n...

.3?

 

Munfififlw ,

3...

$3...

I .3.

o
Lt... -
~e .3
.i(.

an; .
b... “ER“... :3.
«flag UMth
. K . . .nhmw.

.
n
I
a... up...
viii (-
.1.va34.
.u he.
at I.
. “Fusmwanviuuafifi:
5.1 T.” 6. 1.4.

517...! \u

4 .r 1»...qu .

. . fine}... Mafiaflww
. awfthstei§ulah£ 1
..>:3....s.§9mw.x..
.33.: .943 v...
'0. .1. 3m ‘5.
Wt 32‘
an s .

b 3"..-
ahfiau.

.3... .
Sr;.r.&:...a(oi
S» ..
l“‘ I
ska...
.. .Q

.
€1.33: . . 4. :r
1.53.2!!(‘1 his...)

mm...» «.4...

q.

2?. 3.3,... mun

.Kwflti. 5.5! r“. at :9
3.3;! 1}: . .:
‘0.“ alt... I‘ig' s I
1 5.553%!“ 1W.“ S. ......
. 5“: t9 :9 )4} t 1.!
‘il’a‘i‘nuc .4"
. 4"."‘i’fi . ’n .34.
59.11 3.
1515..." 9.3.. 3.3
4‘ Seihuxv [‘3
"3: '3"!
'1...

£1.59.
3. kit43! 4.
:1... 3.
.33 .v.
ire... .2
i w... . a...

 

13m
@507

 

LIBRARY
Michigan State
University

 

This is to certify that the
dissertation entitled

THE ROLE OF THE VP16ADITBP INTERACTION IN
TRANSCRIPTIONAL ACTIVATION

presented by

Dean Daniel Shooltz

has been accepted towards fulfillment
of the requirements for the

PhD. degree in Biochemistry and Molecular
Biology

 

 

 

'7 'Major Proms Qfghafiure

2}; @1245: A” M;

Date

MS U is an Affirmative Action/Equal Opportunity Institution

----¢-.-

A,_ _--.—-—.-.--—----._-—---o-a---nnr----¢-.-s -

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.

 

DAIEDUE

DATEDUE

DAIEDUE

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

6/07 p:/ClRC/Dale0ue.indd-p.1

THE ROLE OF THE VP16AD2TBP INTERACTION
IN TRANSCRIPTIONAL ACTIVATION

By

Dean Daniel Shooltz

A DISSERTATION

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

DOCTOR OF PHILOSOPHY

Department of Biochemistry and
Molecular Biology

2007

ABSTRACT

THE ROLE OF THE VP16AD:TBP INTERACTION
IN TRANSCRIPTIONAL ACTIVATION

By

Dean Daniel Shooltz

The regulation of transcription is central to development, homeostasis, and disease. Key
factors in the regulation of gene expression include transcriptional activator proteins.
These proteins provide the functional link between cis-acting DNA regulatory elements
and transcriptional responses. While the principles of cis-element recognition are well
established, the mechanisms by which transcriptional activation domains function remain
elusive. Intense efforts have been directed at delineating the targets of transcriptional
activation domains, and many potential targets have been identified. However, the in
vivo functional relevance of many of these connections remains poorly characterized.
This dissertation describes efforts to characterize the interaction of the VP16 activation

domain (VP16AD) with the TATA-binding protein (TBP).

The strength of the VP16AD:TBP interaction correlates with the ability of VP16AD to
activate transcription, suggesting that TBP is a relevant target of VP16AD. However,
VP16AD, TBP, and DNA appear unable to form a three-way complex, raising the
question of how VP16AD can activate transcription through interactions with TBP. This
dissertation explored two proposed models for VP16AD-TBP interactions. First, I
examined whether VP16AD could increase the orientational specificity of TBP bound to

its cognate promoter element. The results of this study indicated that TBP is intrinsically

highly oriented on the TATA DNA sequence, and that no TBP—orienting activity could be
detected in VP16AD. The finding that TBP is intrinsically oriented implies that current

models of the assembly of the transcriptional machinery need to be updated.

A second proposed mechanism of VP16AD-TBP interaction posits that a cascade of
interactions with TBP is involved in transcriptional activation. In this model, VP16AD
competes with a negative regulator, binds TBP transiently, and then releases TBP for
DNA association. This dissertation describes my efforts to structurally characterize
VP16AD when bound to TBP. Although a structural model was not determined, the
efforts have resulted in the design and characterization of a protein complex that may
serve as a foundation for further study. The experimental efforts have additionally

resulted in an improved method of TBP purification.

ACKNOWLEDGMENTS

I would like to thank my mentor, Steve Triezenberg, for sharing with me his seemingly
unlimited enthusiasm for science in general and for molecular biology in particular. I
appreciate his dedication to training me to think and function as a scientist. I will fondly

remember my time as a graduate student in his lab.

I would also like to thank the past and present members of the Triezenberg laboratory. In
my early days as a graduate student I relied heavily on the advice of Saren Ottosen,
Yaopan Mao, Billy Yang, Kostas Vlachonasios and Francisco Herrera. I benefited from
their in-depth answers and guidance, and I fondly remember the times we shared in the
lab. I have further enjoyed the camaraderie of the lab as it has continued with Amy Hark,

Kanchan Pavangadkar, Sebla Kutluay, Nate Lord and James Doroghazi.

1 would also like to thank my family and friends. They have been a constant source of
support and encouragement, and they have helped me through the various tough times in
my graduate career. I value their understanding of the time that I needed to spend in the

lab over the past few years.

iv

TABLE OF CONTENTS

LIST OF TABLES ............................................................................................................. vii
LIST OF FIGURES .......................................................................................................... viii
LIST OF ABBREVIATIONS .............................................................................................. x
Chapter I: Literature review ................................................................................................. l
Transcriptional regulation in eukaryotes ................................................................ 1
Information flow: activating the activators ............................................................ 2
Activator association with promoters ..................................................................... 4

Core promoter architecture and activator specificity ............................... 4

Enhancers and upstream activating sequence elements ........................... 5

Actions of activators at promoters .......................................................................... 8
Stepwise recruitment of basal transcription machinery ........................... 8

Recruitment of RNA pol II holoenzyme ................................................ 11

Recruitment of the mediator complex .................................................... 12

Recruitment of chromatin- modifying coactivators ................................ 14

The handoff mechanism ......................................................................... 15

Modulation of TBP orientation .............................................................. 17

Leaving a mark: activator effects after initiation ................................... 18

Structure of transcriptional activators .................................................................. 19
DNA-binding domains ........................................................................... 20

Activation domains ................................................................................ 21

Central hypothesis ................................................................................................ 23

Chapter II: Purification of the TATA binding protein ....................................................... 24
Introduction .......................................................................................................... 24
Materials and methods ......................................................................................... 25

Results and discussion .......................................................................................... 28
Chapter III: The orientation of TBP on DNA .................................................................... 34
Introduction .......................................................................................................... 34
Materials and methods .......................................................................................... 39

Results .................................................................................................................. 44

Discussion ............................................................................................................ 60

Chapter IV: Attempts to determine the structure of VP16C bound to TBP ...................... 65
Introduction .......................................................................................................... 65

VP16 complexed with TBP .................................................................................. 70
Rationale ................................................................................................. 70

Materials and methods ............................................................................ 71

Results and discussion ............................................................................ 74

VP 1 6C-TAND2 bound to TBP ............................................................................ 77
Rationale ................................................................................................. 77

Materials and methods ............................................................................ 79

Results .................................................................................................... 82

Discussion .............................................................................................. 89
VP16C-TAND2-TBP fusion protein .................................................................... 90
Rationale ................................................................................................ 90

Materials and methods ........................................................................... 91

Results .................................................................................................... 94

Discussion ............................................................................................ 103

Chapter V: NMR Analysis of the VP16-TAND2-TBP Complex .................................... 105
Introduction and rationale ................................................................................... 105
Materials and methods ........................................................................................ 107

Results ................................................................................................................. 113
Discussion ........................................................................................................... 130
Chapter VI: Discussion and Prospects for Further Study ................................................ 132
Crystallization of a complex of VP16C and TBP ............................................... 132

NMR analysis of VP16C-TAND2 bound to TBP ............................................... 135

TBP orientation ................................................................................................... I36

TBP purification .................................................................................................. 139
Appendix A: HSQC spectra of VP16C-TAND2 complexed with TBP .......................... 140
Appendix B: Plasmid maps and oligonucleotide sequences ........................................... 146
List of references .............................................................................................................. 150

vi

LIST OF TABLES

Table 1. TBP orientation measurements ........................................................................... 54
Table 2. Intensity of NMR peaks in different spectra ..................................................... 141
Table 3. Comparison of NMR spectra ............................................................................ 142
Table 4. Oligonucleotide sequences ............................................................................... 147

vii

Figure 1.
Figure 2.
Figure 3.
Figure 4.
Figure 5.
Figure 6.
Figure 7.
Figure 8.

Figure 9.

Figure 10.
Figure 11.
Figure 12.
Figure 13.
Figure 14.
Figure 15.
Figure 16.
Figure 17.
Figure 18.
Figure 19.
Figure 20.

Figure 21.

LIST OF FIGURES

Purification of full- length yeast TBP ................................................................ 30
Affinity cleavage methodology ......................................................................... 45
Affinity cleavage patterns on top DNA strand .................................................. 47
Affinity cleavage patterns on bottom DNA strand ............................................ 48
DNA cleavage patterns mapped onto the cleavage probe sequence ................. 50
Method of DNA fragment pattern quantitation ................................................. 52
Anomalous peak in bottom strand DNA fragment analysis .............................. 53
Affinity cleavage pattern in the presence of Gal4-VP16 ................................... 56
Overlay of TBP-OP and TBP-OP (F2271) top strand cleavage patterns ........... 59

Handoff mechanism of TFIID-TFIIA-DNA assembly .................................... 69
Purification of VP16C and TBP. . . . .. .............................................................. 75
Design and purification of a VP16C-TAND2 fusion protein .......................... 83
Association of GST-VP16C-TAND2 and TBP ............................................... 85
Purification of the VP16C-TAND2zTBP complex .......................................... 86
Design and construction of a fusion of VP16C-TAND2-linker—TBP ............. 95
Purification of GST-VP16C-TAND2-linker-TBP fusion proteins .................. 97
Monomeric and dimeric states of VP16C-TAND2- linker-TBP fusions ......... 99
Purification of GST-VPl6C-TAND2-TBP fusion protein ............................ 101
Crosslinking of VP16C-TAND2-Cys to Cys-TBP ....................................... 115

Crosslinking and purification of VP16C-TAND2-TBP complexes .............. 117
HSQC spectrum of fresh sample at 10 mM KCl ........................................... 119

viii

Figure 22.
Figure 23.
Figure 24.
Figure 25.
Figure 26.
Figure 27.
Figure 28.
Figure 29.
Figure 30.

Figure 31.

HSQC spectrum of aged sample at 10 mM KC1 ........................................... 120
Optimization of buffer conditions ................................................................. 123
HSQC spectrum of crosslinked complex in low salt conditions ................... 124
Co-purification of VP16NC and TBP.. ........................................................ 127
Analysis of the stability of a mixture of VP16NC and TBP ......................... 128
HSQC spectrum of aged sample at 10 mM KCl ........................................... 143
HSQC spectrum of fresh sample at 10 mM KCl ........................................... 144
HSQC spectrum of crosslinked complex in low salt conditions ................... 145
Plasmid pDS42-10 ......................................................................................... 148
Plasmid family pDSEl l7-n ........................................................................... 149

AdMLP
AEBSF
BRE
BSA
DPE
DTT

F PLC
GSH
GST
GTF
GVT
HPLC
HSQC
IAAOP
lnr
IPTG
LB
NOESY
PAGE
PCR

PIC

LIST OF ABBREVIATIONS

Adenovirus major late promoter
4-(2-Aminoethyl) benzenesulfonyl fluoride hydrochloride
TFIIB recognition element

Bovine serum albumin

Downstream promoter element

Dithiothreitol

Fast protein liquid chromatography
Glutathione

Glutathione-S-transferase

General transcription factor
GST-VP16C-TAND2 fusion protein

High performance liquid chromatography
Heteronuclear single quantum correlation
5—Iodoacetamido- 1 ,10-phenanthroline

Initiator promoter sequence

Isopropyl-beta- D-thiogalactopyranoside
Luria-Bertani medium

Nuclear Overhauser enhancement spectroscopy
Polyacrylamide gel electrophoresis

Polymerase chain reaction

Pre-initiation complex

PMSF
RNA p01 11
SDS
SEC
TAF
TAND
TBP
TBP-OP
TFII
VP16
VP16AD
VP16C

VP16N

Phenylmethylsulphonylfluoride

RNA polymerase 11

Sodium dodecyl sulfate

Size exclusion chromatography
TBP-associated factor

TAF—l N-terminal domain
TATA-element binding protein

TBP treated with IAAOP
Transcription factor for RNA pol II
Virion protein 16 from herpes simplex virus
VP16 activation domain

VP 1 6 C-terminal activation subdomain

VP 1 6 N-terminal activation subdomain

xi

Chapter 1

Literature Review1

Transcriptional regulation in eukaryotes.

The genome of any organism contains not only the information specifying the primary
sequence of the many proteins required for growth, development and defense, but also
regulatory information for determining when and where a given gene is to be expressed.
The link between cis-acting regulatory DNA sequences and the expression of the
corresponding genes is made by trans-acting transcriptional activator and repressor
proteins. These regulatory proteins must firlfill two functions. They must accurately
recognize which genes they are to regulate, and once associated with their target gene
they must either stimulate (for activator proteins) or inhibit (for repressor proteins)
transcription by RNA polymerase II (RNA pol II). The mechanisms for accomplishing
these two functions are built on principles of modular design and combinatorial logic.
For any given gene, the cis-regulatory elements comprise a set of short DNA sequences
that together dictate the appropriate expression of that gene. The trans-regulatory
proteins that bind such elements bring particular domains that interact with either
chromatin- modifying enzymes or components of the general transcription machinery.
These protein:protein interactions are highly diverse, both in the array of activators and
target proteins present in a particular cell and also in that any given activator might bind

to several target proteins.

 

I Part of this chapter was published in “Herrera FJ, Shooltz DD and Triezenberg SJ (2004). Mechanisms of
transcriptional activation in eukaryotes. Handbook of Experimental Pharmacology. Vol 16623-31”

Information flow: activating the activators

The specificity of gene regulation in a given cell at different times or in response to
various external or internal signals demands that the panoply of transcription factors in
that cell should not all be active all of the time. This obvious observation implies that the
activity of activators must be regulated, and indeed this principle has been demonstrated
in a surprisingly diverse array of mechanisms. Some activators are regulated by control
of their cellular localization from cytoplasm to nucleus. For example, NF —KB, an
activator of cytokine- induced gene expression, is retained in the cytoplasm by interaction
with IKB. Kinases triggered to action by ligand-bound cytokine receptors phosphorylate
IKB, which then releases NF-KB to traverse to the nucleus (reviewed by (40)). The sterol
regulatory element binding protein (SREBP) proprotein is anchored in the endoplasmic
reticulum by two trans-membrane domains. Proteolytic cleavage releases SREBP from
the endoplasmic reticulum and allows its translocation to the nucleus where it activates
genes involved in sterol biosynthesis (reviewed by (138)). Some nuclear hormone
receptors are localized in the cytoplasm until binding to the relevant ligand allows

translocation to the nucleus to activate transcription of target genes (reviewed by (110)).

The activities of transcriptional activators can also be regulated by covalent modifications
including phosphorylation, methylation, and acetylation, and ubiquitinylation. A

particularly impressive example is the tumor suppressor protein p53, which is subjected

to multiple covalent modifications that tightly regulate its function and stability
(reviewed by (12)). Ubiquitinylation of p53 by Mdm2 (an E3 ubiquitin ligase) and
subsequent destruction in the proteasome is responsible for maintaining p53 protein at
low levels. Phosphorylation of p53 in response to ionizing radiation stabilizes the p53

protein. Acetylation of p53 also stimulates its transcriptional activity.

