a number of
possibilities. First, as mentioned previously, long-range corepressors may have the ability
to spread along the template recruiting histone deacetylases and/or other chromatin
modifying activities to a large domain, but short-range repressors may lack the capacity to
spread. Altematively, the differences between long- and short-range corepressors could
relate to the different pr0perties of different histone deacetylases. Groucho has thus far
only been found to bind class I histone deacetylases, whereas CtBP appears to bind both
class I and class II histone deacetylases (Bertos et al., 2001). Perhaps the different
repertoires of histone deacetylases recruited by different corepressors result in different
histone acetylation patterns in the surrounding chromatin. The synthetic enhancer

elements that I have defined in my study can thus serve as useful tools to compare the

163

biochemical activities of these two classes of transcriptional regulators in order to

distinguish between the molecular mechanisms of short-range and long-range repressors.

Next generation Biology: Predictive modeling of differential gene expression as a

function of time and space.

Although computational methodologies have so far focused on the simple
identification of cis-regulatory elements involved in gene regulation, a more challenging
goal has been set to predict from the sequence of a cis-element gene expression patterns
over time and space. While some contextual parameters such as number, relative
affinities, spacing and arrangements of transcription factor binding sites can be tested by
manipulating these parameters individually or in combination, in designing the
experimental cis-element, in vivo gradients of transcription factors, such as the gap
repressor gradient will provide the opportunity to test different concentrations of
transacting factors on the output of a given cis-element. Crucial to the formulation of a
quantitative predictive model for differential gene expression is the ability to quantitate
regulation by the ‘response characteristics’, that is the level of gene expression as a
function of the concentrations of transcription factors (Bower, 2001). We will require the
development of more quantitative approaches to interpret very subtle differences in gene
expression seen with small changes in the sequence of the cis-element. For example, in
order to be able to build a robust predictive tool, quantitative methods to detect subtle
changes in transcriptional readout caused by 10 bp changes in the spacing of transcription

factors sites within a cis-element will have to be established.

164

The first computer experiments fitting analog gene regulation models to real
expression data, were used to understand the network of gap genes expressed in bands (or
domains) along the anterior-posterior axis of the very early (syncitial blastoderm) of the
Drosophila melanogaster embryo (Reinitz and Sharp, 1995). Positional information
along the A-P axis of the syncitial blastoderm is encoded in a succession of different
ways during development. At first the main encoding is a roughly exponential gradient of
Bicoid protein imposed by the mother fly, along with the maternal Hunchback
expression. These provide gene regulation inputs to the gap gene: giant, knirps, lm'ippel,
tailless, and hunchback. These each establish one or two broad domains of expression
along the A-P axis. The gap genes then serve as network inputs to the pair-rule genes,
including even-skipped, which establish narrow, precise stripes of expression along the
A-P length of the embryo. The predictive model of Reinitz and colleagues was concerned
with the establishment of the broad gap gene domains (excluding the extreme ends of the
Anterior-Posterior axis) from maternally supplied initial conditions, by a gene regulation
network in which all gap genes interact with all others and Bicoid provides input to, but
does receive any input from the gap genes. Figure IV-2 shows the experimentally
observed (Figure 1V-2A) and model (Figure IV -2B) fitted curves for gap gene expression
(Reinitz and Sharp, 1995). They are in qualitative agreement, which is the most that
could be expected from the data (expression data and cis-regulatory structure-function)
that was available at the time. The extra dip in Giant expression could not be predicted by
the model, which can be interpreted as an indication of the role of circuit components not
included in the model. Further experimental analyses of this regulatory cascade point to

at least one additional level of complexity that needs to be included in such predictive

165

 

 

 

 

 

 

 

 

Figure IV-2: Data and Model for gap gene circuit.

Horizontal axes are nuclei along lateral from anterior to posterior. Vertical axes are
relative concentrations. (A) Data estimated from immunofluorescence images of gap
gene expression. (B) Output of a circuit model fit to expression data. (Figure adapted
from Reinitz et al., 1995).

166

models of gene regulatory networks, which is a detailed understanding of the cis-
regulatory substructure and its functional significance as we have defined for the short-

range transcriptional repressors (Giant, Knirps and Kriippel) in Chapter 3.

High resolution in vivo and in vitro imaging technologies, combined with large-
scale gene expression assays, are providing vast amounts of quantitative data on gene
regulatory events during development. With genome wide in situ hybridization analysis,
together with global transcription factor binding analysis (Lee et al., 2002; Weinmann
and Famham, 2002; Weinmann et al., 2002) we now have the technology to track the
expression of thousands of genes during the lifetime of a cell, and to trace the interactions
of many of the products of these genes. Thus, by combining quantitative methods to
measure the concentrations of transcription factors over space and time and the
expression of a downstream target gene with extensive input (variations in the internal
design of the cis-element) — output (target gene off or on) patterns we can build a
predictive modeling tool that will allow us not only to identify novel cis- elements that
are regulated by the same suite of transcription factors, in this case the short-range
repressors, but also to predict its expression pattern. As predictions of these models can
be subjected to experimental testing, erroneous predictions maybe overcome by updating
the modeling tool with new empirically obtained corrective structure-function data.
Achieving robust and quantitatively correct predictions is an integrated, iterative process,
with models feeding experimental design, while experimental data in turn feed models.
Thus, the predictive model can be ‘trained’ with input-output data of iteratively

increasing resolution to achieve higher predictive reliability.

167

A simplified example of how exactly such a predictive tool can be used is as
follows: We would like to identify putative targets of short-range transcriptional
repressors in the Drosophila melanogaster genome and predict how these candidate
genes will respond over space and time. Computational algorithms based on binding site
sequence data from transcription factor binding site databases (for example,
TRANSFAC), can be used to scan the genome for clusters of short-range repressor
binding sites together with binding sites for other proteins that are known to work with
them. Combining datasets from a number of large-scale analyses will then help us
prefilter the set of putative regulatory elements obtained. Using gene expression data
from microarray analysis we could ask whether the putative regulatory element is in the
vicinity of a target gene that is known to beregulated by our factors or if they are near
genes that are expressed at approximately the same time as the regulatory factors. Using
data from genome-wide in situ hybridization analysis we could ask if the putative target
genes are expressed in patterns consistent with the pattern of expression of short-range
repressors. We could also probe whether the expression pattern of other factors for which
binding sites occur in the identified regulatory elements correlates with that of our factor
of interest both spatially and temporally and with that of the candidate target gene. Once
the dataset has been scaled down to the most likely candidates, we can compare the
internal organization of the cis-elements with the ‘grammar’ rules defined in this study to
firrther eliminate those that do not seem likely to be regulated by our factor of interest on

the basis of enhancer design.

168

A fully quantitative modeling of animal transcriptional control may appear
unrealistic, but the progress made over the last 30 years has been nothing short of
remarkable. The enormous and growing amount of detailed quantitative information
available from large-scale analysis of the dynamics and localization of transcription
factors, genome-wide analysis of transcription factor binding, improved computational
and phylogenetic methods for identifying cis-regulatory element structure and function,
and genome wide gene expression data may bring us much closer to our goal than
currently seems possible. With a complete understanding of animal transcription systems,
it should be possible to predict biological output directly from regulatory sequence and
the physiological effects on an organism of mutating cis-regulatory sequence or
misexpressing a transcription factor. Ideally,such an understanding of the cis-regulatory
code and the ability to predict the consequences of changes in its sequence, will enable us
to decipher the evolutionary dynamics of transcriptional regulation underlying
morphological complexity and diversity of living organisms. Also, it should allow the
design of novel heterologous promoters that express selected gene products in a specified
group of cells at a controlled level. Such progress would be important both for human
gene therapy and for the development of improved transgenic animals. Quantitative
genome-wide analyses are already changing the way we think of transcriptional control

and will continue to do so in the future.

169

REFERENCES

Arnosti, D. N., Barolo, S., Levine, M. and Small, S. (1996a). The eve stripe 2 enhancer
employs multiple modes of transcriptional synergy. Development 122, 205-14.

Arnosti, D. N., Gray, S., Barolo, S., Zhou, J. and Levine, M. (1996b). The gap protein
knirps mediates both quenching and direct repression in the Drosophila embryo. Embo J
15, 3659-66.

Balasubramanian, P., Zhao, L. J. and Chinnadurai, G. (2003). Nicotinamide adenine
dinucleotide stimulates oligomerization, interaction with adenovirus EIA and an intrinsic
dehydrogenase activity of CtBP. FEBS Lett 537, 157-60.

Barolo, S. and Levine, M. (1997). hairy mediates dominant repression in the Drosophila
embryo. Embo J16, 2883-91.

Barolo, S. and Posakony, J. W. (2002). Three habits of highly effective signaling
pathways: principles of transcriptional control by developmental cell signaling. Genes
Dev 16, 1167-81.

Bergman, C. M. and Kreitman, M. (2001). Analysis of conserved noncoding DNA in
Drosophila reveals similar constraints in intergenic and intronic sequences. Genome Res
11, 1335-45.

Berman, B. P., Nibu, Y., Pfeiffer, B. D., Tomancak, P., Celniker, S. E., Levine, M.,
Rubin, G. M. and Eisen, M. B. (2002). Exploiting transcription factor binding site
clustering to identify cis-regulatory modules involved in pattern formation in the

Drosophila genome. Proc Natl Acad Sci U S A 99, 757-62.

Bertos, N. R., Wang, A. H. and Yang, X. J. (2001). Class II histone deacetylases:
structure, function, and regulation. Biochem Cell Biol 79, 243-52.

Bouallaga, 1., Massicard, S., Yaniv, M. and Thierry, F. (2000). An enhanceosome
containing the Jun B/Fra-2 heterodimer and the HMG-l(Y) architectural protein controls
HPV 18 transcription. EMBO Rep 1, 422-7.

Bower, J. a. B., Hamid. (2001). Computational Modeling of Genetic and Biochemical
Networks: A Bradford Book, The MIT press.

Brantjes, H., Roose, J., van De Wetering, M. and Clevers, H. (2001). All ch HMG
box transcription factors interact with Groucho-related co-repressors. Nucleic Acids Res
29, 1410-9.

Cai, H. N., Arnosti, D. N. and Levine, M. (1996). Long-range repression in the
Drosophila embryo. Proc Natl Acad Sci U S A 93, 9309—14.

Carey, M., Lin, Y. S., Green, M. R. and Ptashne, M. (1990). A mechanism for
synergistic activation of a mammalian gene by GAL4 derivatives. Nature 345, 361-4.