One modification that has garnered much interest lately is the addition of ubiquitin to
activators by ubiquitin ligases. Although long known as a “tag” for protein degradation
by the proteasome, ubiquitin (and a related polypeptide termed SUMO) is increasingly
recognized for non-proteolytic roles as well. Most intriguing has been the suggestion that
ubiquitin is required de facto for the function of activators in yeast and mammalian cells
(reviewed by (115)). One clue is that the protein motif that targets activators for
degradation overlaps with the transcriptional activation domain in proteins including
various nuclear hormone receptors, the tumor suppressor p53, the proto-oncogenes c-myc
and c- fos, and the VP16 protein (148). Mutations in specific E3 ubiquitin ligases can
stabilize an activator protein but reduce its transcriptional activity (147). Moreover,
ectopic addition of a single ubiquitin moiety to an activator can enhance its transcription
activity in a non-proteolytic manner. Counterpart to these clues regarding ubiquitin is
evidence that components of the proteasome can also be found associated with
transcriptional promoters in conjunction with gene regulation (36, 42, 128). Although the
explicit roles of ubiquitin and the proteasome in the mechanism of transcriptional
activation remain uncertain, one model proposes that the addition of ubiquitin may

promote interaction of activators with key target proteins. Subsequent ligation of

additional ubiquitin groups may lead to the degradation of the activator by the

proteosome, providing a rapid downregulation of the activating signals (reviewed by

(97)).

Activator association with promoters

Core promoter architecture and activator specificity

The cis-acting regulatory sequences for genes transcribed by RNA pol II typically
comprise a combination of core promoter elements (reviewed in (156)) which serve as
the binding site for the basal transcriptional factors and define the translational startpoint.
The most obvious feature in many core promoters is the TATA box, located about 30 bp
upstream of the transcription start site. Surrounding the start site itself may be found an
initiator (Inr) element. A downstream promoter element (DPE) has been defined in
Drosophila and in mammalian genes, albeit not in yeast (63). Furthermore, a TFIIB
recognition element (BRE), flanking the TATA box, has been described in some

organisms (156).

Differences among core promoters of various genes, with respect to the presence and
strength of these core elements, have pronounced effects on how those promoters respond
to particular transcriptional activators. This specificity may allow a particular enhancer

to differentially regulate various target genes, and may allow a particular core promoter

to selectively respond to different enhancers. For example, the Spl activation domain
strongly activates core promoters containing TATA elements, Inr elements, or both,
whereas the Gal4- VP16 activator is most effective at a core promoter with both TATA
and Inr (32). In an ”enhancer trapping” study in Drosophila, many enhancers were able
drive expression from both TATA dependent and DPE dependent promoters, but some
enhancers preferentially activated either one or the other core promoter (18). The
mechanistic differences leading to enhancer-core promoter specificity may involve the
recruitment of transcriptional cofactors that display core promoter specificity. For
example, the transcriptional cofactor NC2 represses transcription from TATA dependent
promoters, but activates transcription from DPE dependent promoters (176). Thus, either
direct or indirect recruitment of transcriptional cofactors by an activator may
differentially regulate transcription from different core promoters, leading to activator-

promoter specificity.

Enhancers and upstream activating sequence elements

In addition to the core promoter elements, transcription of many genes depends on cis-
acting regulatory elements termed enhancers or upstream activating sequences, which
provide the binding sites for transcriptional activators. These DNA elements can vary in
sequence and affinity for a particular DNA binding domain, and may exist in promoter-
proximal locations or hundreds to thousands of basepairs upstream or downstream of a

promoter. The consensus sequences of various cis-acting regulatory elements and the

proteins that recognize and bind to those elements have been cataloged in various
databases now available at intemet websites, including the Eukaryotic Promoter Database
at www.epd.isb-sib.ch (132) and the TRANSF AC database at www.gene-regulation.com

(108).

Combinations of activator binding sites may be clustered into more complex regulatory
elements. In such cases, cooperative binding of activators leads to the formation of large
DNA-protein structures termed enhanceosomes, resulting in synergistic effects on
transcription. In the prototypical virus- inducible IFN—B enhanceosome (reviewed in
(112)), a cluster of three different activator binding sites direct the expression of IFN-B in
response to viral infection. However, none of the activator binding sites act alone; only
the combination of all three activator binding sites recapitulates the logic necessary to
drive proper expression and specificity of the IFN-B gene. In this model system, the
enhancer represents not simply the sum of individual activator functions, but rather an
integration of inputs from different sources interpreted by the particular combination of

transcription factors present at the enhancer.

Taking this organizational theme one step further, many genes may have multiple
enhancers, each of which is poised to respond to particular developmental, growth, or
environmental signals. Each of these independent enhancers may be simultaneously
signaling to the core promoter either to stimulate or repress transcription, depending on
the signal inputs received by the regulatory proteins that bind there. This diverse and

sometimes conflicting information must be integrated and interpreted at the promoter to

make a final decision on whether or not transcription is to proceed. This “information
display” model for genes with multiple enhancers has arisen from studies of
developmentally related genes in Drosophila (86), but will likely be relevant for many

genes in humans as well.

With the increasing availability of genomic sequences for prominent experimental
organisms, computational analysis for identifying cis-acting regulatory sequences has
become a growth industry. In some cases, these searches focus on particular cis
regulatory elements, such as the estrogen response elements in mammalian genomes (4).
Other programs are designed for broader application, searching for sites corresponding to
any of the transcription factors in the TRAN SF AC database and combining site searches
to increase the likelihood of identifying legitimate regulatory regions rather than
idiosyncratic consensus sequence matches (10, 47, 69). These computational approaches
yield results that still must be validated by direct evidence of the function of putative
elements in gene regulation and their interaction with specific transcription factors.
Although this is often done on a case-by-case basis, either by mutational analysis of the
cis elements or by in vitro binding assays, more global assessments are also now possible.
For several transcription factors in yeast (141) and in mammalian cells (140, 172),
genomic mapping of transcription factor binding sites has been accomplished by
combining chemical crosslinking and immunoprecipitation (so-called chromatin IP or
“ChIP” assays) with DNA microarrays comprising intergenic or putative regulatory

sequences.

Actions of activators at promoters

Once localized to a promoter, a transcriptional activator can interact with a number of
different targets, including RNA pol II, the basal transcription factors, the mediator
complex, coactivators, and chromatin-remodeling machinery. A common theme in
models of activation is recruitment, where a promoter-bound activator recruits either a
component of the transcriptional machinery or a transcriptional cofactor. This model is
supported by evidence of direct physical interactions of activators with basal transcription
factors, and by activator bypass experiments (reviewed in (134)). In the latter
experiments, a component of the transcriptional machinery is fused directly to a DNA
binding domain, and this artificial recruitment serves to activate transcription.
Alternatively, in a variation of recruitment, a transcriptional activator may modulate the
activity of components of the transcriptional machinery, facilitating the assembly of the

preinitiation complex.

Stepwise recruitment of basal transcription machinery

Transcription of protein-coding genes requires the assembly of a preinitiation complex
(PIC) comprising RNA pol II, the general transcription factors (GTFs), and a number of
associated factors. In one stepwise model of PIC assembly, the TATA-binding protein

(TBP)-containing TFIID complex binds to a promoter, followed by TFIIA, TFIIB, TFIIF

and RNA pol II, and TFIIE and TFIII-l (16). In this model, any step of PIC assembly
might be rate limiting, and the recruitment of GTFs by association with activators may

facilitate assembly.

TFIID, TBP and TAFs

The TFIID protein complex comprises TBP and several TBP-associated factors (TAFs)
(17, 135). TBP binds selectively to the TATA core promoter element, while the TAFs
extend the footprint to include the Inr and DPE elements. Studies in vitro show that

although TBP is sufficient for basal transcription, the TAFs are required for activated

transcription (31, 106, 149).

TBP can bind directly to transcriptional activation domains, as demonstrated by in vitro
binding assays using a wide range of activator proteins, and mutations in activators that
weaken activation also weaken interactions with TBP (56, 153). While these results
suggest that activators simply recruit TBP, other evidence suggests that the mechanism is
more complicated. Two “indirect recruitment” models of activator interaction with TBP

are described below.

Some TAFs have a direct affinity for certain transcriptional activators, suggesting a role
in recruitment or modulation of activity. For example, the glutamine-rich activators Spl,

NF AT, and CREB interact with the TAF 4 protein from Drosophila or human cells (21, 35,

41, 71, 144, 178). The acidic VP16 and p53 activation domains can interact with TAF9
(43, 75, 166). However, these interactions cannot always be interpreted to imply an

effect on TFIID, since a significant number of these TAF proteins are present in other
protein complexes besides TFIID that nonetheless influence transcriptional activation (48,

175).

TFIIA, TFIIB, TFIIF, and TFIIH

TFIIA is a positive cofactor in PIC assembly, as it binds cooperatively with TFHD at
TATA DNA elements. TFIIA also functions as an antirepressor, inhibiting the TBP-
DNA destabilizing actions of Motl and NC2 (reviewed in (135)). The formation of the
ternary TFIID-TFIIA-DNA complex is a rate- limiting step in PIC formation, and
activators can enhance this step (76, 93). Some evidence points to direct association of

the VP16 activation domain (VP16AD) with subunits of TFIIA (76, 77).

TFIIB also stabilizes the TBP-TATA complex, and serves as a docking site for other
components of the PIC. Several activators, including VP16AD, have been shown to bind
TFIIB with a high affinity (96), and activator-TFIIB connections have been implicated in
transcriptional activation (143). Interaction with VP16AD has been shown to alter the
conformation of TFIIB, possibly priming it for incorporation into the PIC (52, 142) or
altering TFIIB-DNA contacts (33). Although this evidence is points to TFIIB as a

potential activator target, other reports have failed to find evidence supporting this

10

association (43, 153), and thus the role for TFIIB as a target for VP16AD remains

controversial.

TF 11H, which contains both protein kinase and nucleic acid helicase activities, also
appears as a target for activation domains. The activation domains of VP16, p53, and
EZF 1 can interact with TFIIH (130, 179) and recruitment of TFIIH may stimulate
promoter escape (87), but no clear evidence exists that activators stimulate the enzymatic
activities of TFIIH. TFIIF and TFIIB might also be targets for activation domains. The
serum response factor (SRF) interacts with the RAP74 subunit of TFIIF (60), and F os-Jun
dimers can interact with both TFIIF and TFIIE (107). Although the mechanistic
implications of these interactions are not fully developed, the dual role of TFIIF as both
an initiation and elongation factor suggests the possibility that activators modulate

promoter escape or elongation in addition to assembly of the preinitiation complex.

Recruitment of RNA pol 11 holoenzyme

Many of the models described in the preceding sections are predicated on the premise
that transcriptional activation involves a sequential recruitment of the basal transcription
factors and RNA pol II to form the PIC at the target promoter. This premise was
challenged, however, by the biochemical purification from yeast cells of an
extraordinarily large protein complex comprising RNA pol II stably associated with a

subset of GTFs together with additional polypeptides from the mediator complex, as

11

described below (74). Certain transcriptional activators were shown capable of recruiting
this “holoenzyme” to a promoter in a manner sufficient to achieve transcriptional
activation in vitro (54, 74). The model arising from these observations is that rather than
separately and sequentially recruiting each general transcription factor, activation might
more simply involve recruitment of the distinct TFIID and holoenzyme complexes.
Similar RNA pol 11 holoenzyme complexes have been also purified from human cells (81)
suggesting that the recruitment of the holoenzyme by activators might be an
evolutionarily conserved mechanism (reviewed by (51)). Kinetic and thermodynamic
questions arising from the two competing models have not been fully resolved. How
could the stepwise assembly occur quickly enough (given diffusion parameters for each
component) for efficient transcriptional activation? And yet, how can a complex the size
of the holoenzyme be translocated to specific genes at specific times quickly enough to

respond to transcriptional activation?

Recruitment of the mediator complex

The mediator complex, first identified as a component of the yeast RNA pol 11
holoenzyme (74), comprises about 20 subunits forming three major domains (Gal 11,
Med9/10 and Srb modules) that wrap around the RNA pol II (reviewed by (11, 116)).
Homologues of the yeast mediator subunits and similar protein complexes have since
been identified in a wide range of organisms. Mammalian protein complexes resembling

yeast mediator were described independently by several laboratories using biochemical

12

purifications of proteins stably bound to different activator proteins (reviewed by (103,
136)). These different purifications lead to very similar complexes (variously termed
TRAP, DRIP, ARC, CRSP, SMCC and PC2) suggesting that different activators bind the
same or highly related human mediator complexes. The slight differences in protein
compositions of these human mediators might represent variations due to differences in
the biochemical purifications or might represent different forms of the mediator complex

that associate with different activators.

Several activators are known to interact physically and functionally with the mediator
complex and the particular mediator subunits involved in these interactions are being
identified. For example, the p53 tumor suppressor protein interacts with the TRAP80
subunit of the human mediator complex whereas the thyroid hormone receptor and
PPARyZ interact with TRAP220 (38, 57). The VP16AD may associate either with
TRAP80 or with ARC92 (57, 114). lnterferon- stimulated transcription depends on an
interaction of STAT3 with the DRIP150 mediator component (89). These and other
examples indicate that the mediator complex can be considered as a modular interface
connecting activators with RNA pol 11 allowing the integration of different signals during

transcriptional activation.

13

Recruitment of chromatin-modifying coactivators

Transcription activation must overcome the physical barriers presented by the packaging
of eukaryotic DNA into chromatin. Chromatin compaction can mask activator binding
sites, mask core promoter elements, and can impede the actions of RNA polymerase. It is
now clear that the histone proteins upon which DNA is wrapped are not only a static
scaffold for the compaction of DNA, but rather they participate actively in the regulation
of gene expression. Transcriptional activators can affect chromatin by recruiting
members from two general classes of chromatin- modifying coactivator proteins. One
class comprises enzymes that covalently modify amino acids within the histones
themselves. These modifications then either directly alter chromatin structure, or serve as
recognition signals for binding additional proteins that modulate that structure (55).
Covalent modifications that have been identified in histone proteins include acetylation,
methylation, phosphorylation, ubiquitinylation, and ADP- ribosylation (reviewed by (8)).
The panoply of such modifications on the various histones has been likened to a “code”
(59, 158) that, when deciphered and integrated, signals whether and how strongly a given

gene is to be expressed.

A second general class of chromatin- modifying transcriptional coactivators comprises the
ATP-dependent chromatin remodeling complexes. The mechanisms of these remodeling
activities are not yet fully understood, but include local and stable alterations of the
DNA-histone contacts leading to sliding of nucleosomes along DNA or transfer of

nucleosomes from one DNA to another in trans. These protein complexes can also alter

14

the superhelicity of DNA in vitro (reviewed by (118)). Repositioning of nucleosomes
may alleviate chromatin- mediated transcription repression, for example by exposing
DNA binding sites for additional activators or by exposing core promoter elements that
might be critical for the binding of general transcription factors and the formation of the

pre-initiation complex.

The handoff mechanism

Some evidence suggests that the interaction of an activator with TBP may involve more
than simply recruiting TBP to a promoter. The acidic activator Gal4 binds TBP
competitively with TATA DNA (180), and TBP and Gal4 do not bind cooperatively to
promoters (181). Additionally, a mutation (L114K) in the DNA binding region of TBP
interferes with activator binding (72), indicating that activators and DNA can bind to
overlapping regions of TBP. Taken together, these observations suggest that competition

for the DNA binding domain of TBP may be involved in some activation mechanisms.

The TFIID complex associates with TATA box DNA more slowly than does isolated
TBP, implying that the TAFs contain inhibitory functions (78). Some activators,
including VP16 and the Zta protein of Epstein- Barr virus, can stimulate the assembly of a
TFIID-TFIIA-DNA complex (76, 93), but do not stimulate a ternary complex when TBP
is used instead of TFIID, suggesting that activators may counteract inhibitory functions

of the TAFs. This ability of activators to stimulate ternary complex assembly appears to

15

be relevant for activation, since mutations in an activation domain that reduce
transcriptional activation potential in vivo also diminish the in vitro TFIID-TFIIA

assembly function (77).

In yeast, two of the TAF inhibitory activities have been mapped to two N-terminal
domains in the largest TAF subunit (84). These TAF N-terminal domains, called
TANDl and TAND2, bind to TBP and competitively inhibit the interactions of TBP with
DNA and TFIIA, respectively (79, 84). Additionally, the C-terminal subdomain of
VP16AD (VP16C) can compete with TANDl for binding to TBP (124). These
observations have led to a handoff model for activation (83) (figure 10, chapter 4), in
which VP16C competes with TANDl, in turn loosening the TAND2-TBP connection,
which then allows TFIIA to associate with TBP. Then, in a step that is still not
understood, TBP is handed-off from VP16C to DNA, leading to TFIID-TFIIA-DNA
complex assembly. The handoff step may involve conformational changes in TFIID or

the formation of additional contacts between TAFs and core promoter elements.