170

Chen, G. and Courey, A. J. (2000). Groucho/TLE family proteins and transcriptional
repression. Gene 249, 1-16.

Chen, G., Fernandez, J., Mische, S. and Courey, A. J. (1999). A functional interaction
between the histone deacetylase de3 and the corepressor groucho in Drosophila
development. Genes Dev 13, 2218-30.

Chi, T., Lieberman, P., Ellwood, K. and Carey, M. (1995). A general mechanism for
transcriptional synergy by eukaryotic activators. Nature 377, 254-7.

Chinnadurai, G. (2002). CtBP, an unconventional transcriptional corepressor in
development and oncogenesis. Mol Cell 9, 213-24.

Chinnadurai, G. (2003). CtBP family proteins: more than transcriptional corepressors.
Bioessays 25, 9-12.

Criqui-Filipe, P., Ducret, C., Maira, S. M. and Wasylyk, B. (1999). Net, a negative
Ras-switchable TCF, contains a second inhibition domain, the CID, that mediates
repression through interactions with CtBP and de-acetylation. Embo J 18, 3392-403.

Deckert, J. and Struhl, K. (2002). Targeted recruitment of de3 histone deacetylase
represses transcription by inhibiting recruitment of Swi/Snf, SAGA, and TATA binding
protein. Mol Cell Biol 22, 6458-70.

Doyle, H. J., Kraut, R. and Levine, M. (1989). Spatial regulation of zerknullt: a dorsal-
ventral patterning gene in Drosophila. Genes Dev 3, 1518-33.

Ellwood, K. B., Yen, Y. M., Johnson, R. C. and Carey, M. (2000). Mechanism for
specificity by HMG-l in enhanceosome assembly. Mol Cell Biol 20, 4359-70.

Fink, R. C. and Scandalios, J. G. (2002). Molecular evolution and structure-function
relationships of the superoxide dismutase gene families in angiosperms and their

relationship to other eukaryotic and prokaryotic superoxide dismutases. Arch Biochem
Biophys 399, 19-36.

Fisher, A. L. and Candy, M. (1998). Groucho proteins: transcriptional corepressors for
specific subsets of DNA-binding transcription factors in vertebrates and invertebrates.
Genes Dev 12, 1931-40.

Flores-Saaib, R. D. and Courey, A. J. (2000). Analysis of Groucho-histone interactions
suggests mechanistic similarities between Groucho- and Tupl-mediated repression.
Nucleic Acids Res 28, 4189-96.

Goto, T., Macdonald, P. and Maniatis, T. (1989). Early and late periodic patterns of

even skipped expression are controlled by distinct regulatory elements that respond to
different spatial cues. Cell 57, 413-22.

171

Gray, S. and Levine, M. (1996). Transcriptional repression in development. Curr Opin
Cell Biol 8, 358-64.

Gray, S., Szymanski, P. and Levine, M. (1994). Short-range repression permits
multiple enhancers to function autonomously within a complex promoter. Genes Dev 8,
1829-38.

Guss, K. A., Nelson, C. E., Hudson, A., Kraus, M. E. and Carroll, S. B. (2001).
Control of a genetic regulatory network by a selector gene. Science 292, 1164-7.

Hader, T., Wainwright, D., Shandala, T., Saint, R., Taubert, H., Bronner, G. and
Jackle, H. (2000). Receptor tyrosine kinase signaling regulates different modes of
Groucho-dependent control of Dorsal. Curr Biol 10, 51-4.

Halfon, M. S., Grad, Y., Church, G. M. and Michelson, A. M. (2002). Computation-
based discovery of related transcriptional regulatory modules and motifs using an
experimentally validated combinatorial model. Genome Res 12, 1019-28.

Halfon, M. S. and Michelson, A. M. (2002). Exploring genetic regulatory networks in
metazoan development: methods and models. Physiol Genomics 10, 131-43.

Hanna-Rose, W., Licht, J. D. and Hansen, U. (1997). Two evolutionarily conserved
repression domains in the Drosophila Kruppel protein differ in activator specificity. Mol
Cell Biol 17, 4820-9.

Harding, K., Hoey, T., Warrior, R. and Levine, M. (1989). Autoregulatory and gap
gene response elements of the even-skipped promoter of Drosophila. Embo J 8, 1205-12.

Hardison, R., Slightom, J. L., Gumucio, D. L., Goodman, M., Stojanovic, N. and
Miller, W. (1997). Locus control regions of mammalian beta-globin gene clusters:
combining phylogenetic analyses and experimental results to gain functional insights.
Gene 205, 73-94.

Hewitt, G. F., Strunk, B. S., Margulies, C., Priputin, T., Wang, X. D., Amey, R.,
Pabst, B. A., Kosman, D., Reinitz, J. and Arnosti, D. N. (1999). Transcriptional
repression by the Drosophila giant protein: cis element positioning provides an
alternative means of interpreting an effector gradient. Development 126, 1201-10.

Jiang, J., Rushlow, C. A., Zhou, Q., Small, S. and Levine, M. (1992). Individual dorsal
morphogen binding sites mediate activation and repression in the Drosophila embryo.
Embo J 11, 3147-54.

Kadosh, D. and Struhl, K. (1998a). Histone deacetylase activity of de3 is important
for transcriptional repression in vivo. Genes Dev 12, 797-805.

Kadosh, D. and Struhl, K. (1998b). Targeted recruitment of the Sin3-de3 histone
deacetylase complex generates a highly localized domain of repressed chromatin in vivo.
Mol Cell Biol 18, 5121-7.

172

Keller, S. A., Mao, Y., Struffi, P., Margulies, C., Yurk, C. E., Anderson, A. R.,
Amey, R. L., Moore, S., Ebels, J. M., Foley, K. et al. (2000). dCtBP-dependent and -
independent repression activities of the Drosophila Knirps protein. Mol Cell Biol 20,
7247-58.

Kim, T. K. and Maniatis, T. (1997). The mechanism of transcriptional synergy of an in
vitro assembled interferon-beta enhanceosome. Mol Cell 1, 119-29.

Kreitman, M. (1996). The neutral theory is dead. Long live the neutral theory. Bioessays
18, 678-83; discussion 683.

Kulkarni, M. M. a. A., D. N. (2003). Information Display by Transcriptional enhancer,
(ed.

Kumar, V., Carlson, J. E., Ohgi, K. A., Edwards, T. A., Rose, D. W., Escalante, C.
R., Rosenfeld, M. G. and Aggarwal, A. K. (2002). Transcription corepressor CtBP is an
NAD(+)—regulated dehydrogenase. Mol Cell 10, 857-69.

Lee, T. I., Rinaldi, N. J., Robert, F., Odom, D. T., Bar-Joseph, Z., Gerber, G. K.,
Hannett, N. M., Harbison, C. T., Thompson, C. M., Simon, 1. et al. (2002).
Transcriptional regulatory networks in Saccharomyces cerevisiae. Science 298, 799-804.

Licht, J. D., Ro, M., English, M. A., Grossel, M. and Hansen, U. (1993). Selective
repression of transcriptional activators at a distance by the Drosophila Kruppel protein.
Proc NatlAcad Sci U SA 90, 11361-5.

Loots, G. G., Locksley, R. M., Blankespoor, C. M., Wang, Z. E., Miller, W., Rubin,
E. M. and Frazer, K. A. (2000). Identification of a coordinate regulator of interleukins
4, 13, and 5 by cross-species sequence comparisons. Science 288, 136-40.

Ludwig, M. Z., Bergman, C., Patel, N. H. and Kreitman, M. (2000). Evidence for
stabilizing selection in a eukaryotic enhancer element. Nature 403, 564-7.

Ludwig, M. Z. and Kreitman, M. (1995). Evolutionary dynamics of the enhancer
region of even-skipped in Drosophila. Mol Biol Evol 12, 1002-11.

Ludwig, M. Z., Patel, N. H. and Kreitman, M. (1998). Functional analysis of eve stripe
2 enhancer evolution in Drosophila: rules governing conservation and change.
Development 125, 949-58.

Lunyak, V. V., Burgess, R., Prefontaine, G. G., Nelson, C., Sze, S. H., Chenoweth, J.,
Schwartz, P., Pevzner, P. A., Glass, C., Mandel, G. et al. (2002). Corepressor-

dependent silencing of chromosomal regions encoding neuronal genes. Science 298,
1747-52.

Maier, D., Preiss, A. and Powell, J. R. (1990). Regulation of the segmentation gene
fushi tarazu has been functionally conserved in Drosophila. Embo J 9, 3957-66.

173

Mannervik, M. and Levine, M. (1999). The de3 histone deacetylase is required for
segmentation of the Drosophila embryo. Proc Natl Acad Sci U S A 96, 6797-801.

Mannervik, M., Nibu, Y., Zhang, H. and Levine, M. (1999). Transcriptional
coregulators in development. Science 284, 606-9.

Markstein, M. and Levine, M. (2002). Decoding cis-regulatory DNAs in the Drosophila
genome. Curr Opin Genet Dev 12, 601-6.

Markstein, M., Markstein, P., Markstein, V. and Levine, M. S. (2002). Genome-wide
analysis of clustered Dorsal binding sites identifies putative target genes in the
Drosophila embryo. Proc Natl Acad Sci U S A 99, 763-8.

Marmorstein, R. (2002). Dehydrogenases, NAD, and transcription--what‘s the
connection? Structure (Comb) 10, 1465-6.

Munshi, N., Agalioti, T., Lomvardas, S., Merika, M., Chen, G. and Thanos, D.
(2001). Coordination of a transcriptional switch by HMGI(Y) acetylation. Science 293,
1133-6.

Nibu, Y., Senger, K. and Levine, M. (2003). CtBP-independent repression in the
Drosophila embryo. Mol Cell Biol 23, 3990-9.

Nibu, Y., Zhang, H., Bajor, E., Barolo, S., Small, S. and Levine, M. (1998a). dCtBP
mediates transcriptional repression by Knirps, Kruppel and Snail in the Drosophila
embryo. Embo J 17, 7009-20.

Nibu, Y., Zhang, H. and Levine, M. ( 1998b). Interaction of short-range repressors with
Drosophila CtBP in the embryo. Science 280, 101-4.

Poortinga, G., Watanabe, M. and Parkhurst, S. M. (1998). Drosophila CtBP: a Hairy-
interacting protein required for embryonic segmentation and hairy-mediated
transcriptional repression. Embo J 17, 2067-78.

Postigo, A. A. and Dean, D. C. (1999). Independent repressor domains in ZEB regulate
muscle and T-cell differentiation. Mol Cell Biol 19, 7961-71.

Read, M. A., Whitley, M. Z., Williams, A. J. and Collins, T. (1994). NF-kappa B and I
kappa B alpha: an inducible regulatory system in endothelial activation. J Exp Med 179,
503-12.

Rebeiz, M., Reeves, N. L. and Posakony, J. W. (2002). SCORE: a computational
approach to the identification of cis-regulatory modules and target genes in whole-

genome sequence data. Site clustering over random expectation. Proc Natl Acad Sci U S
A 99, 9888-93.

Reinitz, J. and Sharp, D. H. (1995). Mechanism of eve stripe formation. Mech Dev 49,
133-58.

174

Roose, J. and Clevers, H. (1999). T CF transcription factors: molecular switches in
carcinogenesis. Biochim Biophys Acta 1424, M23-37.

Rundlett, S. E., Carmen, A. A., Suka, N., Turner, B. M. and Grunstein, M. (1998).
Transcriptional repression by UME6 involves deacetylation of lysine 5 of histone H4 by
RPD3. Nature 392, 831-5.

Shi, Y., Sawada, J., Sui, G., Affar el, B., Whetstine, J. R., Lan, F., Ogawa, H., Luke,
M. P. and Nakatani, Y. (2003). Coordinated histone modifications mediated by a CtBP
co-repressor complex. Nature 422, 7 35-8.

Small, S., Blair, A. and Levine, M. (1992). Regulation of even-skipped stripe 2 in the
Drosophila embryo. Embo J l 1, 4047-57.

Struhl, K. (2001). Gene regulation. A paradigm for precision. Science 293, 1054-5.

Strunk, B., Struffi, P., Wright, K., Pabst, B., Thomas, J., Qin, L. and Arnosti, D. N.
(2001). Role of CtBP in transcriptional repression by the Drosophila giant protein. Dev
Biol 239, 229-40.

Subramanian, T. and Chinnadurai, G. (2003). Association of class I histone
deacetylases with transcriptional corepressor—CtBP. FEBS Lett 540, 255-8.

Sundqvist, A., Sollerbrant, K. and Svensson, C. (1998). The carboxy-terminal region
of adenovirus ElA activates transcription through targeting of a C-terminal binding
protein-histone deacetylase complex. FEBS Lett 429, 183-8.

Szymanski, P. and Levine, M. (1995). Multiple modes of dorsal-bHLH transcriptional
synergy in the Drosophila embryo. Embo J 14, 2229-38.

Thanos, D. and Maniatis, T. (1995). Virus induction of human IFN beta gene
expression requires the assembly of an enhanceosome. Cell 83, 1091-100.

Valentine, S. A., Chen, G., Shandala, T., Fernandez, J., Mische, S., Saint, R. and
Courey, A. J. (1998). Dorsal-mediated repression requires the formation of a
multiprotein repression complex at the ventral silencer. Mol Cell Biol 18, 6584-94.

Weinmann, A. S. and Farnham, P. J. (2002). Identification of unknown target genes of
human transcription factors using chromatin immunoprecipitation. Methods 26, 37-47.

Weinmann, A. S., Yan, P. S., Oberley, M. J., Huang, T. H. and Farnham, P. J.
(2002). Isolating human transcription factor targets by coupling chromatin
immunoprecipitation and CpG island microarray analysis. Genes Dev 16, 235-44.

Xu, C., Kauffmann, R. C., Zhang, J., Kladny, S. and Carthew, R. W. (2000).
Overlapping activators and repressors delimit transcriptional response to receptor
tyrosine kinase signals in the Drosophila eye. Cell 103, 87-97.

175

Yuh, C. H., Bolouri, H. and Davidson, E. H. (1998). Genomic cis-regulatory logic:
experimental and computational analysis of a sea urchin gene. Science 279, 1896-902.

Zhang, H. and Levine, M. (1999). Groucho and dCtBP mediate separate pathways of
transcriptional repression in the Drosophila embryo. Proc Natl Acad Sci U S A 96, 535-
40.

176

APPENDIX A

Composite element representing two different information states exhibits
orientation-independent activity characteristic of enhancers.

As described before in Chapter 2 (Figure II-lF) a composite enhancer element,
containing binding sites for the endogenous short-range repressor Giant, endogenous
activators Twist and Dorsal, and chimeric Gal4 activators, behaves like an ‘information
display’, simultaneously representing contrasting information of activation and repression
at the same time and in the same nucleus. This enhancer element drives lacZ reporter
gene expression from the basal hsp 70 promoter in a complex pattern such that in the
nuclei where the activators and the repressor Giant are co-expressed, transcription is
driven by the cluster of Gal4 activators within the compact regulatory element, while at
the same time the Dorsal and Twist activators within the same element are being actively
repressed by Giant. Thus, this compact regulatory element, has subelements that
simultaneously represent two types of information states, activation and repression,
unlike the binary switch activity observed for many transcriptional enhancers, where
usually a single signal to activate or repress is present. Since this dual activity is unlike
that of ‘real’ enhancers, we tested to see if the element we had created possessed any of
the qualities that are attributed to natural enhancers, namely the ability to function in a
distance- and orientation- independent manner (Banerji et al., 1981), by placing it in
either orientation between the divergently transcribed white (at —265 bp) and lacZ genes
(at —130 bp). We showed that although this element did not function in the binary on or
off mode, its activity was independent of both orientation and distance with respect to the

start of transcription (Figure II-2). The distance- and orientation- independent activity of

177

the element was tested on the transposase lacZ gene, which is different from the hsp 70
basal promoter used in Figure II-l. In order to confirm that these activities were an
inherent property of the element and not specific for the promoter being tested, I also
tested the orientation-independent activity of the element on the hsp 70 lacZ reporter
gene as shown in Figure A-lA, B. A similar pattern of repression and activation of the
hsp 70-lacZ gene is seen when the element is tested in the opposite orientation (Figure A-

1A, B).

178

 

 

Figure A-l: Compact regulatory element displays enhancer-like property of
orientation independence.

(A, B) The regulatory element shown in Figure II-lF was inserted in the opposite
orientation into the UAS-lacZ vector containing the hsp 70 basal promoter. In the
presence of Dorsal, Twist, and Gal4 activators, a composite pattern of gene regulation is
seen as in Figure II-lF with inhibition of Dorsal/Twist and activation by Gal4.

Patterns of gene expression were visualized in 2-4 hour embryos by in situ hybridization
with digU labeled antisense lacZ probes.Embryos are oriented anterior to left; ventral
view (A) and lateral view (B) are shown.

179

MATERIALS AND METHODS

l. Plasmid construction

GAL4 (aa1-93) - GAL4 AD (aa753-881)

A KpnI-Xbal fragment from pSCTEV GAL4 (1-93)- GAL4 (Seipel et al., 1992)
containing the reading frame for the yeast GAL4 activation domain (Gal4 AD) from
amino acid residues 753-881, was cloned into KpnI-Xbal cut pTwiggy (Amosti et al.,
1996b) vector, which contains the twist enhancer (2xPEe-Et) element, twist basal

promoter and the GAL4 DNA-binding domain from residues 1-93.

Reporter gene

The vector M2g5u-lacZ (Chapter 2) was cut with HindIII-SphI enzymes to
remove the five Gal4 sites. These reporters were further modified by introducing oligos
containing two Dorsal (d1) and two Twist (twi) binding sites (DA481/482: 5’AGC TTG
AGG GAT TTT CCC AAA TCG AGG GAA AAC CCA ACT CGC ATA TGT TGA
GCA TAT GGC ATG 3’) (Szymanski and Levine, 1995). Annealed oligos containing a
neutral 55 bp spacer DNA (DA65/66: 5’TCC ATG ATA AAC GCG TGC TAG ACT
ATT GCA GGT ACT GAT CGA ATG CCT CTG CAT G 3’) was placed downstream of
the Dorsal and Twist sites and upstream of the hsp 70 basal promoter (Hewitt et al.,
1999). Five Gal4 binding sites were PCR amplified from the UAS-lacZ vector (Brand and
Perrimon, 1993) using oligos DA487/488 (5’ AAG GAA AAA AGC GGC CGC GCG
CTC GCT AGA GTC 3’) and were introduced upstream of the two Giant binding sites at

the NotI site resulting in the vector M5u2g2dl.twi-55-lacZ.

180

2. P-element transformation, crosses to reporter genes, and whole-mount in situ

hybridization of embryos.

P-element transformation vectors were introduced into the Drosophila gerrnline
by injection of yw67 embryos as described (Small et al., 1992). Embryos were collected
from a cross between a reporter line and a line expressing the GAL4 activator in the
ventral regions of the embryo. The embryos were fixed and stained using digoxigenin-

UTP labeled antisense RNA probes to either lacZ or w as described (Small et al., 1992).

181

REFERENCES

Arnosti, D. N., Gray, S., Barolo, S., Zhou, J. and Levine, M. (1996b). The gap protein
knirps mediates both quenching and direct repression in the Drosophila embryo. Embo J
15, 3659-66.

Banerji, J., Rusconi, S. and Schaffner, W. (1981). Expression of a beta-globin gene is
enhanced by remote SV40 DNA sequences. Cell 27, 299-308.

Brand, A. H. and Perrimon, N. (1993). Targeted gene expression as a means of altering
cell fates and generating dominant phenotypes. Development 118, 401-15.

Hewitt, G. F., Strunk, B. S., Margulies, C., Priputin, T., Wang, X. D., Amey, R.,
Pabst, B. A., Kosman, D., Reinitz, J. and Arnosti, D. N. (1999). Transcriptional
repression by the Drosophila giant protein: cis element positioning provides an
alternative means of interpreting an effector gradient. Development 126, 1201-10

Seipel, K., Georgiev, O. and Schaffner, W. (1992). Different activation domains
stimulate transcription from remote ('enhancer') and proximal ('promoter') positions.
EmboJll, 4961-8.

Small, S., Blair, A. and Levine, M. (1992). Regulation of even-skipped stripe 2 in the
Drosophila embryo. Embo J 11, 4047-57.

Szymanski, P. and Levine, M. (1995). Multiple modes of dorsal-bHLH transcriptional
synergy in the Drosophila embryo. Embo J 14, 2229-38.

182

APPENDIX B

CtBP-dependent and CtBP-independent activities contribute quantitatively to
Knirps repressor function‘.