Interestingly, the handoff mechanism presents a situation where a finely-tuned affinity
between an activator and a target may be beneficial. If an activator interacts too weakly
with TBP, it will compete poorly with the repressive TANDl, but too much affinity for
TBP may interfere with DNA binding. Indeed, this may be the case: Drosophila TANDl ,
which exceeds both VP16C and yTANDl in affinity for TBP, is nonetheless a weaker

activator when fused to a DNA binding domain (83). Thus, the correlation of in vitro

16

affinity and in vivo function may be broken in mechanisms requiring transient

interactions between activators and targets.

Modulation of TBP orientation

In another variation on the theme of recruitment, an activator might function by altering
the orientation of the PIC. TBP can bind specifically to the TATA sequence present on
some promoters, and thus the TBP-DNA complex is thought to serve as a platform for
construction of the PIC. The conserved core region of TBP possesses an approximate 2-
fold symmetry, as does the TATA DNA element. However, the TBP surfaces contacting
other basal transcription factors (TF 11A, TFIIB, and TAFs) are unique, implying that the
orientation of TBP on a promoter defines the orientation of the overall PIC. TBP has
been reported to bind the TATA element in vitro with little or no orientational preference,
although interactions with other basal transcription factors appear to enforce a productive

polarity on TBP orientation (26, 68).

Both VP16AD and a synthetic amphipathic helix have been reported to significantly
increase the orientational specificity of TBP on the TATA DNA element in vitro (67).
This represents a variation on recruitment which might prime the TBP-DNA complex for
productive association with additional general transcription factors, and which might

work against the formation of incorrectly oriented PICs.

l7

Leaving a mark: activator effects after initiation

Although most studies of transcriptional activation have focused on recruitment and
initiation, subsequent steps including promoter escape and elongation can also be
stimulated by activator proteins. Several lines of evidence indicate that activators might
also work in post- initiation steps. One well-characterized example corresponds to the
human and Drosophila gene encoding heat shock protein 70 (hsp70). The uninduced
hsp70 gene contains a paused polymerase near the 5’ end of the gene. In response to heat
shock, not only does the transcriptional initiation rate increase, but the pausing time is
dramatically reduced (13, 145). Transcriptional activators can also stimulate rates of
transcriptional elongation. For example, the heat shock factor-1 (HSF- 1) involved in the
activation of hsp70 gene and the viral activators VP16 and ElA can stimulate elongation
by mechanisms that apparently differ from that of stimulation of initiation (14, 183). The
interaction of activators including HIV tat, c-myc, and NF -kB with the elongation factor
P-TEFb (a kinase that modifies the carboxyl-terminal tail of the largest subunit of RNA

pol II) fithher illustrates this link (5, 22, 65, 66).

Transcription and RNA processing have often been considered as separate and sequential
events, but a more recent perspective views these as a single integrated pathway
(reviewed by (133)). Capping, splicing and polyadenylation are tightly coupled to RNA
pol II through the carboxyl-terminal domain of the largest subunit (reviewed by (104)).

Selection of splice sites in the nascent RNA can be influenced by promoter elements in a

18

manner that is independent of the promoter strength, suggesting that activators might also

regulate alternative splicing decisions (2, 27, 125).

Structure of transcriptional activators

The dual functions of a transcriptional activator protein, cis-element recognition and
transcriptional activation, are typically fulfilled by distinct regions of the protein’s
primary structure. For example, the DNA binding domain of the yeast Gal4 protein (a
prominent activator in the yeast Saccharomyces cerevisiae) resides within the amino-
terminal 100 amino acids, whereas the major transcriptional activation domain resides
within the carboxyl-terminal 120 amino acids. This modular design seems advantageous
both for evolutionary and technological appropriation. In the latter sense, a fusion
protein linking the DNA-binding domain of Gal4 with VP16AD (146) is widely used in
both in vitro and in vivo investigations into the mechanisms of transcriptional activation.
A second example, comprising a fusion of the DNA binding domain of the tetracycline
repressor with VP16AD, allows the regulation of DNA binding by the presence or
absence of the tetracycline ligand (45). This regulatable artificial activator can function

in a wide range of eukaryotes ranging from plants to mammals (44, 46, 173).

19

DNA binding domains

The DNA binding domains of a large number of eukaryotic transcriptional activator
proteins have been extensively characterized by genetic, biochemical, and structural
approaches. Recent reviews catalog the known structures and specificities (101) and
highlight the common themes in structure and recognition (37). In many cases, the
binding activity or specificity of a DNA binding domain may be modulated by ligand
binding, dimerization with other DNA binding proteins, or by association with other

factors (reviewed by (105)).

The principles of proteinzDNA interaction have now been established to a sufficient
degree to permit the design of DNA binding modules of engineered specificity (6, 34).
This is particularly true for the zinc-finger families of transcription factors (24, 64, 151,
177). This ability to tailor novel chimeric transcriptional activators for recognition of
DNA sequences that might not serve as native regulatory elements has profound

implications for potential pharmacological application (95, 100, 184).

20

Activation domains

Features in primary structures

In contrast to DNA binding domains and despite substantial research, relatively little is
known about the structures of transcriptional activation domains. By analysis of primary
sequence, these domains have been broadly classified based on the abundance of
particular amino acids, resulting in acidic, glutamine rich, proline rich, and other classes
(113, 163). Despite these amino acid preferences, however, careful mutational analyses
have indicated that the most critical elements of activation domains are frequently the
patterns of hydrophobic and aromatic amino acids (1, 19, 28, 58, 163). For example, in
the acidic VP16AD, mutation of one or several acidic residues has only a modest effect
on activity, whereas mutation of a single key hydrophobic residue can severely reduce

activity (28, 139, 160).

Secondary and Tertiary Structures

Less is known about the secondary and tertiary structures of activation domains. A
number of biophysical analyses of various regulatory proteins have shown that activation
domains are largely unstructured in solution under physiological conditions (29, 126, 150,
152). However, key amino acids in an activation domain can become conformationally

constrained upon interaction with a target protein, suggesting that the most promising

21

targets for structural studies will be binary complexes between activators and targets. For
example, circular dichroism spectra indicate that the c-Myc transactivation domain is
induced to form a helical structure upon binding to TBP (109). The activation domains
of VP16 and of the estrogen receptor also become conformationally constrained upon
interaction with TBP (153, 171). Furthermore, the VP16AD appears to become helical
upon binding human TAF 32 (166). An amphipathic helix from the p53 activation
domain fills a hydrophobic clefi in the MDM2 oncoprotein (88). An amphipathic helix
structure is also seen in the interface between the activator CREB and its coactivator
protein CBP (137). These examples support the model that an activator target provides a
folding template for an unstructured activation domain, which might allow a particular

activation domain to interact with a number of different target proteins.

In some cases the tables may be turned: that is, an activation domain may provide the
folding template for a potential target. Nuclear hormone receptors have a conserved C-
terrninal activation domain, known as AF -2. Ligand binding leads to a conformational
change in the activator that opens a hydrophobic groove for interaction with
transcriptional co-activators. In this case, the unstructured LxxLL peptide motif present
in several coactivators (53) folds into an amphipathic helix upon binding the AF -2 region

of a hormone receptor (reviewed in (170)).

Although in many cases activation domains seem to adopt a-helical structures, that rule

is not universal. Mutational and biophysical analysis of the Gal4 activation domain

suggested that it might form a B-strand instead (168). The activation domain of E2F-2

22

upon interaction with the Rb tumor suppressor protein assumes a combination of helical
and B-strand conformations (90). Together, these observations suggest that activators
and their target proteins bind to each other using a highly diverse set of interaction

surfaces.

Central Hypothesis

Many potential targets of transcriptional activation domains have been identified, and the
detailed interaction of activators and targets may involve a range of mechanisms.
However, we do not know which targets and mechanisms are relevant for transcriptional
activation in vivo. The strong interaction of VP16AD with TBP and the correlation of
this interaction with VP16AD function lead to the hypothesis that TBP is a relevant in
vivo target for at least one class of activation domains. Additionally, several lines of
evidence suggest that VP16AD and DNA compete for binding to TBP, which argues that
VP16AD cannot stably recruit TBP to a promoter. This leads to the hypothesis that the
VP16AD:TBP interaction in transcriptional activation involves a variation of direct and
stable recruitment. Variations on recruitment include the modulation of TBP orientation
on DNA, discussed in chapter 3, and the transient interactions involved in a handoff

mechanism, discussed in chapter 4.

23

Chapter II

Purification of yeast TATA binding protein

Introduction

The TATA binding protein (TBP) plays a central a central role in all classes of eukaryotic
transcription (25). TBP comprises a highly conserved C-terminal DNA binding domain
and an N-terminal domain that varies considerably among species. TBP specifically
binds to the TATA DNA sequence element (156) found about 30 bases upstream from
the transcription start site in many promoters. Since TBP can specifically bind to the
TATA elements found in some promoters, it is thought to nucleate the assembly of the
large protein complexes that poise RNA polymerases at the transcriptional start sites of

genes (50).

TBP has been purified by a variety of methods, including bulk fractionation (23),
conventional chromatography (9, 20, 23), and fusion to affinity tags. Most large-scale
TBP purifications use a combination of these methods; however, each of these
purification methods has limitations. Conventional chromatography typically relies on
low salt binding conditions, but TBP readily aggregates under low salt conditions
(personal observations). Heparin affinity chromatography provides an alternative to ion
exchange methods, and since TBP can be bound to heparin columns at 200 mM salt (9,
131), low salt conditions can be avoided. However, heparin affinity chromatography

alone is insufficient to purify TBP to homogeneity (131). N—terminal polyhistidine tags

24

and nickel affinity chromatography have frequently been employed (62, 85, 121, 165),
although subsequent dialysis is typically required to remove the elution agent (imidazole
or EDTA), and removal of the extrinsic polyhistidine tag requires proteolysis and the

subsequent removal of protease activity.

In our efforts to stabilize a complex of Saccharomyces cerevisiae TBP bound to the
transcriptional activator VP16C (120, 139, 146), we explored a fusion of VP16C to
another TBP- interacting peptide, TAND2 (3, 79, 102). This VP16C-TAND2 fusion,
based on a precedent by Kotani et al. (83), specifically binds TBP with high affinity in a
salt-dependent manner. We have exploited this interaction as a simple, fast, and effective

means of purifying untagged yeast TBP.

Materials and Methods

Plasmid construction. To generate an expression plasmid encoding a fusion of
glutathione-S-transferase-VP16C-TAND2 (GST-VP16C-TAND2), a DNA fragment
encoding TAND2 (TAF 1 residues 42 to 78) was PCR amplified from yeast genomic
DNA using primers ST805 and ST806. The resulting DNA fragment was tailored by
EcoRI and XhoI and directionally cloned between the EcoRI and XhoI sites of plasmid
pGEX4T-2 (GE Healthcare), yielding plasmid pDS42-5. VP16C (VP16 residues 451-
490) was PCR amplified from plasmid pGEX.VPl6C (119) with primers ST803 and

ST804. The resulting DNA fragment was tailored by BamHI and EcoRI and directionally

25

cloned between the BamHI and EcoRI sites of pDS42-5 to yield pDS42-10. Plasmid
fidelity was verified by sequencing. Plasmid pKA9, encoding Saccharomyces cerevisiae

TBP under control of the T7 promoter, was obtained from Dr. Karen Arndt.

Expression of GS T - VP16C-TANDZ. Plasmid pDS42-10 was transformed into E. coli
strain BL21 (Novagen) and cells containing plasmid were grown in LB medium
containing 80 ug/mL ampicillin. One liter cultures were grown at 37 °C with vigorous
shaking until an OD6oo of about 0.7. Expression of GST-VP16C-TAND2 was induced by
addition of IPTG to a final concentration of 0.5 mM and growth was continued for an
additional 3 hours while the culture temperature was allowed to approach ambient
temperature (25 °C). Cells were harvested by centrifugation and resuspended in l/60
culture volume HEMGT-250 (24 mM HEPES, 0.1 mM EDTA, 12.5 mM MgC12, 10%
glycerol, 0.1% Tween-20, 250 mM KCl, pH 7.9) supplemented with 1 mM PMSF and 5
mM DTT. Cells were lysed by two passages through a French Press (Aminco) operating
at 20,000 psi. Cellular lysates were centrifuged at 10,000 x g at 4 °C for 20 min. The
supernatant was recovered, split into 1 mL aliquots, flash frozen in liquid nitrogen, and

stored at -80 °C.

Expression of yeast T BP. E. coli strain BL21(DE3) Codon Plus (Novagen) was
transformed with plasmid pKA9 which encodes yeast TBP under control of the T7
promoter. Cells containing expression plasmids were grown in TB medium (Terrific
Broth) with 80 mg/L ampicillin and 40 mg/L chloramphenicol. One liter cultures were

grown at 37 °C with vigorous shaking until an OD600 of 1.5. Expression of TBP was

26

induced by the addition of IPTG to a final concentration of 0.1 mM and growth was
continued for 3 h while the culture temperature was allowed to approach ambient
temperature (25 °C). Cells were harvested by centrifiigation and resuspended in 1/30
culture volume of HEGK-400 (25 mM HEPES, 1 mM EDTA, 10% glycerol, 400 mM
KCl, pH 7.5) supplemented with 5 mM DTT. Cells were lysed by sonication (Branson
Sonifier 450, 80% duty cycle, power setting 8, 4 cycles of 30 sec with 2 min rest between
cycles). Cellular lysates were centrifuged at 9000 x g at 4 °C for 25 min. The
supernatant was recovered, split into 1 mL aliquots, flash frozen in liquid nitrogen, and

stored at -80 °C.

Quantitation of GS T - VP] 6C—T ANDZ recovery. Bacterial lysates (10 uL to 100 uL )
were added to 200 uL of a 10% slurry of glutathione (GSH) resin in HEGK-400.

Binding reactions were axially rotated for 45 min at 7 °C. Binding reactions were
centrifuged at 500 x g for l min and the supernatant was removed by aspiration. To wash

the resin, 1.5 mL ice-cold HEGK- 100 was added with enough vigor to ensure resin
resuspension and the resulting diluted slurry was promptly centrifuged at 500 x g for l
min. After a total of 5 washes, 150 uL of elution buffer (0.9 x HEGK-IOO, 0.1 x 100 mM
reduced GSH pH 7.9) was added, and the resin was axially rotated at 25 °C for 30 min.
The resin was centrifuged again and the protein concentration in the supernatant was

measured with a Bradford assay relative to BSA standards.

Purification of TBP. All steps except for elution were performed on ice or in a 7 °C

environment. 400 uL of GVT-containing lysate was mixed with 10 volumes (4 mL) of a

27

TBP-containing lysate and the mixture was incubated on ice for one hour. The lysate
mixture was added over 300 uL GSH resin (equilibrated in HEGK- 100), and the slurry
was gently rotated for 45 min at 7 °C. After resin binding, five washes were performed
as follows: 75 resin volumes of HEGK- 100 were added to the resin, resuspension was
ensured, the slurry was promptly centrifuged at 500 x g for 3 min, and the supernatant
was removed by aspiration. The washed resin was transferred to a polypropylene
chromatography column and the supernatant was drained through the outlet. 2 mL of
HEGK-IOOO buffer at room temperature (25 °C) was carefully layered onto the resin, and
the colurrm was gravity eluted at a flow rate of approximately 1 mL/min. 1 mL fractions
were collected into chilled tubes. TBP typically eluted in the first fraction. Typical

recovery was approximately 1 mg.

Results and Discussion

In our efforts to stabilize a complex of TBP and VP16C, we noted that a previously
developed fusion of VP16C and TAND2 bound tightly to TBP (83). Although the region
of T BP bound by VP16C has not been precisely determined, DNA competition
experiments and TBP mutations have localized the interaction to the concave DNA
binding region of TBP (72, 83, 94). TAND2, an N-terminal domain from the largest TBP
associated factor TAF l (162), interacts with a different region of TBP, on the upper
convex surface (102) and in competition with TFIIA (79). Since the previously

constructed VP16C-TAND2 was shown to bind TBP with high affinity and also appears

28

to preserve the VP16C:TBP interface, we constructed and explored a similar fusion

protein comprising GST, VP16C (VP16 451-490), and TAND2 (TAF 1 42-78).

To form a complex of GST-VP16C-TAND2 (GVT) and TBP, a mixed- lysate approach
was tested. GVT and TBP were expressed in separate E. coli cultures and the soluble
fractions of the cellular lysates were stored at -80 °C. In a first step of lysate
characterization, an amount of GVT-containing lysate sufficient to bind 100 ug protein to
20 uL GSH resin was determined. In a second step, increasing volumes of TBP-
containing lysate were titrated over the constant volume of GVT lysate, the lysate
mixtures were incubated on ice for one hour, and then purified over GSH resin. GVT and

any associated proteins were eluted from the resin by boiling in SDS-PAGE loading
buffer and then resolved by SDS-PAGE (Fig. 1A). This analysis indicated that GVT

could retain TBP though extensive washing of the GSH resin.

In order to quantify the retention of TBP by GVT, high salt conditions were used to
selectively elute TBP from the immobilized complex. Lithium sulfate was found to be
particularly effective at disrupting the TBPzGVT interaction, and was thus used for
analytical-scale elutions. To quantitate purified TBP, 50 uL of 1 M lithium sulfate was
layered onto 20 uL of resin with bound protein complexes, the slurry was rotated axially
for 20 min at room temperature, and 40 uL of the supernatant was assayed in a 1 mL

Bradford assay relative to BSA standards. This method produced a binding curve that

29

>
w

 

 

 

 

 

 

 

’53
3
E‘ 30
9
=————————— GVT 0
§ 20
0.1
-~——-'—TBP m
“‘ 10
0 200 400 600 800 1000
TBP lysate (111)
s E 9::
‘5 0 ”>1
c a «g s
q . 0 'U
'T J- 8 t— 5 8
8 8 2 > o .2
kDa 9.. 9.. U (D m as
..¢ #9
97— g
45‘: ‘9 -— —GVT
31‘W”! - «.- -—TBP
21 .- T
14 g "

 

 

 

Figure 1: Purification of full-length yeast TBP. (A) Titration of increasing amounts of
TBP-containing lysate over a constant amount of GVT-containing lysate. Lysates were
mixed, incubated on ice for one hour, and purified over GSH resin. Bound proteins were
eluted by boiling in SDS loading buffer, resolved by SDS PAGE, and stained with C00-
massie R-250. (B) Quantitation of purified TBP. TBP was released from immobilized
GVTzTBP complexes by addition of 1 molar lithium sulfate. Eluted proteins were quanti-
tated by Bradford assay. (C) Purification of isolated TBP monitored by SDS-PAGE.
Lanes 1-2, TBP culture before and after induction of TBP expression. Lane 3, lysate
containing TBP. Lane 4, lysate containing GST-VP16C-TAND2. Lane 6, proteins
retained on GSH resin after washing. Lane 8, TBP released by addition of buffer contain-
ing 1 molar potassium chloride.

30

closely resembled the SDS-PAGE data (Fig. 18). Notably, the recovery profile of TBP
indicated that with low amounts of TBP lysate, the recovery was linear with increasing
TBP, while with higher amounts of TBP lysate the recovery approached saturation and
then diminished. This indicated that lysate ratio can be tuned for different purposes. To
make optimal use of GSH resin, saturating amounts of TBP lysate should be used.

Conversely, to make optimal use of TBP lysate, lower amounts should be used.

To optimize the binding reaction, the salt concentration in the lysate mixture was
explored. Since TBP readily aggregates under low salt conditions and the GVTzTBP
interaction is weakened at high salt concentrations, a moderate salt concentration was
expected to be optimal. The optimal salt concentration was determined by performing a
parallel series of analytical scale purifications where the salt concentration in the lysate
mixture was the only variable. The results from this experiment indicated that the
recovery of N-terminally truncated TBP derivatives was best when the lysate mixture
contained 400 mM KCl, which yielded nearly twice as much TBP as a lysate mixture
containing 100 mM KCl. The salt sensitivity of full- length TBP was less pronounced,
but also showed that recovery was optimal from a lysate mixture containing 400 mM KCl.
Based on these results, and to simplify subsequent purification, the TBP lysis buffer has

been defined to contain 400 mM KC1.

Since the lithium sulfate that was used in the analytical-scale elutions may interfere with
downstream applications, potassium chloride was tested as an elution buffer for

preparative scale purifications of TBP. To perform this test, 180 uL of elution buffers

31

with various concentrations of KCl were added to 20 uL beds of protein loaded resin.
The samples were rotated axially for 20 min at different temperatures, and the
supernatant was assayed for eluted protein. From this analysis, it was determined that l
M KCl was sufficient to elute TBP at room temperature, while at least 1.25 M KCl was

required for elution at 7 °C.

To scale up the purification method for preparative use, a column-based elution format
was employed. Lysate mixtures were scaled up to supply protein for 300 pL of GSH
resin. The resin was washed as in the small scale purifications, and then transferred to a
plastic chromatography column. The resin was settled and the supernatant was drained,
and 2 mL of (25 °C ) HEGK- 1000 was carefully layered onto the resin. The column was
gravity-eluted at approximately 1 mL/min flow, and fractions were collected into tubes
chilled in ice. TBP eluted primarily in the first fraction, and approximately 1 mg of TBP

was recovered (Fig. 1C).

The TBP purified by this method has been successfully used in several experiments.
Purified full- length TBP has been found to specifically recognize the TATA DNA
sequence (Chapter 3), demonstrating proper function. A purification of the conserved
180 amino acid core region (residues 61-240) was crystallized (Chapter 4), indicating that
the protein was properly structured. A preparation of TBP (residues 49-240) has also
been used as a standard for gel filtration experiments (Chapter 5), where it eluted with an
apparent molecular mass of about 34 kDa, in rough agreement with the expected dimeric

state. Additionally, the gel filtration analysis showed no material eluting with a higher

32

apparent molecular weight, indicating that the preparation was not detectably denatured

or aggregated.

This method of TBP purification has several distinct advantages over current methods.
Foremost, pure TBP can be obtained in under three hours and without the need for
chromatographic equipment. Additionally, the resulting material is monodisperse,
structured, and functional. Importantly, TBP can be purified without the need for
extrinsic affinity tags or any subsequent proteolysis or cleanup. The method can also be
performed at any scale, providing for rapid purification based on need. Finally, since
typically over 100 mg of GVT can be purified from a single liter of E. coli culture, a
single large preparation of GVT—containing lysate (separated into small aliquots and

characterized once) can provide for many TBP purifications.

33

Chapter III

The orientation of TBP on the TATA DNA sequence

Introduction

The core promoter region of a protein coding gene localizes the assembly of a large
multiprotein complex comprising RNA polymerase II (polII) and several general
transcription factors (GTFs). The resulting pre- initiation complex (PIC) poises polII in
the proper location and direction for subsequent transcription of the downstream DNA.
In vitro, the PIC can be assembled onto DNA by the ordered addition of GTFs,
suggesting a stepwise PIC assembly pathway (16). However, a large multiprotein
complex comprising pollI and several additional factors can be purified from cells,

suggesting that PIC formation may instead involve the recruitment of a larger polII

“holoenzyme” (91 ).

The TATA binding protein (TBP), either alone or with TBP-associated factors (TAFs) in
the TFIID complex (162), can bind to the TATA element present in many promoters.
TBP is thus thought to mediate the initial protein-DNA contacts during PIC formation.
The highly conserved DNA-binding region of TBP comprises two imperfect direct
sequence repeats, and the tertiary structure possesses a high degree of pseudodyad
symmetry (123). However, the contacts between TBP and the other basal factors are not
symmetrical. TFIIA binds specifically to a region in the N-terminal repeat in TBP (39,

161), while TFIIB interacts with residues in the C-terminal repeat (122). The asymmetric

34

connections between TBP and other basal factors suggest that the orientation of TBP
bound to a promoter dictates the orientation of the overall PIC and thus the location and

direction of transcription.

TBP specifically recognizes the TATA element located about 30 bases upstream from the
transcriptional start site in many promoters. TBP binds in the minor groove of DNA at
the TATA sequence, inducing a large distortion in the DNA structure (70, 73). Notably,
the minor groove of the TATA sequence provides few structural features to guide TBP
toward a particular orientation. Nonetheless, the TATA sequence is found in a preferred
orientation in promoters (15, 156), suggesting both that the TATA sequence can orient
TBP, and that the TATA-directed orientation of TBP at a promoter serves an important

role in directing transcription.

The intriguing symmetry of TBP and the TATA element have led to several studies that
attempt to decipher their role in determining transcriptional directionality. All crystal
structures of TBP bound to TATA DNA sequences show oriented binding, suggesting
that TBP may bind to the TATA element in only one orientation. However, crystal
lattice constraints may influence this outcome, possibly by trapping a slightly favored
orientation. Efforts to determine the orientation of isolated TBP on DNA in solution by
nuclease protection analysis (footprinting assays) are hampered by the symmetry of TBP.
However, the larger TFIID complex extends the TBPzDNA complex asymmetrically and
can thus be employed in DNA footprinting experiments to determine directionality of the

TFIID:DNA complex.

35

F ootprinting experiments with TF 11D bound to an isolated TATA sequence revealed
asymmetrical protection patterns directed by the orientation of the TATA sequence (16,
117, 167). Additionally, in vitro transcription reactions have indicated that an isolated
TATA sequence can direct oriented (though weak) transcription (127). Although this
evidence suggests that the orientation of the TATA sequence dictates the orientation of
the PIC and transcription, other evidence indicates that this is not a strict rule. For
example, an upstream activator can dominantly direct downstream transcription,

regardless of the orientation of an intervening TATA element (127).

While these studies indicate that the TATA sequence orientation can orient the PIC and
transcription, and also that this TATA-directed orientation can be overridden by an
upstream activator, the results come from complicated contexts. The TFIID footprinting
assays rely on TAFs, and the transcription reactions involve many additional factors.
These additional factors provide additional protein- DNA contacts that may modulate the
orientational specificity of TBP, and additional protein-protein contacts that may mediate

connections to other factors.

To more directly assess the intrinsic directionality of TBP binding to the TATA sequence,
Schepartz and colleagues developed an in vitro system to measure the orientation of TBP
in the absence of other factors (26). By attaching the copper chelator 5-Iodoacetamido-
1,10-phenanthroline (IAAOP) (129, 155) directly to TBP, they converted TBP into a site-

specific nuclease. The IAAOP moiety was positioned such that the orientation of TBP

36

directed the nuclease activity to positions either upstream or downstream of the TATA
sequence. Utilizing this affinity-cleavage methodology, it was determined that isolated
TBP binds to the Adenovirus major late promoter (AdMLP) TATA sequence with only a
modest degree of orientational preference. Inclusion of either TFIIA or TFIIB increased
the orientational specificity (26), and inclusion of both TF 11A and TFHB led to virtually
unidirectional TBP binding (68). These results suggested a model of PIC assembly in
which isolated TBP is unable to orient itself at a promoter, and thus the orientation of the

PIC and the direction of transcription are dependent on other factors.

Subsequent experiments using similar affinity-cleavage assays indicated that the
prototypical acidic activator Gal4-VP16 (146) can enhance the orientational specificity of
TBP bound to the TATA element (67). This result suggested a novel mechanism of
transcriptional activation, in which an activator functions by promoting forward-oriented

PIC assembly, or possibly by working against reverse-oriented PIC assembly.

Intriguingly, the TBP-orienting mechanism implies a ternary interaction between
VP16AD, TBP and DNA, although several other lines of evidence suggest that acidic
activation domains and DNA compete for TBP binding, precluding such a ternary
interaction. For example, the activation domains of VP16 and Gal4, when free in
solution, disrupt the interaction of TBP and DNA (94, 180, 181). A mutant in the
concave DNA binding region of TBP (L114K) abrogates TBP interactions with both
VP16AD and the Gal4 activation domain (72), suggesting that activators and DNA bind

to overlapping regions of TBP. Additionally, VP16AD has been shown to compete with

37

a TAF 1 inhibitory domain (TANDl) for binding to the DNA binding region of TBP
(124). Since VP16AD and DNA apparently compete for binding to TBP, it is unclear

how VP16AD might modulate the orientation of the TBP-DNA complex.

To firrther explore the proposed TBP-orienting activity of VP16AD, we have replicated
and extended the affinity cleavage experiments performed in the Schepartz lab.
Unexpectedly, and contrary to the results from the Schepartz lab, we found that TBP
binds to the TATA DNA element with a high degree of orientational specificity. We
have also found that Gal4- VP16AD does not alter the intrinsic orientational specificity of
TBP. In addition to our studies of wild-type TBP, we examined the orientational
specificity of a TBP mutant that can direct forward transcription from a reversed TATA
sequence in vivo (Dr. K. Amdt, pers. com.). This mutant TBP has a slightly relaxed
orientational specificity, which may explain the in vivo phenotype. Taken together, our
affinity cleavage results indicate that the current model of TBP orientation needs to be

revised.

38

Materials and Methods

Cleavage probe sequences. Plasmid pGSMLT, encoding five Gal4 binding sites fused to
bases -50 to +10 of the Adenovirus major late promoter (AdMLP) and a 390 bp G- free
sequence, was obtained from Dr. M. Carey, UCLA. The sequence of the TATA element
in pGSMLT was altered by site-directed mutagenesis (QuikChange method, Stratagene).
Oligonucleotides ST1130 and ST1131 were used to create plasmid pDSB51-1, encoding
the symmetric TATA sequence (TATATATA). Oligonucleotides ST1132 and ST1133
were used to create plasmid pDSBS 1-5 encoding the reverse TATA sequence

(C‘1“l'[ 'l'ATA). Plasmid sequences were verified by sequencing.

Cleavage probe generation. F luorophore- labeled Oligonucleotides ST903-6F AM and
ST904-HEX (synthesized by Integrated DNA technologies, www.idtdna.com) were used
to prime PCR amplification of a 304 bp DNA segment from plasmid pGSMLT or
derivative plasmids. 6F AM denotes 6-carboxyfluorescein attached to the 5’ position of
the oligonucleotide, and HEX denotes hexachlorofluorescein attached to the 5’ position
of the oligonucleotide. The resulting DNA segment, encompassing five Gal4 binding
sites, an AdMLP TATA sequence, and a portion of the G- free sequence, was spin-column

purified (Qiagen) and stored at —— 80 °C.

T BP expression. Plasmids encoding full length Saccharomyces cerevisiae TBP and

mutant derivatives were obtained from Dr. Karen Arndt (pKA9: wild-type; pJV SS6:

K97C, F2271; pJVSS7: E188C, F2271; pJVSSB: K97C; pJVSS9: E188C). TBP

39

expression plasmids were transformed into E. coli strain BL21(DE3)codon-plus. Cells
containing expression plasmids were grown in LB medium containing 80 mg/L
ampicillin and 40 mg/L chloramphenicol. 5 mL cultures were inoculated into 1 L of TB
medium (Terrific Broth) containing 80 mg/L ampicillin, and cultures were grown at 37
°C with vigorous shaking to an ODfiOO of about 1.5. Expression of TBP was induced by
the addition of IPTG to a final concentration of 0.1 mM, the ambient temperature was
shifted to 25 °C, and growth was continued for 3 additional hours. Cells were harvested
by centrifugation and resuspended in 30 mL HEGK-400 (24 mM HEPES, 1 mM EDTA,
10% v/v glycerol, pH 7.5) with 400 mM KCl and 5 mM DTT. Resuspended cells were
lysed by sonication (Branson Sonifier 450, 80% duty cycle, power level 8, 4 cycles at 30
sec with 3 min rest between cycles). Cellular lysates were centrifirged at 9,000 x g at 4
°C for 25 min. The lysate supernatant was collected, flash frozen in liquid nitrogen, and

stored at -80 °C.

T BP purification. TBP was purified using an intrinsic affinity method described in
chapter 2. Briefly, bacterial lysates containing TBP and GST-VP16C-TAND2 were
mixed and incubated on ice for one hour. Lysate mixtures were added over glutathione
resin and the resulting slurry was gently rotated at 7 °C for 45 min. The resin was
collected by centrifugation at 500 x g and the lysate mixture was removed. Five washes
were performed as follows: 75 bed volumes of ice-cold HEGK—IOO were added, resin
resuspension was ensured, and the dilute slurry was immediately centrifuged at 500 x g.
After the final wash, the resin was transferred to a fritted column, 2 mL of HEGK- 1000

(25 °C) was layered onto the resin bed, and column elution fractions were collected.

40

Eluted material was flash frozen in liquid nitrogen and stored at -80 °C. The
concentration of eluted TBP was measured by the Bradford assay relative to BSA

standards.

IAAOP preparation. A 3 mg sample of IAAOP was obtained from Dr. A. Schepartz
(Yale). IAAOP is light sensitive, and thus was handled under minimal light conditions.
The IAAOP sample was suspended in 88 uL dimethylformamide (DMF) and stored at -
80 °C as a master stock. Working stocks (10 mM IAAOP in DMF) were diluted from the

master stock.