In this study I have defined the contextual elements of cis-regulatory enhancers
that allow short-range repressors to mediate repression effectively. However, we do not
yet understand the physical and biochemical changes in promoter complexes that
accompany short-range repression. Chromatin Immunoprecipitation (ChIP) assays in
Drosophila embryos using the simple genetic switch elements defined in this study will
facilitate our understanding of the molecular mechanisms of repression employed at
developmentally regulated genes. These experiments can be designed to directly test
mechanistic questions about the nature. of short-range transcriptional repression,
specifically, whether repression blocks activator access to the DNA alters chromatin
structure, or whether changes in the composition or presence of the basal machinery are

observed.

Chromatin Irnmunoprecitation (ChIP) has been extensively used to study
transcriptional regulation in cell culture, where a homogeneous population of cells is
available for analysis. In the Drosophila embryo however, the endogenous targets of the
short-range repressors such as the pair-rule gene even-skipped, are expressed in discrete
domains in the early blastoderm embryo (Frasch et al., 1987; Harding et al., 1986;
Macdonald and Struhl, 1986) and therefore, are not useful targets for ChIP experiments

as such assays would result in high background noise. Using transgenes, such as those

 

1This data is included in the following manuscript submitted to Development: Paolo Struffi, Maria Corado,
Meghana M. Kulkami and David N. Amosti. Quantitative Contributions of CtBP-dependent and —
independent repression activities of Knirps.

183

described in Chapter 2 and 3 of this study would allow one to regulate the expression of
the reporter transgene throughout the embryo using inducible and ubiquitous expression

of activators and the repressors.

The first step towards performing Drosophila embryo ChIP assays is to obtain a
homogeneous population of nuclei/cells. This involves engineering embryos to carry a
transgene that can be activated in all (or almost all) nuclei and then repressed in all (or
almost all) nuclei at a given time. In the embryo however, the repressors are expressed in
spatially localized domains (Gergen and Wieschaus, 1986; Hoch et al., 1992; Petschek et
al., 1987; Rothe et al., 1989) and are therefore, typically active on a given gene only in a
small fraction of the nuclei. Thus a CM? analysis of the whole embryo will represent a
mixture of promoters that are not repressed and those that are actively repressed. To
overcome this difficulty and to increase the signal to noise ratio, I decided to utilize
transgenic flies that express different versions (full length and N-tenninal dCtBP-
independent repression domain) of recombinant Knirps proteins, ubiquitously and under
a heat-shock inducible promoter (constructed and generated by Paolo Struffi in the lab).
As a first step, it was important to determine if the recombinant proteins were able to
function as repressors in vivo on an endogenous Knirps target. I looked at the expression
pattern of one of Knirps targets, the pair-rule gene even-skipped (eve) in transgenic
embryos before and after heat shock. Knirps is required for correct regulation of the eve
stripe 3/7 and 4/6 enhancers, as demonstrated by the expression patterns of lacZ reporter
genes in kni mutant embryos (Fujioka et al., 1999; Small et al., 1996). It has been
previously demonstrated that (Keller et al., 2000) the posterior portion of the eve stripe 3

pattern is not derepressed in a CtBP mutant, consistent with the CtBP-independent

184

activity of Knirps on this enhancer. In contrast, Knirps repression of eve stripe 4/6 is
compromised in a CtBP mutant background, indicating that the CtBP-independent
repression activity of Knirps is insufficient to regulate this enhancer (Struffi et al.,
submitted). Therefore, depending on which portion of the eve gene is bound by the
Knirps protein, its repression activity is either dependent or independent of the CtBP

cofactor.

2-4 hour old transgenic embryos carrying transgenes for either the full-length
recombinant Knirps protein or the dCtBP-independent repression domain were heat-
shocked at 38 ° C for 30 minutes, recovered for 30 minutes in a water bath at room
temperature before fixing with formaldehyde. In situ hybridization was performed on
both heat-shocked (Figure B-lB, D) and nOn heat-shocked transgenic (Figure B-lA, C)
embryos using antisense mRNA probe to eve. The N-terminal region of Knirps (kni l-
330) is a weak repressor compared to the full-length protein (kni 1-429), but is able to
repress the previously known dCtBP-dependent target, the eve stripe 4+6 enhancer, when
over-expressed (Figure B-lD). This suggests that increasing the dose of the repressor
may be sufficient to overcome the requirement for the dCtBP cofactor and that multiple
repression activities within a single protein represent quantitative effects on gene
expression. Upon heat-shock both the full-length (Figure B-lB) and the dCtBP-
independent (Figure B-lD) versions of the recombinant Knirps protein were able to

abolish the expression of eve stripes 3/7 and the eve stripes 4/6.

185

CtBP

l

PMDLSMK

 

 

 

 

1 429 1 330
hisl 0301 I JFLAG hisLDBD I lFLAG
332-337
C
A trgt'fi‘i- .- arr-"r"
‘34s:
1,, _,
,,. l: " .‘. .3, 1
l 2 a s s ' ’ " g: 5 '
I . D , .
0' j.
i.
i.
1132557

a},

Figure B-l: Pattern of endogenous eve expression in embryos expressing full length
Knirps 1-429 (B) and CtBP-independent region of Knirps 1-330 (D).

Structure of proteins expressed from hsp70 promoter: 1-429, full-length Knirps protein;
1-330, CtBP-independent Knirps repression domain are depicted above the corresponding
embryos. Endogenous eve pattern is visualized by in situ hybridization using digoxigenin
labeled eve mRN A probe before and after heat-shock of embryos carrying transgenes that
express either the hill-length recombinant Knirps protein or the CtBP-independent Knirps
repression domain. Embryos are oriented anterior to right and lateral views are shown.

(A, C) Endogenous eve expression in non heat-shocked embryos carrying the transgene
for firll-length recombinant Knirps protein (A) or the CtBP-independent Knirps
repression domain (C).

(B) Endogenous eve expression in heat-shocked embryos expressing the full-length
recombinant Knirps protein. Overexpression of full-length Knirps abolishes the
expression of all eve stripes except for stripe 5.

(D) Endogenous eve expression in heat-shocked embryos expressing the CtBP-
independent Knirps repression domain. Overexpression of CtBP-independent Knirps
repression domain abolishes the expression of eve stripes 3, 4, 6, and 7.

186

 

The experiments described above also demonstrated that recombinant Knirps
protein expressed ubiquitously throughout the embryo under heat shock conditions was

functional as a repressor and could be used in future ChIP assays.

MATERIALS AND METHODS

Transgenic flies expressing recombinant Knirps proteins

Transgenic flies carrying heat-shock inducible versions of the Knirps proteins,
thni 1-429 (full-length protein) and thni 1-330 (dCtBP-independent domain) with a
hexahistidine tag at the N-tenninus and a double FLAG tag at the C-terrninus were
obtained from Paolo Struffi in the lab. These proteins can be expressed throughout the
blastoderm embryo under conditions of heat-shock. Construction of these constructs is

described in Struffi et al., (submitted).

Heat-shock treatment

To induce expression of recombinant Knirps proteins, 2-4 hour old embryos
collected on apple-juice plates at room temperature (22-23°C) were incubated for 30
minutes at 38°C in a 10-liter water bath to ensure rapid and even heating. After induction,
embryos were allowed to recover in a water bath at room temperature for 30 minutes

prior to fixation.

187

In situ hybridization

In situ hybridizations were performed using digoxigenin-UTP-labeled antisense
RNA probes to eve on both heat-shocked and non-heat-shocked transgenic embryos

(Small et al., 1992).

188

REFERENCES

Frasch, M., Hoey, T., Rushlow, C., Doyle, H. and Levine, M. (1987). Characterization
and localization of the even-skipped protein of Drosophila. Embo J 6, 749-59.

Fujioka, M., Emi—Sarker, Y., Yusibova, G. L., Goto, T. and Jaynes, J. B. (1999).
Analysis of an even-skipped rescue transgene reveals both composite and discrete
neuronal and early blastoderm enhancers, and multi-stripe positioning by gap gene
repressor gradients. Development 126, 2527-38.

Gergen, J. P. and Wieschaus, E. (1986). Dosage requirements for runt in the
segmentation of Drosophila embryos. Cell 45, 289-99.

Harding, K., Rushlow, C., Doyle, H. J., Hoey, T. and Levine, M. (1986). Cross-
regulatory interactions among pair-rule genes in Drosophila. Science 233, 953-9.

Hoch, M., Gerwin, N ., Taubert, H. and Jackle, H. (1992). Competition for overlapping
sites in the regulatory region of the Drosophila gene Kruppel. Science 256, 94-7.

Keller, S. A., Mao, Y., Struffi, P., Margulies, C., Yurk, C. E., Anderson, A. R.,
Amey, R. L., Moore, S., Ebels, J. M., Foley, K. et al. (2000). dCtBP-dependent and -
independent repression activities of the Drosophila Knirps protein. Mol Cell Biol 20,
7247-58.

Macdonald, P. M. and Struhl, G. (1986). A molecular gradient in early Drosophila
embryos and its role in specifying the body pattern. Nature 324, 537-45.

Petschek, J. P., Perrimon, N. and Mahowald, A. P. (1987). Region-specific defects in
l(1)giant embryos of Drosophila melanogaster. Dev Biol 119, 175-89.

Rothe, M., N auber, U. and Jackle, H. (1989). Three hormone receptor-like Drosophila
genes encode an identical DNA-binding finger. Embo J 8, 3087-94.

Small, S., Blair, A. and Levine, M. (1992). Regulation of even-skipped stripe 2 in the
Drosophila embryo. Embo J 11, 4047-57.

Small, S., Blair, A. and Levine, M. (1996). Regulation of two pair-rule stripes by a
single enhancer in the Drosophila embryo. Dev Biol 175, 314-24.

189

APPENDIX C

Relative potencies of short-range repressor function

The analyses of short-range transcriptional repressors, Giant, Knirps and Kruppel,
have so far indicated that they function in very similar ways. These proteins are capable
of repressing the activities of enhancer elements when bound within approximately 100
bp of key activator sites, or of a basal promoter element when cognate sites are
introduced close to the start of transcription (Amosti et al., 1996; Gray and Levine, 1996;
Gray et al., 1994; Hewitt et al., 1999; Keller et al., 2000). Another common property of
short-range transcriptional repressors is their interaction with the evolutionarily
conserved corepressor CtBP (C-tenninal Binding Protein) (Nibu et al., 1998a; Nibu et al.,
1998b) Drosophila short-range repressors also possess CtBP-independent repression
activities and the requirement for dCtBP can vary on a gene-to-gene as well as on an
enhancer-to-enhancer basis. We have successfully shown that increasing the dose of the
Knirps repressor may be sufficient to overcome the requirement for the dCtBP cofactor
and that multiple repression activities within a single protein represent quantitative
effects on gene expression (Paolo Struffi, Maria Corado, Meghana M. Kulkami and
David N. Amosti. Quantitative Contributions of CtBP-dependent and —independent

repression activities of Knirps. Manuscript submitted).