T BP derivitization. TBP derivatization was performed under low light conditions. 230
pg (8.5 nmol) of TBP was diluted into a final volume of 600 uL in HEGK- 1000 (pH 8.2).
Ten equivalents of IAAOP (85 nmol) in DMF was added and the solution was gently
mixed. Derivatization reactions continued at 25 °C for one hour in the absence of light.
To remove unreacted IAAOP, the derivatization reactions were buffer- swapped by
dilution into 20 volumes of HGK-400 (24 mM HEPES, 10% v/v glycerol, 400 mM KCl,
pH 7.9) and concentration with a centrifugal concentration device (Millipore, 10 kDa
molecular weight cut-oft). After two cycles of buffer swap, the recovered material was
split into single-use aliquots, flash frozen in liquid nitrogen, and stored at -80 °C. The
concentration of the resulting TBP-OP was estimated by a Bradford assay relative to

BSA standards.

Gal4- VP16AD. Gal4-VP16 was purified as described (119).

41

 

Cleavage reactions. Cleavage reactions were carried out as previously described (67)
with modifications. 100 ng of fluorophore- labeled probe DNA was incubated with 50 ng
TBP-OP, 1 ug poly (dG-dC)°poly(dG-dC), and 12 ug BSA in 300 rd. binding buffer (4
mM Tris-Cl, 60 mM KCl, 5 mM MgC12, 4% glycerol, 0.1% NP-40, pH 8) for 35 min at
25 °C while protected from light. To initiate DNA cleavage, 60 1.1L of binding buffer
containing 30 mM 3-mercaptopropionic acid, 0.06% w/v H202, and 300 uM CuSO4 was

added to the TBP:DNA binding reactions. DNA cleavage was continued for 3 h at 25 °C

while protected from light. Cleavage reactions were terminated by ethanol precipitation.

F ootprinting reactions. A solution comprising 67 uM Methidiumpropyl-EDTA, 133 uM
F e(NH4)2(SO4)2, and 17 mM DTT was prepared immediately before use. 90 uL of this
solution was added to 300 uL of the TBP:DNA binding reactions. Cleavage was

continued for 5 min at 30 °C and. subsequently terminated by ethanol precipitation.

DNA fragment analysis. DNA cleavage and footprinting reactions were ethanol
precipitated, washed twice with 70% ethanol, and dissolved in 18-20 uL formamide
mixed with in- lane size standards (Applied Biosystems ROX-500 mixture or equivalent).
Samples were analyzed by an Applied Biosystems 3100 genetic analyzer (30 cm capillary,

60 sec injection time) at the Research Technology Support Facility, MSU.

Quantitation of DNA fragment data. Raw chromatographic data was extracted from the

ABI Prism data files with the program Batchextract.exe (fip.ncbi.nih.gov/pub/

42

forensicsfbatchextracU). Orientation measures were performed by the program BSS-3.C
as follows. In-capillary size marker positions were used to define the center of the
cleavage pattern. From this center position, raw fluorescence intensity data were
summed in regions spanning 13 bases upstream or downstream from the center position.
A signal baseline was approximated with a straight line connecting minima found in 17
base regions flanking the quantitated regions. Baseline areas were subtracted from the
raw data sums. TBP orientation was measured from the baseline-corrected fluorescence
data as the summed signal in one quantitation region divided by the sum from both

regions.

43

Results

To explore the TBP orienting activity of the VP16 activation domain we have replicated
and extended the affinity cleavage experiments performed by the Schepartz laboratory
(26, 67). In these experiments, the DNA cleavage agent IAAOP was attached to an
engineered cysteine (K97C or E188C) located in one of the two stirrup loops of TBP,
resulting in affinity-cleavage proteins K97C-OP and E188C-OP. In the TBP:DNA
complex, the stirrup loops lie in the minor groove of DNA at positions flanking the 8 bp
TATA sequence. Thus, attachment of IAAOP to a stirrup loop position on TBP localizes
DNA cleavage activity to one side of the TBP:DNA complex. Subsequent analysis of the
DNA fragment distribution allows the determination of the orientation of TBP bound to

DNA (Fig. 2).

Our TBP orientation measurements utilized the same DNA cleavage methods as the
Schepartz group, but we have used a different method of DNA fragment analysis. The
Schepartz group used radioactively end-labeled DNA cleavage probes, resolved the
affinity-cleaved DNA fragments on polyacrylamide sequencing gels, and then quantified
gel images to determine the distribution of DNA cleavage events in the vicinity of the
TATA element. In contrast, we used fluorophore end-labeled DNA cleavage probes and
resolved affinity-cleaved DNA fragments using an Applied Biosystems 3100 genetic
analyzer. This instrument couples capillary electrophoresis with fluorescence detection,
allowing the simultaneous separation and detection of different uniquely labeled DNA

fragment populations. By attaching different fluorophores to the top and bottom strands

44

 

Lys 97 Glu 188

6-FAM
* TATAAAAG
ATATTI'TC :1:
HEX

 

 

 

 

 

 

 

{JilllJLi E

 

5 x Gal4 TATA

Figure 2: Affinity cleavage methodology. (A) Locations of K97 and E188 shown as
spheres mapped onto a ribbon representation of the conserved core of TBP. (B) Location
of fluorophores on each strand of the cleavage probe DNA. 6-FAM denotes 6-
carboxyfluorescein, and HEX denotes hexachlorofluorescein. (C) Schematic diagram of
304 basepair cleavage probe drawn to scale.

45

of the cleavage probe DNA, we were able to simultaneously determine the DNA cleavage
patterns on both probe DNA strands (Fig. ZB). Since each DNA separation was
performed in an individual capillary, the DNA fragment sizes had to be determined
relative to an array of in-capillary DNA size standards. To provide for precise fragment
sizing, we calibrated a set of generic size standards by comparison to cleavage probe
DNA that was digested by restriction enzymes. The restriction enzymes BamHI and
BstUl were used to cleave probe DNA at positions 28 bases upstream or 16 bases
downstream from the center of the TATA sequence, respectively, providing accurate size

standards flanking the region of interest.

To determine the intrinsic orientational preference of TBP binding to the TATA element,
K97C-OP and E188C—OP were used to cleave probe DNA encoding the AdMLP TATA
element. An analysis of the DNA fragment distribution showed specific cleavage
localized to the region of the TATA sequence. The affinity-cleavage patterns were
clearly asymmetrical, with more cutting on one side of the TATA sequence. E188C-OP
cleaved the top strand of the probe DNA preferentially on the upstream side of the
AdMLP TATA sequence (Fig. 3A), while K97C-OP cleaved the top strand preferentially
on the downstream side (Fig. 3B). These cleavage preferences match the orientation of
TBP that is observed in the crystals structures of all TBP:DNA complexes, where the N-
terminal repeat of the TBP core region faces downstream, toward the transcriptional start
site. The cleavage patterns on the bottom strand of the probe DNA were consistent with
the top strand patterns, again showing that E188C-OP preferentially cleaved upstream

from the TATA sequence while K97C-OP cleaved downstream (Figs. 4A, 4B).

46

 

 

A E188C-OP B K97C-OP
TATAAAAG TATAAAAG

 

 

 

 

 

 

 

 

 

C E1 88C-OP D K97C-OP
TATATATA

  

TATATATA

 

 

 

 

 

 

 

 

 

 

 

 

 

E E188C-OP F K97C-OP
CIIIIATA CIIIIATA

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Figure 3: Affinity cleavage patterns on top DNA strand. Representative top strand cleavage
data are shown in 60 basepair windows centered on the cleavage patterns. The horizontal
axis is DNA position, and the vertical axis is scaled fluorescence units. The sequence of the
TATA element is shown on each panel. The vertical bars indicate positions at the center and
13 bases to either side of the cleavage patterns.

47

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

1 fire-A

 

A E188C-OP B K97C-OP
TATAAAAG TATAAAAG
C El 88C-OP D K97C-OP
TATATATA TATATATA
N
E E1 88C-OP K97C-OP
C I I I IATA CTTTTATA

 

Figure 4: Affinity cleavage patterns on bottom DNA strand. Representative bottom strand
cleavage data are shown in 60 basepair windows centered on the cleavage patterns. The
horizontal axis is DNA position, and the vertical axis is scaled fluorescence units. The top
strand sequence of the TATA element is shown on each panel. The vertical bars indicate
positions at the center and 13 bases to either side of the cleavage patterns.

Importantly, the cleavage patterns determined from the top DNA strand and the bottom
DNA strand qualitatively agreed. Since the fluorophore label on each strand was located
at the 5’ position, the agreement between top strand and bottom strand cleavage patterns
required that shorter DNA fiagments in one strand correspond with longer fragments in
the other strand, indicating that the asymmetrical cleavage patterns were not a result of

over-cutting of the probe DNA.

The highly oriented binding of TBP to the AdMLP TATA sequence prompted further
exploration using variant TATA sequences. The AdMLP TATAAAAG sequence was
replaced with the symmetrical TATATATA sequence. TBP-OP affinity cleavage
directed by the symmetrical sequence led to equally distributed amounts of cleavage on
either side of the TATA sequence (Figs. 3C,D; 4C,D), providing validation of the
experimental techniques and indicating that the sequences flanking the AdMLP TATA
sequence do not dominantly orient TBP. When the TATA sequence was reversed to
CTTTTATA the DNA cleavage patterns were correspondingly reversed (Figs. 3E,F;
4E,F), indicating that the 8 basepair AdMLP TATA sequence is sufficient to direct highly

oriented binding of TBP to DNA.

To quantify the degree of TBP orientation, raw fluorescence data were summed in
adjacent l3 basepair regions centered on the cleavage patterns. The 13 basepair
quantitation regions are sufficiently wide to capture the specific cleavage signals. It
should be noted that the cleavage patterns were not precisely centered on the TATA

sequence, but were shifted in the 5’ direction on each strand (Fig. 5), and thus the centers

49

2500 -

 

 

 

 

1500 -
500
:3 l—___l
<1: l—l
-500 -
-1500 -
-2500 -
RAW DATA POSITION

Figure 5: DNA cleavage patterns mapped onto the cleavage probe sequence. Probe DNA
encoding a symmetrical TATA sequence was cleaved with K97C-OP, and the resulting
capillary electrophoresis data was aligned to the probe DNA sequence. The eight basepair
TATATATA sequence is represented by the box on the horizontal axis. top strand data is
plotted as positive arbitrary units (AU) and bottom strand data is plotted as negative AU.

of the cleavage patterns were determined by direct inspection. To correct for background
DNA cleavage, a signal baseline was approximated by a straight line connecting
fluorescence minima in regions upstream and downstream from the quantitated regions.
TBP orientation was defined as the baseline-corrected signal in one quantitation region
divided by the total baseline-corrected signal (Fig. 6). Orientation comparisons were
performed using top strand data, due to the occasional presence of an anomalous signal in
the bottom strand cleavage pattern. The source of this anomalous signal has not been
determined. Indeed, in side-by-side affinity cleavage reactions, built from common
stocks and processed in parallel, the anomalous signal was prominent in one sample but
absent in the other (Fig. 7), precluding any obvious correlation with the affinity cleavage

experimental procedures.

Using a quantitation of top strand cleavage patterns, El 88C-OP was found to be 90%
oriented on the forward TATA sequence, but only 78% oriented on the reversed TATA
sequence (Table 1). This difference may indicate over-digestion of the probe DNA, since
a bias toward shorter top strand fragments would increase the apparent orientation of
E188C-OP on the forward TATA sequence and correspondingly reduce the apparent
orientation on the reversed TATA sequence. A similar shift in the K97C-OP orientation
measures can be explained by a bias toward shorter fragments. Additionally, the
cleavage patterns from a TBP mutant (F2271, discussed below) are also consistent with a
bias toward shorter DNA fragments. If the cleavage reactions have a slight bias toward
shorter DNA fragments, then actual measure of TBP orientation will lie between the

positively and negatively shifted orientation measures. With this reasoning, the degree of

51

2000

 

‘12

E 1500

LL}

D

Z

8

a: 1000

o

:2

-—1

LL

3 500
0
4000 4200 4400 4600 4800 5000

RAW DATA POSITION

Figure 6: Method of DNA fragment pattern quantitation. Raw fluorescence units were
summed in 13 bp regions B-C and CD. A signal baseline (slanted line) was approximated
by a straight line connecting minima found in 17 bp regions A-B and D-E. The area below
the baseline was subtracted from the raw data sums from regions B-C and OD. Orientation
was calculated as the ratio of the baseline-corrected signal in B-C or CD divided by the
total baseline corrected signal in B-D.

52

 

FLUORESCENCE UNITS

 

 

 

RAW DATA POSITION

 

 

 

FLUORESCENCE UNITS

W W

RAW DATA POSITION

 

 

 

 

 

 

Figure 7: Anomalous peak in bottom strand DNA fragment analysis. Vertical bars indicate
13 base quantitation regions centered on the cleavage patterns. Cleavage reactions (A)
and (B) were built from common stocks and performed in parallel. In (B), an extra peak
is present in the first quantitation region.

53

Table l: Affinity cleavage measurements.

 

TBP orientation measurements

 

 

TBP TATA Top strand“ Standard Calculatedb
version version orientation Samples deviation orientation

K97C-OP Forward 71% 10 2.4% 75%
Reverse 79% 3 4.3%
Symmetric 50% 3 1 .6%

E188C-OP Forward 90% 10 2.2% 84%
Reverse 78% 3 3.6%
Symmetric 50% 3 0.9%

K97C-OP (F 2271) Forward 67% 7 3.5% 69%
Reverse 71% 3 0.7%
Symmetric 47% 3 0.2%

E188C-OP (F 2271) Forward 85% 6 3.5% 81%
Reverse 77% 4 2.9%
Symmetric 53% 3 l .2%

 

a Orientation relative to the direction of the TATA sequence. For symmetrical TATA
sequences, orientation is relative to TATA flanking regions.
Orientation calculated as the average of orientation measures obtained from forward
and reversed TATA sequences.

54

TBP orientation on the AdMLP TATA sequence has been defined as the average of the
top- strand orientation measures from forward and reversed TATA sequences. .
Accordingly, El88C-OP has been determined to be 84% oriented on the AdMLP TATA
sequence, and K97C-OP has been determined to be 75% oriented on the AdMLP TATA
sequence (Table 1). Our measures of TBP orientation are significantly different from the

values of 60%-64% (26) and 51% (67) reported by the Schepartz lab.

Since our measurements indicate that TBP can bind the TATA element with a high
degree of orientational specificity, there would appear to be little opportunity for TBP
orientational enhancement by Gal4-VP16. Nonetheless, the report that Gal4-VP16 can
significantly increase the orientational specificity of TBP orientation (67) led us to
replicate these reactions. In these experiments, Gal4-VP16 was bound to the probe DNA
prior to addition of TBP-OP. To ensure Gal4-VP16 association with the cleavage probe
DNA, a single preparation of DNA, TBP, and Gal4-VP16 was split into separate affinity
cleavage and footprinting reactions. DNA footprinting with methidium-propyl EDTA-
iron clearly showed protection of the Gal4 binding sites, indicating the presence of Gal4-
VP16 on the probe DNA (Fig. 8A). An analysis of affinity-cleaved probe DNA indicated
reduced amounts of cleavage at the TATA sequence, but the ratio of upstream and
downstream cutting was not significantly altered (Fig. 8B). These results indicated that

Gal4-VP16 does not alter the orientation of TBP binding to the AdMLP TATA sequence.

55

>

 

 

 

 

 

 

 

 

 

2500‘
m
E 2000' p I
z 1500, ‘
LL]
U
E; i
1
O 1000q .
D
A
1.1..
500'
E 1__H_H H H A
0 fi I I I I I
2000 2500 3000 3500 4000 4500 5000
RAW DATA POSITION
B
m
[-t
E 800-l
m 1
L2) 500 '1
m -
U
E; 1
200 ' '
O 1 Wk t
:1 , M 01 I ' ,
Bi 1‘1“ 111 . ALB w l‘“'“+‘kl‘-
“100 I I I I I I
3800 3900 4000 4100 4200 4300 4400

RAW DATA POSITION

Figure 8: Affinity cleavage pattern in the presence of Gal4-VP16. (A) Footprinting of
cleavage probe bound to Gal4-VP16. Bottom strand data is shown. Top trace, absence of
Gal4-VP16. Bottom trace, presence of Gal4-VP16. The location of the Gal4 binding sites
is indicated by the cluster of five boxes on the graph. The smaller, isolated box indicates
the position of the TATA sequence. (B) Affinity cleavage patterns in the absence (black
line) and presence (grey line) of Gal4-VP16. The cleavage data from a reaction including
Gal4-VP16 was scaled by a factor of three and overlayed with data from a reaction lacking
Gal4-VP16.