Using synthetic enhancer elements where the identity, stoichiometry, and the
exact arrangement of activator and repressor binding site are well-defined, we
demonstrated that the previously held simple notion that short-range repressors block the

activity of all protein complexes within a 100 bp is incorrect. The manipulation of these

190

 

composite elements in terms of the number of activator and repressor binding sites,
relative affinities, spacing and distribution of these binding sites further allowed us to

define the contextual parameters that dictate repression effectiveness.

The contextual dependencies of repression described in Chapter 3 were developed
for the Giant repressor. To determine if similar rules applied to other types of repressors,
we carried out parallel evaluations of the short-range repressors Giant, Knirps, and
Kruppel. To test quantitative similarities or differences between these factors we created
reporters that would compare repressor activity on genes that represented permissive or
non-permissive contexts for the Giant protein. The short-range repressor Knirps was
unable to inhibit lacZ expression driven by the Gal4 activator from five high affinity Gal4
sites indicating a similar limitation of repression on even proximally bound activators as
Giant (Figure C-lE and F). The Kruppel protein was also unable to repress the activity of
five high affinity Gal4 sites (Figure C-lG), although many precellular embryos showed a
narrowing in the lacZ expression pattern (Figure C-lH; arrow) in the central regions
where Kruppel is expressed. Thus, in general, the short-range repressor Kruppel appears
to be a more potent transcriptional repressor than Giant or Knirps. The Giant (Figure C-
lA; arrows) and Kruppel (Figure C-lD; arrow) factors exhibited repression activity in the
corresponding regions of the embryo when tested against three Gal4 sites. The Knirps
repressor was also active in this context although in general the levels of repression
appeared to be lower (Figure C-lB and C; arrows).

The results described above indicate that although Giant, Knirps, and Kriippel

function similarly in most respects, they may have different repression potencies. This

191

 

difference may reflect subtle differences in the mechanism of repression employed or
may likely represent differences in the levels of the protein or DNA binding specificity.
Whatever the case, these observations are important with ramifications not only in the
efficient design of reporters for ChIP assays but also for computational modeling of gene

regulatory networks involving the short-range repressors.

MATERIALS AND METHODS (SEE CHAPTER 3) l

 

192

 

 

is ."
g. , .- -“ K
B

- r.
’ ‘it if} Q . at
Knl K
C G

' V9} ’ 3‘ ‘ f

’1 v‘ 9 s .9!
KM K Kt
D H
"' t t

  

5. as“ ‘f?
“Lin-n J’”: * ' . I“

repGau so up Gan-so

Figure C-l: Relative potencies of the short-range repressors Giant, Knirps and
Kriippel

A schematic of the reporters used is shown below. The reporters contain two binding

sites for either one of the short-range repressors Giant, Knirps or Kruppel and five/three
high affinity Gal4 binding sites.

All three of the short-range repressors Giant (A), Knirps (B, C), and Kruppel (D) are able
to repress the activity of the Gal4-activator in the context of three Gal4 binding sites
(arrows). However, the level of repression mediated by the Knirps protein was in general
lower.

All the short-range repressors Giant (E), Knirps (F), and Kruppel (G, H) are unable to
repress in the context of five high affinity 0314 sites. However, in the case of the Kruppel
protein, we observe a narrowing in the lacZ expression pattern in the cennal regions of
the embryo where the repressor is present (H).

Embryos are oriented with anterior to left; lateral views are shown. LacZ expression was

visualized in 2-4 hr old embryos using in situ hybridization with digU labeled antisense
mRNA probes to LacZ.

193

REFERENCES

Arnosti, D. N., Gray, S., Barolo, S., Zhou, J. and Levine, M. (1996). The gap protein
knirps mediates both quenching and direct repression in the Drosophila embryo. Embo J
15, 3659-66.

Gray, S. and Levine, M. (1996). Short-range transcriptional repressors mediate both
quenching and direct repression within complex loci in Drosophila. Genes Dev 10, 700-
10.

Gray, S., Szymanski, P. and Levine, M. (1994). Short-range repression permits
multiple enhancers to function autonomously within a complex promoter. Genes Dev 8,
1829-38.

Hewitt, G. F., Strunk, B. S., Margulies, C., Priputin, T., Wang, X. D., Amey, R.,
Pabst, B. A., Kosman, D., Reinitz, J. and Arnosti, D. N. (1999). Transcriptional
repression by the Drosophila giant protein: cis element positioning provides an
alternative means of interpreting an effector gradient. Development 126, 1201-10.

Keller, S. A., Mao, Y., Struffi, P., Margulies, C., Yurk, C. E., Anderson, A. R.,
Amey, R. L., Moore, S., Ebels, J. M., Foley, K. et al. (2000). dCtBP-dependent and -
independent repression activities of the Drosophila Knirps protein. Mol Cell Biol 20,
7247-58.

Nibu, Y., Zhang, H., Bajor, E., Barolo, S., Small, 8. and Levine, M. (1998a). dCtBP
mediates transcriptional repression by Knirps, Kruppel and Snail in the Drosophila
embryo. Embo J 17, 7009-20.

Nibu, Y., Zhang, H. and Levine, M. (1998b). Interaction of short-range repressors with
Drosophila CtBP in the embryo. Science 280, 101-4.

194

 

 

APPENDIX D

Analysis of other Drosophila transcriptional repressors

Studies on transcriptional repression in the Drosophila embryo indicate that there
may be two basic forms of repression: Long-range and Short-range repression. Long-
range repressors, for example typified by the Drosophila Hairy protein, function over
distances of at least 500 bp to silence the transcriptional complex or to inhibit upstream
activators that are bound to promoter-proximal regions. In principle, long-range
repressors can function in a dominant fashion to block multiple enhancers in a complex
modular promoter (Barolo and Levine, 1997; Gray and Levine, 1996). Short-range
repressors present in the early Drosophila embryo include the products of the gap genes
snail (sna), krilppel (Kr), giant (gt), and knirps (kni). These proteins are capable of
repressing the activities of enhancer elements when bound within approximately 100 bp
of key activator sites, or of a basal promoter element when cognate sites are introduced
close to the start of transcription (Amosti et al., 1996b; Gray and Levine, 1996; Gray et
al., 1994; Hewitt et al., 1999; Keller et al., 2000). The mechanisms by which short-range
and long-range Drosophila repressors inhibit transcription are poorly understood,
although one model of repression in the embryo suggests that the short-range/long-range
distinction results from the recruitment of distinct cofactors (Nibu et al., 1998a; Nibu et
al., 1998b; Zhang and Levine, 1999). Short-range repressors recruit the corepressor CtBP
to mediate repression whereas long-range repressors have been shown to interact with the
Groucho corepressor (Barolo and Levine, 1997; Fisher and Caudy, 1998; Mannervik et
al., 1999; Poortinga et al., 1998). An interesting observation is that both the short-range

corepressor, CtBP, and the long-range corepressor, Groucho, have been shown to interact

195

with and recruit Histone deacetylase complexes (Flores-Saaib and Courey, 2000; Nibu et
al., 2001; Sundqvist et al., 1998).

In addition to having a longer range of action, we find that the long-range
repressor Hairy is more potent than Giant even on a local scale and can block the activity
of a greater number of activators as compared to Giant. However, contrary to previous
studies that suggest that long-range repressors function by spreading along the chromatin
to silence a large chromosomal locus, we find that the repression mediated by Hairy is
transient similar to the activity of the short-range repressors and is easily reversed as the
embryo deve10ps.

We decided to compare the activities of other Drosophila transcriptional
repressors to determine if we could detect similar such differences and similarities in
their repression function. Since several of these transcriptional repressors are not well
characterized, we first decided to test the activity of different repressors on genes that
represented permissive contexts for the Giant protein. We constructed different hsp 70-
lacZ reporters containing two binding sites for the repressor, three Gal4 sites and a spacer

similar to the M2g3u2x-lacZ construct used in Chapter 3.

Snail (sna): Snail is a zinc finger protein and is a key mesoderm determinant in
the Drosophila embryo. It is activated in the ventral regions of the embryo by the
transcription factor Dorsal. Snail has been characterized to be a short-range
transcriptional repressor like Giant, Knirps and Kruppel and has been shown to inhibit
activators within enhancers or the basal promoter itself over distances of a 100 bp (Gray

and Levine, 1996; Gray et al., 1994). The Snail repression domain contains both a

196

 

conserved copy of the P-DLS-k motif, as well as the slightly divergent sequence P-DLS-
R. Like other short-range repressors Snail, interacts with and requires the CtBP

corepressor to mediate repression.

However, we find that in a context where the other short-range repressors are able
to mediate repression, Snail is unable to block the activity of three high affinity Gal4 sites
(Figure D-lA, B). We consider several possibilities for this lack of Snail repressor
function: First, the binding sites do not bind the protein efficiently. The binding sites used
here were previously used in the context of the rhomboid NEE enhancer (Gray and
Levine, 1996; Gray et al., 1994; Nibu et al., 1998a; Nibu et al., 1998b), which is an
endogenous target of Snail. In this context Snail was shown to mediate repression of the
linked lacZ gene. It is possible that the rhomboid NEE enhancer has cryptic Snail binding
sites that cooperate with the synthetic Snail sites used in this study to mediate repression
by Snail. Second, Snail protein levels may be low compared to that of Giant, Knirps and
Kriippel and therefore two Snail binding sites may be insufficient. Third, Snail is a weak
transcriptional repressor and the contextual ‘rules’ defined for the short-range repressors

may not apply to the Snail repressor.

Engrailed, Runt and Even-skipped: Runt is a potent repressor of the pair-rule
gene even-skipped (Jimenez et al., 1996; Manoukian and Krause, 1993) and the
segmentation gene engrailed and is expressed in seven transverse stripes in the cellular
blastoderm stage embryo (Wheeler et al., 2002). Runt has also been shown to activate the

expression of the sex lethal gene in Drosophila (Kramer et al., 1999). Runt has been

197

 

 

Figure D-l: The activity of the Gal4 activators from three high affinity sites is not
repressed by Snail.

Schematic of the reporter gene is shown below. The reporter contains two Snail binding
sites (sna), three high affinity Gal sites (colored circles), a neutral spacer (open circles),
hsp 70 basal promoter driving lacZ expression.

(A) Snail is unable to repress the Gal4 activators in early blastoderm embryo.
(B) Snail is unable to repress the Gal4 activators even at later stages of embryogenesis.

LacZ expression is activated throughout the embryo by ubiquitous expression of the full
length Gal4 activator protein under the control of the actin5C enhancer. Patterns of gene
expression were visualized in 2-4 hour embryos by in situ hybridization with digU
labeled antisense lacZ probes. Embryos are oriented anterior to left; lateral view (A) and
field shot of embryos at various stages in development (B) are shown.

198

 

shown to interact physically and genetically with the corepressor Groucho (Aronson et
al., 1997). Interestingly, only a subset of genes, for example even-skipped, repressed by
Runt require the recruitment of Groucho for their repression (Aronson et al., 1997;
Jimenez et al., 1996).

engrailed encodes for a homeodomain protein and is expressed in fourteen stripe
(each one cell wide) along the anterior-posterior axis and plays a key role during
segmentation in the Drosophila embryo. Engrailed has been characterized as a long-range
transcriptional repressor and has been shown to interact with and recruit the corepressor
Groucho to mediate repression of its target genes such as even-skipped (Jimenez et al.,
1997). Engrailed like Runt has also been shown to have Groucho-independent repression
activities (Tolkunova et al., 1998).

even-skipped encodes a homeodomain transcription factor that is required for the
expression of both odd- and even-numbered Engrailed stripes, which are activated by
distinct mechanisms (DiNardo and O'Farrell, 1987; Howard and Ingham, 1986). Previous
data suggests that the role of Eve in the activation of engrailed might be, at least in part,
indirect. Early Eve stripes repress paired at high concentrations, and sloppy-paired, a
repressor of engrailed (Cadigan et al., 1994a; Cadigan et al., 1994b; Grossniklaus et al.,
1992) at low concentrations (Fujioka et al., 1995). Eve interacts with and recruits the
Groucho corepressor to mediate repression of both paired and sloppy-paired. Eve also
represses another repressor of engrailed, odd-skipped in a Groucho-independent manner
(Fujioka et al., 1995; Fujioka et al., 2002; Manoukian and Krause, 1992).

Thus, all three proteins Runt, Engrailed and Even-skipped mediate repression via

Grouch-dependent and —independent pathways like the CtBP-dependent and -

199

independent activities of the short-range repressors. However, when tested against three
high affinity Gal4 sites, reporters carrying binding sites for these repressors did not show
corresponding repression of lacZ expression (Figure D-2A, B, C, D, E, F, G, and H).
Again, it is possible that the failure to mediate repression because the binding sites used
do not bind the protein efficiently or because protein levels are insufficient to mediate

repression in this context.

The experiments described above tested the activity of repressors that recruit the
CtBP corepressor or repressors that recruit the Groucho corepressor to mediate
repression. In Drosophila there are at least two known transcription factors (Brinker and
Suppressor of Hairless) that interact with and require both CtBP and Groucho to mediate

repression.

Brinker (Brk): Brinker is a transcription repressor and is expressed in
ventrolateral regions of the embryo abutting the dorsal decapentaplegic (dpp) expression
domain. Brinker is itself one of the downstream targets of dpp signaling and in turn
represses some Dpp-responsive genes (Campbell and Tomlinson, 1999; Jazwinska et al.,
1999a; Jazwinska et al., 1999b; Minami et al., 1999). It has been demonstrated that
Brinker harbors a functional and transferable repression domain, through which it recruits
the corepressors Groucho and CtBP. The mechanism of Brinker repression is dependent
on promoter context and requires either or both Groucho and CtBP for switching of some
target genes, whereas for silencing of others, it requires neither of these cofactors in

which case it might be involved in direct competition with activators for DNA binding

200

 

sites (Hasson et al., 2001; Seller and Bienz, 2001). When tested against three high
affinity Gal4 sites, no repression was observed in ventolateral stripes, leading to the

conclusion that Brinker was unable to block lacZ expression (Figure D-3A, B).

 

201

 

 

 

meals-so

Figure D-Z: The activity of the Gal4 activators from three high affinity sites is not
repressed by Engrailed (en), Runt (run) or Even-skipped (eve).

Schematic of the reporter gene is shown below the corresponding embryos. The reporter
contains two Engrailed (en) or two Rant (run) or two Even-skipped (eve) binding sites,
three high affmity Gal sites (colored circles), a neutral spacer (open circles), hsp 70 basal
promoter driving lacZ expression

(A, B, C) Engrailed is unable to repress the Gal4 activators at the difl‘erent stages of
embryogenesis examined.

(D, E, F) Rum is unable to repress the Gal4 activators at various stages of
embryogenesis.

(G, H) Even-skipped is unable to repress the Gal4 activators at various stages of
embryogenesis.

LacZ expression is activated by the Gal4-Gal4 AD chimeric protein expressed in the
ventral regions of the embryo. Patterns of gene expression were visualized in 2-4 hour
embryos by in situ hybridization with digU labeled antisense lacZ probes. Embryos are
oriented anterior to left; lateral view (A, B, C, D, E, F, and G) and field shot of embryos
at various stages in development (H) are shown.

202

Suppressor of Hairless Su(H): Notch is the receptor for a conserved signaling
pathway that regulates numerous cell fate decisions during development (Qi et al., 1999).
Signal transduction involves the presenilin-dependent intracellular processing of Notch
and the nuclear translocation of the intracellular domain of Notch, NICD (Lecourtois and
Schweisguth, 1998; Struhl and Adachi, 1998; Struhl and Greenwald, 1999). NICD
associates with Suppressor of Hairless [Su(H)], a DNA binding protein, and Mastermind
(Mam), a transcriptional coactivator (Petcherski and Kimble, 2000). In the absence of
Notch signaling, Su(H) acts as a transcriptional repressor (Barolo et al., 2000; Dou et al.,
1994; Morel and Schweisguth, 2000). It has been shown that Hairless, an antagonist of
Notch signaling (Bang et al., 1995; Bang et al., 1991; Bang and Posakony, 1992;
Schweisguth and Posakony, 1994), and is required to repress the transcription of the
singleminded gene in the Drosophila embryo. Hairless forms a DNA-bound complex
with Su(H). Furthermore, it directly binds the Drosophila C-tenninal Binding Protein
(dCtBP) and Groucho, which act as transcriptional corepressors (Barolo et al., 2002;
Morel et al., 2001). In the case of Su(H), we see some interesting patterns of lacZ
expression when the protein is tested against three high affinity Gal4 sites (Figure D-3C,
D, and E). In some embryos we see a posterior patch of lacZ expression in the dorsal
region of the embryo (Figure D-3C; arrow). This may reflect activation by Su(H) in the
absence of the Gal4 activator in nuclei where Notch signaling is active. In other embryos
(Figure D-3D and E; arrow), we observe more intense lacZ staining, with a gap in
expression in the central regions of the embryo that is wider in the more ventral regions.

In this case we believe that lacZ

203

 

 

Figure D-3: The activity of the transcription factors Brinker (brk) and Suppressor
of Hairless [Su(H)] on a reporter with three high affinity Gal4 sites.

Schematic of the reporter gene is shown below the corresponding embryos. The reporter
contains two Brinker (brk) or two Suppressor of Hairless [Su(H)] binding sites, three high
affinity Gal sites (colored circles), a neutral spacer (open circles), hsp 70 basal promoter
driving lacZ expression.

(A, B) Brinker expressed in ventrolateral stripes in the blastoderm embryo is unable to
repress the Gal4 activators at the different stages of embryogenesis examined.

(C, D, E) Activity of Suppressor of Hairless [Su(H)]on hsp 70-lacZ reporter. (C)
Posterior patch of lacZ expression (arrow) may result from activation by Su(H) alone in
the absence of Gal4 activators. (D, E) lacZ staining is more intense, suggesting that
Su(H) is synergizing with the Gal4 activator to drive lacZ expression. In the nuclei that
show lacZ expression presumably Notch signaling is active. LacZ staining shows a gap in
the central regions of the embryo (arrow). The gap is much wider in the ventrolateral and
ventral regions of the embryo. This gap in lacZ expression may be mediated by Su(H)
acting as a repressor presumably because Notch signaling is absent in the nuclei in this
region.

LacZ expression is driven by full length Gal4, which is expressed ubiquitously
throughout the embryo under the control of the actin5C enhancer. Patterns of gene
expression were visualized in 2-4 hour embryos by in situ hybridization with digU
labeled antisense lacZ probes. Embryos are oriented anterior to left; lateral view (A, C,
D, and E)) and field shot of embryos at various stages in development (B) are shown.

204

staining is synergistically activated by both Gal4 and Su(H) in nuclei where Notch
signaling is active, and in regions where lacZ expression is absent Su(H) is actively
repressing the Gal4 activator in the absence of a Notch signal. Notch signaling in early
embryogenesis is not well characterized. Therefore further investigation of Su(H) activity
on this reporter construct will have to be examined later in development, for example in

the wing imaginal disc where Notch signaling is well studied.

In addition to the repressors described above, I also tested the activity of the
transcription factors Sloppy-paired, Tramtrack and Cubitus interruptus on the reporter
with three high affinity Gal4 sites. The activity of these proteins as repressors is not well

characterized and not much is known about their mode of action in gene regulation.

Sloppy-paired (Slp 1): Sloppy-paired (Slp 1) is a forkhead domain protein and
has been recently shown to be involved in setting up the anterior border of eve stripe 2
expression in the Drosophila embryo (Andrioli et al., 2002). Two kinds of embryos were
frequently seen in experiments testing the activity of this protein (Figure D-4A and B).
We observed embryos that showed lacZ expression in a broad domain in the anterior
region of the embryo, (Figure D-4A) which is consistent with the pattern in which
Sloppy-paired is expressed. We also observed embryos showing lacZ expression in a
continous ventral swath, presumably driven by the Gal4 activator, in addition to the
anterior broad domain (Figure D-4B). These results indicate that Slp 1 activates rather

than represses in the context tested.

205

 

Figure D-4: The activity of the transcription factors Sloppy-paired (Slpl),
Tramtrack (ttk) and Cubitus interruptus (Ci) on a reporter with three high affinity
Gal4 sites.