56

In an extension of our use of the affinity-cleavage method, we have participated in a
collaborative effort to study a TBP mutation that may lead to altered orientational
specificity. The laboratory of Dr. Karen Amdt has performed a yeast genetic screen that
selected for TBP variants supporting elevated levels of forward transcription from a
reversed TATA sequence in vivo. This screen identified the TBP mutant F2271. Three
simple models can be put forth to explain the ability of F2271 to exceed the ability of
wild type TBP at directing forward transcription from a reversed TATA sequence. First,
the F 2271 mutation may lead to an increased affinity for DNA, but not to changes in
orientational specificity. In this case, the amount of TBP facing forward on the reversed
TATA sequence would be increased simply due to more TBP occupancy at the promoter.
Second, the DNA affinity of F 2271 may be unaffected, but orientational specificity may
be relaxed or reversed. In this case the fraction of TBP facing forward would be
increased, again allowing increased forward transcription from a reversed TATA
sequence. Third, F 2271 may not affect either DNA affinity or orientational specificity,
suggesting that the effects of F2271 are mediated by altered interactions with other
components of the transcriptional machinery. The Amdt lab has determined that F 2271
has a normal overall affinity for DNA, implying either that the orientational specificity of

F2271 is altered or that interactions between F2271 and other factors are altered.

To further explore the properties of the F2271 TBP variant, we have used the affinity-
cleavage method to assess the orientational specificity of F 2271 bound to the AdMLP
TATA sequence. Affinity-cleavage reactions were performed using K97C-OP (F 2271)

and E188C-OP (F 2271) as the cleavage agents. The resulting DNA cleavage patterns

57

were quite similar to those from the wild type versions of TBP-OP, indicating a similar
high degree of orientational specificity. However, a careful comparison of the DNA
cleavage data showed that the F2271 mutation conferred a slightly relaxed orientational
preference. This relaxed orientation was most visible in an overlay of top-strand affinity
cleavage results from wild-type and F 2271 TBP-OP (Fig. 9). Several repetitions of the
affinity cleavage reactions have indicated that the difference, although small, is
reproducible. A quantitation of the wild type and F 2271 cleavage patterns indicated that
F2271 reduced the orientational specificity of El 88C-OP bound to the AdMLP TATA
sequence from 84% to 81% (Table 1). A Student’s t-test of the orientation measures
derived from forward TATA sequences indicated that the two sets of orientation
measures did not likely overlap (p<0.01). Similarly, the F 2271 mutation reduced the
forward orientation of K97C-OP from 75% to 69% and again the measurements from the

forward TATA sequence did not likely overlap (p<0.05).

58

21>

 

 

 

 

    
 

 

\
t“-
.MA —- E188C-OP
‘19 1'05“ — E188C-OP (F2271)
g ,a/ \‘if
z \
m \
B I!
a A
o t’ “\'
:2 .0 '2’ .\.
t-J I... 1 “'1' \\
a. 1' \,
I ‘ \ new.
RAW DATA POSITION
B
m — K97C-OP
E — K97C-OP (F2271)
1.1.}
U
z
LU
U
E”
8
E.) "I"; .N‘? MTV/‘1

 

f"

  

 

 

RAW DATA POSITION

Figure 9: Overlay of T BP-OP and TBP-OP (F 2271) top strand cleavage patterns. TBP-OP
and TBP-OP (F 2271) were used to cleave a forward-oriented AdMLP TATA sequence.
DNA cleavage patterns from several independent experiments were normalized to an
equal baseline-corrected area in the predominant cleavage region, and the resulting plots
were overlayed. Vertical bars indicate the upstream and downstream cleavage regions.
(A) E188C-OP top strand cleavage patterns. (B) K97C-OP top strand cleavage patterns.

59

Discussion

The previous report of a TBP-orienting activity residing in the VP16 activation domain
raised questions about the mechanistic details and potential relevance of this novel
mechanism of activation. For instance, given a panel of activation domain mutants with a
variable degree of functionality, how well would the in vitro orienting activity correlate
with in the in vivo transcriptional activation function? And how could the implicit
ternary character in the TBP orientation function be reconciled with other evidence
indicating that VP16AD, TBP and DNA could not form a ternary complex? To approach
these questions we have replicated and extended the affinity cleavage reactions initially
developed by the Schepartz lab. Unexpectedly, and contrary to the report from the
Schepartz lab, we found that isolated TBP bound to the AdMLP TATA sequence with a
high degree of orientational specificity. A review of the published data from the
Schepartz lab suggests that the precedent measures of TBP orientation may have been
affected by a significant and variable background contribution to the cleavage patterns,
leading to an incorrect conclusion that TBP is only modestly oriented on the AdMLP
TATA sequence. The clear signals and internal consistency in our affinity cleavage data
suggests that TBP in fact binds to the AdMLP TATA sequence with a high degree of

orientational specificity.

Although the affinity cleavage patterns produced by K97C-OP and E188C-OP
qualitatively agree, these two versions of TBP-OP appear oriented to different degrees.

This disagreement raises the possibility that neither version of TBP may provide an

60

accurate measure of TBP orientation. Since the stirrup loops lie in the minor groove of
DNA, the cysteine mutation and OP derivitization may interfere with DNA binding or the
intrinsic directionality of TBP. Indeed, other mutations in the stirrup loops have been
shown to alter the DNA binding properties of TBP. For instance, in the N-terrninal
stirrup loop of TBP, an A100P mutation conferred increased DNA affinity to TBP, and
also lent TBP the ability to recognize a reversed TATA sequence in vivo (157). A
substitution in the symmetrically related position in the C-terminal stirrup loop (P191A)
also altered DNA recognition, but in this case DNA binding was reduced (157).
Additionally, mutations of L189 and F190 abrogated DNA binding, although position
E188 tolerated mutations without a loss of TBP:DNA binding ability (131). These results
indicate that TBP is sensitive to mutation at positions located only 1 to 3 residues away
from both the K97C and the E188C. Since the current affinity cleavage methods attach a
bulky (~300 Da) moiety to a cysteine substitution flanking critical residues, it is possible
that the resulting TBP-OP proteins suffer from defects in DNA association. Nonetheless,
both K97C-OP and E188C-OP produce highly asymmetric and qualitatively similar DNA
cleavage patterns on the forward and reversed TATA sequences, suggesting that any
TBP-OP defects are modest, and that TBP binds to the AdMLP TATA sequence with a

high degree of orientational specificity.

Our studies also showed that the transcriptional activator Gal4-VP16 did not enhance the

orientational specificity of K97C-OP bound to the AdMLP TATA sequence. This result
is in disagreement with the precedent report from the Schepartz lab (67). In efforts to

reconcile these different results, the gel images from the precedent study were examined.

61

From sequencing gel images (figure 2 in (67)), it is apparent that a significant

background signal existed in some affinity-cleavage reactions and not others. This
background signal, when present, would dilute the specific affinity-cleavage bands, thus
reducing the apparent orientation of TBP. Since the background signal was reduced
when Gal4-VP16 was present, the apparent orientation of TBP increased upon addition of
Gal4-VP16. A visual correction for the background signal in the published gel images
suggests that the orientation of TBP is unchanged in the presence of Gal4-VP16. Thus
the conclusion from the Schepartz lab that Gal4-VP16 increased the orientation of TBP
may have arisen from a significant and variable background contribution to the specific

signals.

While our findings indicate that enhancement of TBP orientation is not a function of the
VP16 activation domain, the strong correlation of in vivo VP16AD activity with in vitro
VP16AD:TBP affinity suggests that TBP is nonetheless a relevant target of VP16AD
(120). Other lines of evidence show that VP16AD competes with other factors for
binding to the DNA-binding region of TBP. These competitive interactions, while
precluding the ternary interaction required for altering TBP:DNA orientation, might

allow for activation through a multi- step handoff mechanism (reviewed in Chapter 2).

Overall, the F 2271 mutation appears to have an effect on TBP orientational specificity,
reducing the forward oriented population of TBP by about 4-5%. Although this is a small
difference, it was detected when the cleavage activity localized to either stirrup of TBP,

and also when using either forward or reversed TATA sequences. Additionally, enough

62

data was acquired from forward TATA cleavage reactions to show that the effects of

F 2271 on orientational specificity were statistically significant. The modest reduction in
F 2271 orientational specificity raises the question of whether such a small change could
lead to the transcriptional differences detected in the in vivo assays performed by the
Amdt laboratory. In this respect, an accurate measure of the absolute degree of TBP
orientation would allow the determination of the relative change in forward-facing TBP
bound to a reversed TATA sequence. For instance, if TBP naturally binds to the AdMLP
TATA sequence with 84% forward orientation, then the F2271 mutation would be
expected to increase the forward- facing TBP on the reversed TATA sequence from 16%
to 21%, which may account for the increased transcription from directed by F 2271.
However, a limitation in this analysis is that we cannot precisely determine the absolute
degree of TBP orientation using the current affinity cleavage methodology, and thus we

cannot determine the relative change in TBP orientation conferred by the F2271 mutation.

Our finding that TBP is highly oriented on the AdMLP TATA sequence has implications
for models of PIC assembly. The current view of PIC assembly (reviewed in (50, 156))
assumes that TBP is essentially randomly oriented on the TATA element and that other
factors, such as TFIIB, are required to direct the orientation of TBP on a promoter. This
model has arisen primarily from two different lines of evidence. First, studies of the
relatively simple but homologous transcription apparatus of Archaea found that the
highly symmetric archaeal TBP is incapable of establishing transcriptional directionality,
and that the TFIIB-related factor TF B is required to define the orientation of transcription

(7, 82, 98). Second, the results from the Schepartz lab suggested that eukaryal TBP, like

63

archaeal TBP, is also incapable of directional recognition of the TATA sequence, and
that other factors are required to establish a productive orientation of the eukaryal
TBP:DNA complex (26, 67, 68). In the PIC assembly model arising from these
observations, the orientation of the TATA sequence is assumed to be a modest factor in
overall promoter architecture. Our results, contrary to the report from the Schepartz lab,
show that the eukaryal TBP can recognize the TATA sequence in a highly directional
manner, suggesting that the archaeal and eukaryal TBP homologs may be functionally
divergent, and that the orientation of the TATA sequence in eukaryal promoters may be a

more significant variable in promoter architecture than currently believed.

64

Chapter IV

Attempts to determine the structure of VP16C bound to TBP

Introduction

The transcription of eukaryotic protein-coding genes can be stimulated by cis-acting
regulatory sequences that may be located from tens to thousands of bases away from the
core promoter region and the transcriptional start site. The link between cis-acting
regulatory sequences and transcription is made by trans-acting proteins known as
transcriptional activators. The dual functions of transcriptional activators, DNA
recognition and transcriptional activation, typically comprise distinct domains of the
protein structure. While the principles of DNA recognition by transactivators are well

characterized, the mechanisms and targets of activation domains remain elusive.

Studies of transcriptional activation domain function frequently employ the unusually
potent activation domain of the VP16 transactivator from the herpes simplex virus as a
model system (146). The VP16 activation domain (VP16AD) resides in the C-terminus
of the VP16 protein (residues 411-490) (49, 164, 174), and can be further divided into
independently acting subdomains VP16N (residues 412 to 456) and VP16C (residues
457-490) (139, 160, 169). NMR analyses of the isolated VP16AD have indicated that it
exists in an unstructured state in solution (61, 126), although it can assume a helical
conformation under certain conditions (30) and when bound to TAF 32 (166) or PC4 (61).

The VP16AD has a preponderance of acidic residues, suggesting that the VP16AD may

65

exist as an “acidic blob” (154), enabling nonspecific interactions with a variety of targets.
Thorough mutational analyses of the activation domain have shown, however, that the
acidic character of the VP16AD is not critically required for activity. Mutation of one or
a few acidic residues to uncharged residues had only a modest effect on in vivo function,
whereas bulky hydrophobic residues in a few key positions are critically required for
activity (28, 139, 160). The requirement for bulky hydrophobic residues in key positions
suggests that a hydrophobic core and tertiary structural elements play a role in VP16AD
structure. Whether these bulky hydrophobic residues are involved internally to a

VP16AD fold or in interfaces with target proteins remains unknown.

Various studies have identified a wide range of proteins that interact either directly or
indirectly with VP16AD. Among these are several of the generally-required basal
transcription factors, including TFIIA (76, 77), TFIIB (43, 96, 169), both the TBP (56,
153, 159) and TAF9 (75) subunits of TFIID, and TFIIH (179). Additionally, components
of the RNA polymerase 11 associated mediator complex have been shown to interact with
VP16AD (92, 114, 182). Associations with chromatin- modifying coactivators, which
alter chromatin structure either covalently or noncovalently, have also been reported.
However, in this panoply of possible VP16AD interactions, little evidence establishes
which particular interactions are functionally relevant in the complicated process of in

vivo transcriptional regulation.

TBP stands out as likely being a relevant target of the VP16AD. TBP, a component of

the multisubunit TFIID complex, is a central player in all classes of eukaryotic

66

transcription, and is thought to mediate the initial protein-DN A contacts in the assembly
of the large pre- initiation complex (PIC) that poises RNA polymerase II at the
transcriptional start site. A quantitative assessment of TBP interaction with a spectrum of
VP16C mutants revealed tight binding of TBP to wildtype VP16C, and a robust
correlation between in vitro binding affinity and in vivo function, consistent with the
hypothesis that TBP interaction is important for VP16C function (120). Additionally,
time-resolved fluorescence anisotropy techniques have shown that critical hydrophobic
residues in both VP16 activation subdomains become structurally constrained and
protected from solvent when bound to TBP, indicating an acquisition of structural

elements in the otherwise unstructured VP16AD (152, 153).

While TBP appears to be a relevant target of the VP16AD, the detailed activation
mechanism probably involves more than a simple recruitment of TBP to a promoter.
Acidic activation domains in general do not bind cooperatively with TBP at promoter
DNA sequences (94, 180, 181). Additionally, a mutation in the concave DNA binding
region of TBP (L114K) abrogates interactions with VP16AD (72), indicating that the
activator and DNA interact with overlapping regions of TBP. However, despite this
competition of acidic activators and DNA for binding to TBP, acidic activators can
nonetheless stimulate formation of the TBP-containing TFIID-TFIIA-DNA complex (76,
77). Notably, in the TFIID complex, the intrinsic ability of TBP to bind DNA is inhibited.
Two inhibitory activities have been mapped to N-terminal domains of the TFIID
component TAF 1 (80). These domains, TANDl and TAND2, have been shown to bind

TBP in competition with DNA (80, 84, 99) and TF 11A (3, 79), respectively. The

67

observation that VP16C binds TBP competitively with the negative regulator TANDl
(124) suggests a cascade of interactions with the concave surface of TBP, with TANDl
being displaced by VP16C, and VP16C subsequently displaced by DNA. In this
“handoff model” (83) (Fig. 10), an activator functions as an anti- repressor when
interacting with TBP in the context of TF 11D. Importantly, the ability of VP16C to
stimulate TFIID-TFIlA-DNA assembly in vitro correlates with the ability of VP16C to
activate transcription in vivo (77), suggesting that TFIID-TFIIA-DNA assembly is a
relevant mode of transcriptional activation in vivo. Furthermore, in support of the model
that competition with TANDl can activate transcription, TANDl itself functions as an

activator in vivo when fused to a DNA binding domain (83).

68

  
 

 

 

  

 

 

Activation
Domain \ TAND2 / \ TBP
PNf‘ TANDl
brndrng \
ITATAI
—>
ITATAI

 

 

 

Figure 10: Handoff mechanism of TFIID-TFIIA-DNA assembly. In the upper panel TBP
is embedded in the TFIID complex, with TANDl competing with DNA binding and
TAND2 competing with TFIIA binding. In the middle panel, the VP16C has evicted
TANDl, which may facilitate TFIIA competition with TAND2 for binding TBP. In the
bottom panel, DNA has replaced VP16C, and TFIIA and TFIID are bound to the core
promoter.