Schematic of the reporter gene is shown below the corresponding embryos. The reporter
contains two Sloppy-paired (Slpl) or two Tramtrack (ttk) or two Cubitus interruptus (Ci)
binding sites, three high affinity Gal sites (colored circles), a neutral spacer (open
circles), hsp 70 basal promoter driving lacZ expression.

(A, B) Activity of the forkhead domain protein Sloppy-paired. (A) LacZ expression is
activated in a broad anterior domain consistent with the pattern of Slpl expression in the
blastoderm embryo (arrow). (B) LacZ expression is also activated in a ventral swath by
the Gal4 activator. Thus, Slpl appears to activate rather than repress in this context.

(C, D) Activity of Tramtrack (ttk) on hsp 70-lacZ reporter. Tramtrack is unable to repress
the Gal4 activators at various stages of embryogenesis.

(D, E) Activity of Cubitus interruptus (Ci) on hsp 70-lacZ reporter. Ci is unable to
repress the Gal4 activators at various stages of embryogenesis.

LacZ expression in A, B, C, and D is driven by the Gal4-Gal4 AD chimeric activator in
ventral regions where it is expressed and in E and F by full length 6314, which is
expressed ubiquitously throughout the embryo under the control of the actin5C enhancer.
Patterns of gene expression were visualized in 2-4 hour embryos by in situ hybridization
with digU labeled antisense lacZ probes. Embryos are oriented anterior to left; lateral
view (A, C, D, and E)) and field shot of embryos at various stages in development (B) are
shown.

206

 

SlplL
“Diff

MET

sitzk

our:

h ”or

lifSS

HG

 

 

,1
D M
0
{mm
t - O" ‘1'” a?
\ In a «a- n/ ‘

 

u. i I'm

Cl Gal4

:4.

Figure D-4: The activity of the transcription factors Sloppy-paired (Slpl),
Tramtrack (ttk) and Cubitus interruptus (Ci) on a reporter with three high affinity
Gal4sites.

207

Tramtrack (TTK): TTK has been proposed to act as a maternally provided
repressor of several pair-rule genes, such as even-skipped (eve). eve contains in its
promoter region binding sites for the trithorax-like transcription factor GAGA and TTK.
In transient expression experiments, it was demonstrated that GAGA activates
transcription from the eve stripe 2 promoter element and that TTK inhibits this GAGA-
dependent activation. Repression by TTK of the eve promoter requires its activation by
GAGA and depends on the presence of the POZ/BTB domains of TTK and GAGA
(Pagans et al., 2002). In the context of three high affinity Gal4 sites, TTK failed to block
lacZ expression (Figure D-4C, D). It is possible that GAGA-TTK interaction is required

for TTK function in vivo.

Cubitus interruptus (Ci): The Drosophila Gli homolog Cubitus interruptus (Ci)
controls the transcription of Hedgehog (Hh) target genes. A repressor form of Ci arises in
the absence of Hh signalling by proteolytic cleavage of intact Ci, whereas an activator
form of Ci is generated in response to the Hh signal. These different activities of Ci
regulate overlapping but distinct subsets of Hh target genes (Muller and Basler, 2000).
We did not observe any activity mediated by Ci at the stages of embryogenesis examined
(Figure D-4E, F). It is possible that in order to observe Ci activity, would require
investigation at later stages in development for example in the imaginal discs where Hh

signaling has been well characterized.

208

MATERIALS AND METHODS
1. GAL4 (aa1-93) - GAL4 AD (aa753-881)

A KpnI-Xbal fragment from pSCTEV GAL4 (1-93)— GAL4 (Seipel et al., 1992)
containing the reading frame for the yeast GAL4 activation domain (Gal4 AD) from
amino acid residues 753-881, was cloned into KpnI-Xbal cut pTwiggy (Amosti et al.,
1996a; Amosti et al., 1996c) vector, which contains the twist enhancer (2xPEe-Et)

element, twist basal promoter and the GAL4 DNA-binding domain from residues 1-93.

2. Fly stocks

Flies expressing the full-length yeast transcriptional activator Gal4 ubiquitously
throughout the embryo under the control of the actin5C enhancer, act5cGAL4/Cy0
(Stock # 4414) were also obtained from Bloomington. In order to obtain ubiquitous
activation of the lacZ reporter gene in the early (2-4 hour old) embryo, act5cGAL4/Cy0

females were crossed to males carrying the reporter transgene.

3. Reporter genes
The vector M2g3u2x-lacZ was modified to remove the two Giant binding sites, replacing

them with two sites for other transcriptional repressors with the following oligos:
SNAIL (Gray and Levine, 1996)
DA696: S’GGC CGC CAG CAA GGT GGT ACT AGA CAT CAG CAA GGT GA 3’

DA697: S’AGC TTC ACC TTG CTG ATG TCT AGT ACC ACC TTG CTG GC 3’

209

EVEN-SKIPPED (Li and Manley, 1998)

DA698: S’GGC CGC TCA ATT AAA TGA GTA CTA GAC ATC AAT TAA ATG

AA3’

DA699: 5’AGC TTT CAT TTA ATT GAT GTC TAG TAC TCA TTT AAT TGA GC

3’

TRAMTRACK (Kamashev et al., 2000)

DA700: 5’GGC CGC GGT CCT GCG TAC TAG ACA GGT CCT GCA 3’

DA701: 5’AGC TTG CAG GAC CTG TCT AGT ACG CAG GAC CGC 3’

BRINKER(Sa11er and Bienz, 2001)

DA702: S’GGC CGC GAG GCG CCA CCG TAC TAG ACA GAG GCG CCA CCA

3’

DA703: 5’AGC TTG GTG GCG CCT CTG TCT AGT ACG GTG GCG CCT CGC 3’

CUBITUS INTERRUPTUS (Muller and Basler, 2000)

DA704: 5’GGC CGC ACG GGC GGT CTG TAC TAG ACA ACG GGC GGT CTA

39

DA705: 5’AGC TTA GAC CGC CCG TTG TCT AGT ACA GAC CGC CCG TGC 3’

SUPPRESSOR OF HAIRLESS (Guss et al., 2001)

DA706: S’GGC CGC CGT GGG AAG TAC TAG ACA CGT GAG AAA 3’

210

DA707: S’AGC TTT TCT CAC GTG TCT AGT ACT TCC CAC GGC 3’
ENGRAILED (TRANSF AC database)

DA708: 5’GGC CGC ACT AAT TAG CGT ACT AGA CAA CTA ATT AGC A 3’
DA709: S’AGC TTG CTA ATT AGT TGT CTA GTA CGC TAA TT A GTG C 3’
RUNT (Kramer et al., 1999)

DA710: 5’GGC CGC TGC GGT CGT ACT AGA CAT GCG GTC A 3’ _

DA711: 5’AGC TTG ACC GCA TGT CTA GTA CGA CCG CAG C 3’

SLOPPY PAIRED (Yu et al., 1999)

DA714: S’GGC CGC TCT TCG ATG TCA ACA CAC CGA CCC TCT TCG ATG

TCA ACA CAC CA 3’

DA715: 5’AGC TTG GTG TGT TGA CAT CGA AGA GGG TCG GTG TGT TGA

CAT CGA AGA GC 3’

4. P-element transformation, crosses to reporter genes, and whole-mount in situ

hybridization of embryos.

P-element transformation vectors were introduced into the Drosophila gerrnline
by injection of yw"7 embryos as described (Small et al., 1992a). Embryos were collected
either directly from each transgenic reporter line or from a cross between a reporter line
and a line expressing the GAL4-activator chimeric proteins in the ventral regions or

ubiquitously throughout the embryo. The embryos were fixed and stained using

211

digoxigenin-UTP labeled antisense RNA probes to either lacZ or w as described (Small et

al., 1992b).

212

 

 

REFERENCES

Andrioli, L. P., Vasisht, V., Theodosopoulou, E., Oberstein, A. and Small, S. (2002).
Anterior repression of a Drosophila stripe enhancer requires three position-specific
mechanisms. Development 129, 4931—40.

Arnosti, D. N., Barolo, S., Levine, M. and Small, S. (1996a). The eve stripe 2 enhancer
employs multiple modes of transcriptional synergy. Development 122, 205-14.

Amosti, D. N., Gray, S., Barolo, S., Zhou, J. and Levine, M. (1996b). The gap protein
knirps mediates both quenching and direct repression in the Drosophila embryo. Embo J
15, 3659-66.

Arnosti, D. N., Gray, S., Barolo, S., Zhou, J. and Levine, M. (1996c). The gap protein
knirps mediates both quenching and direct repression in the Drosophila embryo. Embo J
15, 3659-66.

Aronson, B. D., Fisher, A. L., Blechman, K., Caudy, M. and Gergen, J. P. (1997).
Groucho-dependent and —independent repression activities of Runt domain proteins. Mol
Cell Biol 17, 5581-7.

Bang, A. G., Bailey, A. M. and PosakOny, J. W. (1995). Hairless promotes stable
commitment to the sensory organ precursor cell fate by negatively regulating the activity
of the Notch signaling pathway. Dev Biol 172, 479-94.

Bang, A. G., Hartenstein, V. and Posakony, J. W. (1991). Hairless is required for the
development of adult sensory organ precursor cells in Drosophila. Development 111, 89-
104.

Bang, A. G. and Posakony, J. W. (1992). The Drosophila gene Hairless encodes a novel
basic protein that controls alternative cell fates in adult sensory organ development.
Genes Dev 6, 1752-69.

Barolo, S. and Levine, M. (1997). hairy mediates dominant repression in the Drosophila
embryo. Embo J16, 2883-91.

Barolo, S., Stone, T., Bang, A. G. and Posakony, J. W. (2002). Default repression and
Notch signaling: Hairless acts as an adaptor to recruit the corepressors Groucho and
dCtBP to Suppressor of Hairless. Genes Dev 16, 1964-76.

Barolo, S., Walker, R. G., Polyanovsky, A. D., Freschi, G., Keil, T. and Posakony, J.
W. (2000). A notch-independent activity of suppressor of hairless is required for normal
mechanoreceptor physiology. Cell 103, 957-69.

Cadigan, K. M., Grossniklaus, U. and Gehring, W. J. (1994a). Functional redundancy:
the respective roles of the two sloppy paired genes in Drosophila segmentation. Proc Natl
Acad Sci USA 91, 6324-8.