69

VP16 complexed with TBP

Rationale

Our current understanding of activation domain structure/function relationships is limited
by an absence of structural data. The evidence that intrinsically unstructured activation
domains acquire structural features upon binding to target proteins suggests that
complexes of activators and targets represent the best candidates for structural analysis.
The plausibility of the handoff model and the evidence for structural elements in VP16C
when bound to TBP suggests that the VP16C:TBP protein complex represents a
structured intermediate in a relevant mechanism of transcriptional activation. A high
resolution structural model of this complex would clearly localize the VP16C binding site
on TBP, providing insights into the handoff mechanism. It would provide a basis for the
prediction and analysis of activation domain mutations, and might guide a precise
dissection of the interface of VP16C and TBP. An activation domain structure allows the
definition of a structure-based activator motif, and also enables a computational docking
approach to search for other targets among known protein structures. Taken together, the
wealth of information available from a high resolution structural model warranted our

attempts to study the structure of the VP16C:TBP complex by x-ray crystallography.

70

Materials and methods

Purification of VP16C. The expression vector pGEX.VPl6C (119), encoding a fusion of
GST, a thrombin cleavable linker, and VP16 residues 452-490 under control of the tac
promoter, was transformed into BL21(DE3)Codon-plus E. coli cells. Cells containing the
expression plasmid were grown in LB medium containing 80 mg/L ampicillin and 40
mg/L chloramphenicol. One L cultures were grown at 37 °C with vigorous shaking until
reaching an OD600 of 0.6-0.9. Expression of the GST-VP16C fusion protein was induced
by addition of IPTG to a final concentration of 0.2 mM and growth was continued for an
additional 3 h. Cells were harvested by centrifugation and resuspended in 10 mL
HEMGT buffer (24 mM HEPES, 0.1 mM EDTA, 12.5 mM MgC12, 10% v/v glycerol,
0.1% v/v TWEEN-20, pH 7.9) with 250 mM KCl. Resuspended cells were lysed by
three passages through a French Press (Aminco) operating at 20,000 psi. Cellular lysates
were centrifuged at 10,000 x g at 4 °C for 10 min to sediment insoluble matter. The
soluble lysate fraction was diluted lO-fold with HEMGT-250 containing 5 mM DTT,
mixed with 1 mL GSH resin, and gently rotated at 4 °C for 3 h. The resin was collected
and washed with two 5 mL aliquots of HEMGT-250, two 5 mL aliquots of HEMGT- 100,
and two 5 mL aliquots of thrombin digestion buffer (20 mM Tris-Cl, 150 mM NaCl, 2.5
mM CaC12, pH 8.4). The resin was resuspended in 2 mL thrombin digestion buffer and
biotinylated thrombin was added to liberate VP16C from the resin-bound fusion protein.
Digestion was performed at 4 °C for 4 h with gentle rotation. The supernatant and a 1

mL wash were collected and incubated with 0.2 mL streptavidin agarose beads to trap the

71

biotinylated thrombin. The supernatant and a 0.5 mL wash were collected and stored at ~

80 °C.

The VP16C peptide was further purified by anion-exchange HPLC. VP16C solutions
were diluted tenfold with 20 mM Tris buffer (pH 8.5) to lower the salt concentration and
bound to a DEAE ion-exchange HPLC column (TSK DEAE-SPW, 2.15 x 15 cm,
Beckman). The column was developed with a 275 mL linear gradient from 0 to 750 mM
NaCl in 20 mM Tris pH 8.5 at a 5 mL/min flow rate. VP16C typically eluted at about
435 mM NaCl. VP16C concentration was estimated by absorbance at 280 nm and by a

bicinchoninic acid assay (BCA assay, Pierce) relative to BSA standards.

Purification of yeast T BP. The plasmid trc-TBP, encoding the fusion of a histidine tag, a
thrombin-cleavable linker, and Saccharomyces cerevisiae TBP residues 61-240 under
control of a T7 promoter, was obtained from Dr. J .H. Geiger. This plasmid was
transformed into E. coli strain BL21(DE3)Codon-plus. Cells containing the plasmid were
grown in LB medium containing 80 mg/L ampicillin and 40 mg/L chloramphenicol. One
liter cultures were grown at 37 °C with vigorous shaking to an OD600 of about 1.0.
Expression of TBP was induced by addition of IPTG to a final concentration ofO. 15 mM,
cultures were transferred to 18 °C (ambient), and growth was continued for 12-14 h with

vigorous shaking. Cells were harvested by centrifugation and stored at ~80 °C.

Cell pellets from three liters of culture were resuspended in 50 mL lysis buffer (50 mM

sodium phosphate, 20% v/v glycerol, 500 mM NaCl, 1 mM EDTA, 5 mM DTT, 0.1%

72

Triton X- 100, 5 mM imidazole, pH 7.5). The resuspended cells were lysed by 2 passages
through a French Press (Aminco) operating at 20,000 psi. Cellular lysates were
centrifuged at 45,000 x g at 4 °C for 25 min to sediment insoluble matter. The soluble
lysate fraction was collected and mixed with 2.5 mL of nickel-affinity resin

(Ni-NTA Superflow, Qiagen) and gently rotated at 4 °C for 1.5 hours. The resin was
collected, washed 3 times with 10 mL lysis buffer with no imidazole, and 4 times with
lysis buffer containing 40 mM imidazole. Bound proteins were eluted with lysis buffer
containing 150-200 mM imidazole and collected in 1 mL fractions. TBP typically eluted

in the first 3 elution fractions.

TBP was further purified by ion-exchange F PLC. TBP solutions were mixed with a
diluent (20 mM Tris, 10% glycerol, 1 mM DTT, pH 7.5) to lower the salt concentration
to 150 mM. Diluted TBP was passed through a 1 mL Q-Fast F low anion-exchange
column (Pharmacia) and loaded onto a 1 mL SP-Fast Flow cation-exchange column
(Pharmacia). After loading, the Q column was disconnected, and the SP column was
developed with a linear gradient from buffer A (20 mM Tris-Cl, 10% glycerol, 150 mM
NaCl, 1 mM DTI‘, pH 7.5) to buffer B (Buffer A with l M NaCl) over 20 column

volumes at 1 mL/min flow. TBP typically eluted at around 540 mM NaCl.

Protein crystallization. VP16C and TBP were mixed at a 1.4 to 1 molar ratio, and the
mixture was buffer swapped to crystallization buffer (10 mM Tris-Cl, 10% v/v glycerol,

250 mM NaCl, 2 mM DTT, pH 8) and concentrated to 4 mg/mL in a YM~3 centrifugal

concentrator (Amicon).

73

Results and discussion

The VP16C activation domain (VP16 residues 452-490) was expressed as a GST fusion
protein, purified over GSH sepharose, cleaved from GST by thrombin digestion, and
subsequently HPLC purified over anion exchange resin (Fig. 11A) (119). The yield of
purified protein was approximately 1 mg per liter of initial culture. The conserved core
region of yeast TBP (residues 61-240) was expressed as a histidine tagged fusion protein,
and purified using nickel affinity chromatography (Fig. 11B,C). Notably, the expressed
protein was mostly insoluble, but nonetheless about 1 mg of purified TBP could be
recovered per liter of initial bacterial culture. The purity of the both proteins was judged

as acceptable for attempts at protein crystallization.

Purified VP16C and TBP were mixed at a 1.4:] molar ratio and concentrated to 4 mg/mL
total protein, as assayed by a Bradford dye binding assay. Because low salt conditions
lead to TBP instability, a protein buffer containing 250 mM NaCl was chosen.
Concentrated protein solutions were mixed with equal volumes of a diverse set of
precipitating mixtures, and 2 uL drops were deposited on plastic cover slips, inverted,
and sealed over the corresponding precipitating mixtures in a standard hanging-drop
vapor diffusion crystallization configuration. Plates were incubated at room temperature

and 4 °C and observed for crystal growth.

74

 

 

 

 

 

 

 

 

d)
A .c 8 5.25
0 .29 t: 0
338-1: 93
-o o 171‘
285 8 to:
3.50009:
ma 5;:
434:3, 5'1
m -
29 — ~ ' g. — GST-VP16C
18"" ~-——VP16C
14-;: £3
6.2—1“'~“
12 3 4 5 6
B C 3%
8 ., :2:
:1 o -‘-c
'8 o gB—«va r:
a? afiasaes
3 reacting
kDa 3 kDa 08333381
83
47 -
32
.-
25 -
- -—TBPC
16.5 ‘5
.5 Q!
1

N
03
A
LII
O\
\I

Figure 11: Purification of VP 16C and TBP. Protein samples were resolved by SDS-PAGE
and stained with Coomassie R-250. (A) Purification of VP16C. Lanes 1-3, cellular
extracts. Lane 4, proteins retained on GSH resin. Lane 5, products of thrombin digestion.
Lane 6, thrombin-liberated protein after HPLC purification. (B) Expression of the 6xHis-
tagged core region of yeast TBP (yTBPc) in E. coli, monitored by whole-cell extracts. (C)
Nickel-affinity purification of 6xHis-yTBPc. Lanes 1-2, cleared lysate before and after
incubation with nickel resin. Lanes 3-6, proteins released by washing with 40 mM imidaz-
ole. Lane 7, protein released by 200 mM imidazole.

75

Short-lived needle-shaped protein crystals were observed in a precipitating condition
comprising 100 mM sodium acetate, 200 mM ammonium sulfate, and 25% PEG 4000,
pH 4.6. Subsequent refinement of the three components of this crystallization buffer led
to a condition comprising 100 mM sodium acetate, 350 mM ammonium sulfate, 25%
PEG 4000, pH 5.0. While these conditions produced improved crystal quality and
longevity, crystal growth was still primarily one dimensional and the resulting crystals
were not suitable for diffraction studies. Notably, a protein solution containing TBP but
omitting VP16C led to visually indistinguishable protein crystals under otherwise
identical conditions, suggesting that the crystals obtained from the VP16C:TBP mixture
were likely to comprise only TBP, probably in a dimerized form (123). At this point
crystallization efforts were redirected to a tighter protein complex that apparently

preserves the VP16C:TBP interaction, described in the next section.

76

VP16C-TAND2 bound to TBP

Rationale

TBP exists as a dimer in solution, with the concave DNA binding regions buried in the
dimer interface (23). As VP16C likely binds to a region of TBP overlapping the DNA
binding region, it is likely that TBP dimerization competes with the VP16C:TBP
interaction. Since the dissociation constant of TBP dimers is in the low nanomolar range
(23) and the dissociation constant of the VP16C:TBP interaction is about 40 nM (120,
153), TBP dimerization is favored over TBP:VP16C interactions. Increasing the strength
of the VP16C:TBP interaction would work to push the equilibrium away from TBP

dimerization and toward the desired VP16C:TBP complex.

In domain swap experiments aimed at demonstrating functional equivalence between
VP16C and TANDl , Kotani and coworkers (83) constructed a fusion of VP16C and
yTAND2. This fusion displayed strong binding to TBP, and was able to bind and retain
TBP through extensive washing, whereas neither TANDl, TAN D2, nor VP16C alone
could retain significant amounts of TBP. The VP16C-TAN D2 fusion also appeared to
preserve a relevant VP16C:TBP interface. The TBP mutant L114K, which is deficient in
interacting with VP16C, did not bind well to the VP16C-TAND2 firsion. Furthermore,
VP16C, in the context of the same fusion to TAND2, could complement a TANDl-
deletion phenotype in yeast, but a transcriptionally-defective mutant of VP16C failed to

substitute for TANDl, indicating a sensitivity to critical mutations of VP16C.

77

Since the VP16C-TAND2 fusion binds TBP more tightly than VP16C alone, and since
this fusion appears to preserve the transcriptionally relevant VP16C:TBP interface, it is
an attractive target for structural characterization (Fig. 12A). The following sections
describe attempts to crystallize a complex between a VP16C-TAND2 fusion and the

conserved core region of yeast TBP.

78

Materials and methods

Plasmid construction. TAND2 (TAF 1 residues 43-73) was PCR amplified from S.
cerevisiae genomic DNA using adapter-primers ST805 and ST806. The resulting
amplified DNA segment was tailored by EcoRI and X7101 restriction enzyme digestion
and directionally cloned between the EcoRI and XhoI sites of plasmid pGEX4T-2
(Pharmacia), resulting in plasmid pDS42-5. VP16C (VP16 residues 452-490) was PCR
amplified from plasmid pGEX-VP16C using primers ST803 and ST804. The resulting
amplified DNA segment was tailored by BamHI and EcoRI restriction enzyme digestion
and directionally cloned between the BamHI and EcoRI sites of plasmid pDS42-5 to
create plasmid pDS42~ 10, encoding a direct fusion of VP16 Gly 490 to TAND2 Gly 41
as a GST fusion protein under control of the tac promoter. Plasmid sequences were

verified by sequencing.

Expression of GST— VP16C-TAND2. Plasmid pDS42-10 was transformed into E. coli
strain BL21 and cells containing the expression plasmid were grown in LB medium
containing 80 mg/L ampicillin. One liter cultures were grown at 37 °C with vigorous
shaking until an OD600 of about 0.7. Expression of the fusion protein was induced by
addition of IPTG to a final concentration of 0.5 mM and grth was continued for an
additional 3 h. Cells were harvested by centrifugation and cell pellets were resuspended
in 10 mL HEMGT buffer (24 mM HEPES, 0.1 mM EDTA, 12.5 mM MgC12, 10% v/v
glycerol, 0.1% v/v TWEEN-20, pH 7.9) with 250 mM KCl. Resuspended cells were

lysed by two passages through a French Press (Aminco) operating at 20,000 psi and

79

insoluble matter was sedimented by two steps of centrifugation at 10,000 x g at 4 °C for
10 min. The soluble lysate fraction was collected, split into 1 mL aliquots, flash frozen in

liquid nitrogen, and stored at ~80 °C.

Expression of yeast TBP. 6xHis-tagged yeast TBP was expressed in E. coli as described
above. The soluble fraction of the bacterial lysate was split into 1 mL aliquots, flash

frozen in liquid nitrogen, and stored at ~80 °C.

Purification of the VP16C-TANDZ: TBP complex. The VP16- TAND22TBP complex was
typically prepared by mixing 6-10 volumes of lysate containing TBP per volume of lysate
containing GST-VP16C-TAND2. Lysate mixtures were diluted with ice-cold HEMGT
buffer (24 mM HEPES, 0.1 mM EDTA, 12.5 mM MgC12, 10% v/v glycerol, 0.1% v/v
Tween-20) to reduce the salt concentration to 100 mM and incubated on ice for 1 h. The
lysate mixture was added to GSH resin and the mixture was gently rotated for 45 min at 4
°C. The resin was washed 4 times with 75 resin volumes of ice-cold HEMGT with 100
mM KCl. The washed resin was resuspended to a 50% slurry and biotinylated thrombin
protease was added. Digestion was allowed to proceed for 14-16 h at 4 °C with gentle
rotation. Liberated proteins were collected by column-based elution, including a wash
step supplementing the protein solution to final concentrations of 5 mM DTT and 0.05%
w/v NaN3. Eluted proteins were incubated with streptavidin agarose to trap the
biotinylated thrombin and the remaining protein was concentrated in a centrifugal

concentrator (Amicon) to 5- 10 mg/mL total protein.

80

HPLC analysis and mass spectrometry. To prepare samples for reverse-phase HPLC,
purified proteins were diluted in water with 0.1% trifluoroacetic acid (TF A).
Alternatively, protein crystals were harvested and dissolved in acetonitrile with 0.1 %
TFA. Samples were bound to a C18 reverse-phase HPLC column (Vydac, 5 mm bore,
300 A particles). The column was developed with a linear gradient from Buffer A (water
with 0.1% TF A) to buffer B (90% acetonitrile, 10% water, 0.1% TF A), and eluted
material was observed by absorbance at 214 nm. The molecular mass of the proteins in
the HPLC peak fractions was determined by matrix-assisted laser desorption/ionization-

time of flight mass spectrometry.

81

Results

The design of the fusion protein comprising VP16C and yeast TAND2 was based on a
VP16C-TAND2 fusion developed by Kotani and coworkers (83). The fusion protein
incorporated the same junction point as the precedent, but was carefully terminated at the
boundaries of the domains interacting with TBP. For VP16C, an alanine scanning study
indicated that residues 457 and 458 were involved in activation in both yeast and
mammalian activation assays (160), and thus the N-terminus of VP16C was defined as
residue 452 to allow a native context for these residues. A deletion analysis of TAND2
(79) indicated that yeast TAF 1 residues 10~73 were sufficient for a tight connection to
TBP, whereas inclusion of 15 more residues conferred little additional binding strength.
Thus, the fusion protein design includes VP16C (452-490) fused directly to TAND2(41~

73).