213

 

Cadigan, K. M., Grossniklaus, U. and Gehring, W. J. (1994b). Localized expression
of sloppy paired protein maintains the polarity of Drosophila parasegments. Genes Dev 8,
899-913.

Campbell, G. and Tomlinson, A. (1999). Transducing the Dpp morphogen gradient in
the wing of Drosophila: regulation of Dpp targets by brinker. Cell 96, 553-62.

DiNardo, S. and O'Farrell, P. H. (1987). Establishment and refinement of segmental
pattern in the Drosophila embryo: spatial control of engrailed expression by pair-rule
genes. Genes Dev 1, 1212-25.

Don, S., Zeng, X., Cortes, P., Erdjument-Bromage, H., Tempst, P., Honjo, T. and
Vales, L. D. (1994). The recombination signal sequence-binding protein REP-2N
functions as a transcriptional repressor. Mol Cell Biol 14, 3310-9.

Fisher, A. L. and Caudy, M. (1998). Groucho proteins: transcriptional corepressors for
specific subsets of DNA-binding transcription factors in vertebrates and invertebrates.
Genes Dev 12, 1931-40.

Flores-Saaib, R. D. and Courey, A. J. (2000). Analysis of Groucho-histone interactions
suggests mechanistic similarities between Groucho- and Tupl-mediated repression.
Nucleic Acids Res 28, 4189-96.

Fujioka, M., Jaynes, J. B. and Goto, T. (1995). Early even-skipped stripes act as
morphogenetic gradients at the single cell level to establish engrailed expression.
Development 121, 4371-82.

Fujioka, M., Yusibova, G. L., Patel, N. H., Brown, S. J. and Jaynes, J. B. (2002). The
repressor activity of Even-skipped is highly conserved, and is sufficient to activate

engrailed and to regulate both the spacing and stability of parasegment boundaries.
Development 129, 4411-21.

Gray, S. and Levine, M. (1996). Short-range transcriptional repressors mediate both
quenching and direct repression within complex loci in Drosophila. Genes Dev 10, 700-
10.

Gray, S., Szymanski, P. and Levine, M. (1994). Short-range repression permits
multiple enhancers to function autonomously within a complex promoter. Genes Dev 8,
1829-38.

Grossniklaus, U., Pearson, R. K. and Gehring, W. J. (1992). The Drosophila SIOppy
paired locus encodes two proteins involved in segmentation that show homology to
mammalian transcription factors. Genes Dev 6, 1030-51.

Guss, K. A., Nelson, C. E., Hudson, A., Kraus, M. E. and Carroll, 8. B. (2001).
Control of a genetic regulatory network by a selector gene. Science 292, 1164-7.

214

Hasson, P., Muller, B., Basler, K. and Paroush, Z. (2001). Brinker requires two
corepressors for maximal and versatile repression in Dpp signalling. Embo J 20, 5725-36.

Hewitt, G. F., Strunk, B. S., Margulies, C., Priputin, T., Wang, X. D., Amey, R.,
Pabst, B. A., Kosman, D., Reinitz, J. and Amosti, D. N. (1999). Transcriptional
repression by the Drosophila giant protein: cis element positioning provides an
alternative means of interpreting an effector gradient. Development 126, 1201-10.

Howard, K. and Ingham, P. (1986). Regulatory interactions between the segmentation
genes fushi tarazu, hairy, and engrailed in the Drosophila blastoderm. Cell 44, 949-57.

Jazwinska, A., Kirov, N., Wieschaus, E., Roth, S. and Rushlow, C. (1999a). The
Drosophila gene brinker reveals a novel mechanism of Dpp target gene regulation. Cell
96, 563-73.

Jazwinska, A., Rushlow, C. and Roth, S. (1999b). The role of brinker in mediating the
graded response to Dpp in early Drosophila embryos. Development 126, 3323-34.

Jimenez, G., Paroush, Z. and Ish-Horowicz, D. (1997). Groucho acts as a corepressor
for a subset of negative regulators, including Hairy and Engrailed. Genes Dev 11, 3072-
82.

Jimenez, G., Pinchin, S. M. and Ish-Horowicz, D. (1996). In vivo interactions of the
Drosophila Hairy and Runt transcriptional repressors with target promoters. Embo J 15,
7088-98.

Kamashev, D. E., Balandina, A. V. and Karpov, V. L. (2000). Tramtrack protein-DNA
interactions. A cross-linking study. J Biol Chem 275, 36056-61.

Keller, S. A., Mao, Y., Struffi, P., Margulies, C., Yurk, C. E., Anderson, A. R.,
Amey, R. L., Moore, S., Ebels, J. M., Foley, K. et al. (2000). dCtBP-dependent and -
independent repression activities of the Drosophila Knirps protein. Mol Cell Biol 20,
7247-58.

Kramer, S. G., Jinks, T. M., Schedl, P. and Gergen, J. P. (1999). Direct activation of
Sex-lethal transcription by the Drosophila runt protein. Development 126, 191-200.

Lecourtois, M. and Schweisguth, F. (1998). Indirect evidence for Delta-dependent
intracellular processing of notch in Drosophila embryos. Curr Biol 8, 771-4.

Li, C. and Manley, J. L. (1998). Even-skipped represses transcription by binding TATA
binding protein and blocking the TFIID-TATA box interaction. Mol Cell Biol 18, 3771-
81.

Mannervik, M., Nibu, Y., Zhang, H. and Levine, M. (1999). Transcriptional
coregulators in development. Science 284, 606-9.

215

 

Manoukian, A. S. and Krause, H. M. (1992). Concentration-dependent activities of the
even-skipped protein in Drosophila embryos. Genes Dev 6, 1740-51.

Manoukian, A. S. and Krause, H. M. (1993). Control of segmental asymmetry in
Drosophila embryos. Development 118, 785-96.

Minami, M., Kinoshita, N., Kamoshida, Y., Tanimoto, H. and Tabata, T. (1999).
brinker is a target of Dpp in Drosophila that negatively regulates Dpp-dependent genes.
Nature 398, 242-6.

Morel, V., Lecourtois, M., Massiani, 0., Maier, D., Preiss, A. and Schweisguth, F.
(2001). Transcriptional repression by suppressor of hairless involves the binding of a
hairless-dCtBP complex in Drosophila. Curr Biol 11, 789-92.

Morel, V. and Schweisguth, F. (2000). Repression by suppressor of hairless and
activation by Notch are required to define a single row of single-minded expressing cells
in the Drosophila embryo. Genes Dev 14, 377-88.

Muller, B. and Basler, K. (2000). The repressor and activator forms of Cubitus
interruptus control Hedgehog target genes through common generic gli-binding sites.
Development 127, 2999-3007.

Nibu, Y., Zhang, H., Bajor, E., Barolo, S., Small, 8. and Levine, M. (1998a). dCtBP
mediates transcriptional repression by Knirps, Kruppel and Snail in the Drosophila
embryo. Embo J 17, 7009-20.

Nibu, Y., Zhang, H. and Levine, M. (1998b). Interaction of short-range repressors with
Drosophila CtBP in the embryo. Science 280, 101-4.

Nibu, Y., Zhang, H. and Levine, M. (2001). Local action of long-range repressors in the
Drosophila embryo. Embo J 20, 2246-53.

Pagans, S., Ortiz-Lombardia, M., Espinas, M. L., Bernues, J. and Azorin, F. (2002).
The Drosophila transcription factor trarntrack (TTK) interacts with Trithorax-like
(GAGA) and represses GAGA-mediated activation. Nucleic Acids Res 30, 4406-13.

Petcherski, A. G. and Kimble, J. (2000). Mastermind is a putative activator for Notch.
Curr Biol 10, R471-3.

Poortinga, G., Watanabe, M. and Parkhurst, S. M. (1998). Drosophila CtBP: a Hairy-
interacting protein required for embryonic segmentation and hairy-mediated
transcriptional repression. Embo J 17, 2067-78.

Qi, H., Rand, M. D., Wu, X., Sestan, N., Wang, W., Rakic, P., Xu, T. and Artavanis-
Tsakonas, S. (1999). Processing of the notch ligand delta by the metalloprotease
Kuzbanian. Science 283, 91-4.

216

Saller, E. and Bienz, M. (2001). Direct competition between Brinker and Drosophila
Mad in Dpp target gene transcription. EMBO Rep 2, 298-305.

Schweisguth, F. and Posakony, J. W. (1994). Antagonistic activities of Suppressor of
Hairless and Hairless control alternative cell fates in the Drosophila adult epidermis.
Development 120, 1433-41.

Seipel, K., Georgiev, O. and Schaffner, W. (1992). Different activation domains
stimulate transcription from remote ('enhancer') and proximal ('promoter') positions.
EmboJll, 4961-8.

Small, S., Blair, A. and Levine, M. (1992a). Regulation of even-skipped stripe 2 in the
Drosophila embryo. Embo J 11, 4047-57.

Small, S., Blair, A. and Levine, M. (1992b). Regulation of even-skipped stripe 2 in the
Drosophila embryo. Embo J 11, 4047-57.

Struhl, G. and Adachi, A. (1998). Nuclear access and action of notch in vivo. Cell 93,
649-60.

Struhl, G. and Greenwald, I. (1999). Presenilin is required for activity and nuclear
access of Notch in Drosophila. Nature 398, 522-5.

Sundqvist, A., Sollerbrant, K. and Svensson, C. (1998). The carboxy-terrninal region
of adenovirus ElA activates transcription through targeting of a C-terminal binding
protein-histone deacetylase complex. FEBS Lett 429, 183-8.

Tolkunova, E. N., Fujioka, M., Kobayashi, M., Deka, D. and Jaynes, J. B. (1998).
Two distinct types of repression domain in engrailed: one interacts with the groucho
corepressor and is preferentially active on integrated target genes. Mol Cell Biol 18,
2804-14.

Wheeler, J. C., VanderZwan, C., Xu, X., Swantek, D., Tracey, W. D. and Gergen, J.
P. (2002). Distinct in vivo requirements for establishment versus maintenance of
transcriptional repression. Nat Genet 32, 206-10.

Yu, Y., Yussa, M., Song, J., Hirsch, J. and Pick, L. (1999). A double interaction screen
identifies positive and negative ftz gene regulators and ftz-interacting proteins. Mech Dev
83, 95-105.

Zhang, H. and Levine, M. (1999). Groucho and dCtBP mediate separate pathways of

transcriptional repression in the Drosophila embryo. Proc Natl Acad Sci U S A 96, 535-
40.

217