The GST-VP16C-TAND2 (GVT) fusion protein (Fig. 12A,B) was expressed in E. coli.
Soluble lysate fractions were bound to GSH resin and washed thoroughly. SDS-PAGE
analysis of resin-bound proteins indicated recovery of a pure species with little sign of
truncation or degradation (Fig. 12C). An elution of bound material with reduced GSH
followed by protein quantitation indicated that about 60 mg of GVT could be recovered

per liter of bacterial culture.

82

 

 

 

 

B | GST H VP16C | TAND2 |
C kDa

83 —

47 _ GST-VP16C-TAND2

-O/
32 — -' ——GST—VP16C
- - "'
25 _ \GST
16.5 —

 

 

 

Figure 12: Design and purification of a VP16C-TAND2 fusion protein. (A) Predicted
arrangement of VP16C and TAND2 when bound to TBP. (B) Design of the GST-VP16C-
TAND2 fusion protein. The link between GST and VP16C encodes a thrombin recogni-
tion sequence. (C) SDS-PAGE gel of GST, GST-VP16C, and GST-VP16C-TAND2
proteins purified over GSH resin. Each pair of lanes represents elution of resin-bound
proteins (left) by 10 mM reduced GSH and (right) by boiling in SDS-PAGE loading
buffer. Proteins stained with Coomassie R-250.

83

To test the ability of GVT to associate with TBP, a bacterial lysate containing 6xHis-
yTBPc was added to the GVT-containing lysate prior to purification of GVT. GVT
effectively bound TBP and retained it through the subsequent washing steps, resulting in
a clean purification of both GVT and TBP (Fig. 13A). The GVTzTBP association was
found to be stable over a range of conditions, but could be weakened by a combination of
elevated salt concentration and elevated temperature (Fig. 13B,C). Subsequent
experiments have exploited the specificity and salt sensitivity of this interaction as an

effective means of purifying untagged TBP (Chapter 2).

To saturate the GVT with TBP, TBP-containing lysate was titrated against a constant
volume of GVT-containing lysate, and bound proteins were analyzed by SDS-PAGE.
Typically 6 to 10 volumes of TBP-containing lysate were sufficient to saturate the TBP
recovery (data not shown). A thrombin digestion was used to liberate the VP16C-
TAND2:TBP complex from the resin and also to remove the hexahistidine tag from
yTBPc (Fig. 14A). The resulting protein was recovered and concentrated to between 5
and 10 mg/mL total protein. Reverse-phase HPLC resolved the purified protein mixture
into two distinct peaks (Fig. 14B). Mass spectrometry of the material eluted in the HPLC

peaks indicated masses corresponding to intact VP16C-TAND2 and TBP.

84

 

kDa
83 —
47 — GST-VP16C-TAND2
- - /
32 - 2: S: — GST—VP16C
25 _ ’ C \
a. . GST
16.5 — \
TBP

 

 

 

 

 

B o o
0 C 25 C Temperature
100 200 400 100 200 400 KCl (mM)
15 90 15 90 15 90 15 90 15 90 15 90 minutes
bé-n-vuun-JA.. ‘NIII-cho- '—'TBP

 

 

 

 

 

 

Figure 13: Association of GST-VP16C-TAND2 and TBP. Protein samples were resolved
by SDS-PAGE and stained with Coomassie R-250. (A) E. coli lysates containing GST,
GST-VP16C, or GST-VP16C-TAND2 were incubated with an E. coli lysate containing
6xHis-TBPc, and the resulting mixtures were purified over GSH resin. (B) Resin-bound
complexes were subjected to washing under different conditions as indicated.

85

Bound Proteins
Digest slurry
Digest SN

 

GST-VP16C—TAND2

 

 

 

 

"' _/ GST
-
-- 6xHis-TBPc
TBPc
B
120 —
100 - VP16C-TAND2 TBP

    

\

214 nm absorbance (AU)
m
o

 

Minutes

Figure 14: Purification of the VP16C-TAND22TBP complex. (A) E. coli lysates containing
GST-VP16C-TAND2 and 6xHis-TBPc were mixed, incubated, and purified over GSH
resin. SDS-PAGE was used to resolve the bound proteins, products of thrombin digestion,
and liberated proteins. Proteins were stained with Coomassie R-250. VP16C-TAND2,
which migrates near the dye front on SDS-PAGE gels, is not shown. (B) Reverse-phase
HPLC analysis of proteins liberated by thrombin digestion. HPLC peak fractions were
identified by mass spectrometry.

86

Concentrated VP16C-TAND2:yTBPc was subjected to crystallization trials using both
hanging drop and under-oil methods. The growth of needle-shaped crystals was observed
with a precipitating mixture comprising 100 mM Tris, pH 8.5, 200 mM lithium sulfate,
and 30% PEG 4000 (data not shown). A series of refinements of the crystallization
variables led to improvements in crystal quality, but no conditions were found to promote
crystal grth in more than one dimension. As crystal growth was promoted by
increasing lithium sulfate concentrations, the effects of this salt on the VP16C~
TAND2:TBP interaction was examined. By washing immobilized GST-VP16C~
TAND2:TBP complexes with buffers containing various salts it was determined that
lithium sulfate reduced the amount of retained TBP (data not shown), which suggested
that lithium sulfate was effective at disrupting the TBP association. This observation
suggested that the crystals obtained in the presence of elevated concentrations of lithium

sulfate comprised TBP dimers liberated from the VP16C-TAND2 fusion protein.

Additional screening of crystallization conditions revealed that precipitant mixtures
comprising 20% PEG 8000 with either 100 mM HEPES, pH 7.5, or 100 mM CHES, pH
9.5, could produce plate~ like crystals. Several rounds of refinement led to precipitant
mixtures comprising 100 mM CHES or HEPES, pH between 8.0 and 8.6, and PEG 4000
or PEG 8000 at between 4% and 12.5% v/v. These precipitant mixtures led to crystals
that were well~ grown in three dimensions (data not shown). These crystals were
amenable to cryoprotection by slow stepwise washing into a mother liquor mimic
containing 30% glycerol. Initial diffraction data indicated a unit cell of 120x90x60

angstroms and diffraction peaks consistent with crystallized protein. However, reverse-

87

phase HPLC analysis of dissolved crystals indicated that they comprised TBP and a
lesser amount of an unidentified component, but not VP16C-TAN D2 (data not shown).
In parallel with the peptide analysis, a full set of diffraction data were collected from the
crystals and solved by molecular replacement, confirming that the crystals comprised

TBP dimers and no VP16C-TAND2.

Since degradation of VP16C-TAND2 may have destabilized the binary complex and
released TBP for subsequent crystallization, the mother liquor of a crystal-bearing
condition was analyzed for intact proteins. Reverse-phase HPLC indicated a complex
mixture, but no clear peak corresponding to VP16C-TAND2 (data not shown). Although
VP16C-TAND2 was not detected, it must be noted that the assessment of the mother
liquor was made 35 days after the crystallization condition was assembled, and 25 days
after TBP crystals were harvested, allowing an extended window for VP16C-TAND2
degradation. A conclusive measure of VP16C-TAND2 integrity during the

crystallization period would require further experimentation.

88

Discussion

The VP16C-TAND fusion, modeled after the precedent from Nakatani and coworkers,
binds the core region of yeast TBP with high enough affinity for a co-purification of TBP.
The success of the co-purification technique, coupled with the salt-sensitivity of the

intermolecular interaction, has been developed as an effective means of TBP purification

(Chapter 2).

The TBP-only crystals derived from the VP16C-TAND2 complex indicate that the
complex releases TBP over time, which may indicate degradation of VP16C-TAND2.
Although VP16C-TAND2 is intact at times soon after preparation, an analysis of stability
over long term incubation has not yet been completed. Any further attempts to crystallize
this complex should be prefaced by assurances that VP16C-TAND2 is indeed stable over

the time periods required for protein crystallization.

Alternatively, the TBP crystals could have arisen from dissociation of the intact VP16C-
TAND22TBP complex. This could occur by a kinetically limited rearrangement of the
heterodimeric complex to thermodynamically favored TBP homodimers. Additionally, a
TBP crystal lattice may provide a thermodynamically stable “sink” for any free TBP that
exists in equilibrium with the heterodimeric complex. In either of these cases, increasing
the stability of the complex may be of benefit. One approach to enhancing the stability of

the complex, a direct fusion of VP16V~TAND to TBP, is explored in the next section.

89

VP16C-TAND2-TBP fusion protein

Rationale

Since protein crystals grown from a complex of VP16C-TAND2 and TBP contained only
TBP, it may have been that the complex, while stable enough to retain TBP through
purification, still allowed enough TBP to escape and for crystallization of TBP dimers to
occur. One method for further tightening the complex was to provide a direct tether by
encoding it as a single peptide chain, thus ensuring a high local concentration of VP16C-
TAND2 and TBP. Other advantages of this approach included a simplified purification,
and a resulting complex that would likely prevent the crystallization of dimerized TBP.
The risks included the incorporation of a flexible linker that may have masked protein
surfaces otherwise useful as lattice contacts, the imposition of steric constraints that may
have disrupted proper folding of VP16C or TAND2, and the possibility that the direct

fusion would interfere with proper folding of the TBP core domain.

90

Materials and methods

VP16-TAND2-TBP plasmid construction. A DNA fragment encoding VP16C-TAN D2
was PCR amplified from plasmid pDS42~10 using primers ST906 and ST907. The
resulting amplified DNA segment was tailored by Neal and BamHI restriction enzyme
digestion and directionally cloned between the Neal and BamHI sites of plasmid pET28a
to create plasmid pDSE104~ 1. A DNA fragment encoding TBPc (residues 61—240) was
excised from the plasmid TEV-TBP (obtained from J .H. Geiger) by BamHI and Xhol
restriction enzyme digestion, gel purified, and directionally cloned between the BamHl
and Xhol sites of plasmid pDSElO4-l to create plasmid pDSElO4-5. This plasmid
encodes VP16C(452~490) fused to TAND2(41~73), a single glycine, and TBP(61~240),

under control of a T7/lac promoter.

Linker extension: Cohesive Oligonucleotides ST908 and ST909 were phosphorylated
with T4 polynucleotide kinase. Plasmid pDSE104~5 was linearized with BamHI
digestion for 1 h, and then T4 DNA ligase, ATP, and phosphorylated Oligonucleotides
ST908 and ST909 were added directly to the digestion reaction. Extension of the
linearized plasmid by linker polymerization was allowed to proceed for 2 h at 12 °C.
Subsequently, BamHI and MluI were added directly to the ligation reaction, and digestion
was allowed to proceed at 37 °C for l h. The DNA segment corresponding to linker-
extended VP16C-TAND2 was gel purified and ligated to a BamHl-Mlul fragment of
pDSElO4-5 encoding the remainder of the parental plasmid sequence, in order to

recapitulate the original plasmid sequence with the linker extensions of TAND2.

91

Ligated plasmids were transformed into E. coli strain DHSor and transformed cells were
selected on LB with 50 ug/mL kanamycin. Transforrnant colonies were screened by
colony PCR using primers ST805 and ST925 that flank the extension site, and candidates
leading to PCR products of expected length were confirmed by sequencing, resulting in a

family of plasmids denoted pDSE106-n, where n represents the number of insertions of

the GSGS tetrapeptide sequence between TAND2 and TBP.

From the pDSElO4-5 and the pDSElO6-n series of plasmids, fragments encoding
TAND2-linker-TBP were excised by EcoRI and XhoI restriction digestion, gel-purified,
and directionally cloned between the EcoRI and XhoI sites of plasmid pDSC2-2
(encoding GST-VP16C-TAND2 under control of a T7/lac promoter), yielding sequences
encoding GST-VP16C-TAND2- linker-TBP inserted between the Neal and XhoI sites of
plasmid backbone pRSF ~DUET~1 (N ovagen). This plasmid family is denoted pDSE112~
n. Subsequently, the Ncol-Xhol fragments encoding the entire fusion protein were
subcloned into the pETDuet-l (Novagen) plasmid background, resulting in plasmid

family pDSEl 17-n.

Protein expression and purification. Plasmids encoding VP16C-TAND2-linker-TBP
were transformed into E. coli strain BL21(DE3)-codon plus. Cells containing the
plasmid were grown in LB medium containing 40 mg/L chloramphenicol and either 80
mg/L ampicillin or 50 mg/L kanamycin. One liter cultures were grown at 25 °C with

vigorous shaking until an OD600 of 1.0. Expression of the fusion proteins was induced by

92

addition of IPTG to a final concentration of 0.1 mM. Cultures were grown an additional
4 h. Cells were harvested by centrifugation and then resuspended in 20 mL wash buffer
(20 mM Tris-Cl, 5% v/v glycerol, 0.1 mM EDTA, 500 mM KCl, pH 7.9) with 5 mM
DTT. Resuspended cells were lysed by three passages through a French Press (Aminco)
operating at 20,000 psi. Insoluble matter was sedimented by centrifugation at 16,000 x g
for 30 min at 4 °C. The soluble lysate fraction was recovered and mixed with 0.75 mL
GSH resin and rotated gently for l h at 4-7 °C. The GSH resin was recovered and
washed 5 times with 75 resin volumes of ice-cold wash buffer. Thrombin was added to
liberate the fusion proteins from resin-bound GST, and digestion was allowed to proceed
for 8 to 14 h at 7 °C with gentle rotation. Eluted proteins were collected and further

purified by size exclusion chromatography.

Size exclusion chromatography. Preparative and analytical size exclusion
chromatography (SEC) was performed over Superdex S~200 prep grade media in a
Pharmacia XK-series 75/ 100 column using a Pharmacia F PLC system, operating at 7 °C,

with 1 mL/min flow rate and 280 nm absorbance detection of eluted material.

93

Results

In an attempt to stabilize a complex of VP16C-TAND2 and TBP a direct fusion of the
two proteins was explored. The fusion protein linkage was designed both to incorporate
a minimal length linking peptide segment and to insulate VP16C from linker- imposed
steric constraints. Since the TAND2 domain interacts with the convex upper surface of
TBP (79, 102), it is likely to be closer to TBP’s N-terminus than is VP16C, and thus
TAND2 provided a reasonable point for attachment to TBP. Thus, the C-terminus of the
VP16C-TAND2 fusion was bridged to the N-terminus of the core region of yeast TBP
(Fig. 15A,B). To allow for flexibility and a hydrophilic character to the linker peptide, it
was designed to encode a series of glycine-serine repeats. Since the location of the

TAN D2 C-terminus in unknown, 3 series of fusion proteins with increasingly longer

linker peptides was generated.

To generate this family of peptide linkers of varying length, complementary
oligonucleotides encoding Gly-Ser-Gly-Ser were designed (Fig. 15C). The annealed
oligonucleotide pairs have 5’ GATC overhangs on each end, allowing for polymerization
in the presence of DNA ligase activity. The sequences were designed such that only a
head-to-head ligation of the annealed oligo pairs creates a BamHI restriction site. In this
way, a polymer of randomly-oriented units, when digested by BamHI, results in polymer
of forward-oriented units of varying length, terminating with a BamHI- generated 5’-

overhang (Fig. 15D). This polymerization strategy was used to extend the DNA

94

A

   

Engineered linker /

 

L GST H VP16C | TAND2J(GSGS)L TBPcore

 

5 ’— GATCAGGCTCTG _
TCCGAGACCT AG -5’ —

.
MW%E%A%E%

 

Figure 15: Design and construction of a fusion of VP16C-TAND2-linker-TBP. (A)
Predicted arrangement of a fusion of VP16C-TAND2 bound to the core region of yeast
TBP, and placement of an engineered linker. (B) Primary structure of the fusion protein.
(GSGS) represents a variable number of GSGS tetrapeptide segments. (C) Cohesive
oligonucleotides that anneal to form a single extension unit. (D) Extension of the DNA
sequence encoding TAND2 by polymerization of multiple randomly—oriented extension
units. Only a head—to-head ligation of extension units encodes a BamHI recognition
sequence.

95

sequence encoding TAND2, and the extended sequence was ligated to the sequence
encoding the core region of yeast TBP. Resulting plasmids encoding different
characteristic lengths of the peptide linker were identified by colony PCR with primers
flanking the linker sequence, and candidate clones were verified by sequencing. This
procedure generated sequences encoding VP16C-TAND2-TBP firsions with 1 to 7
insertions of the GSGS linker segment, all resulting from a single execution of the

oligonucleotide extension reaction.

The firsion protein constructs were initially encoded in pRSF-based plasmids, which
confer kanamycin resistance and have a copy number of about 100 per cell. After
expression in E. coli, GST-VP16C-TAND2- linker-TBP was purified from the bacterial
lysates using GSH resin. The overall yield of purified protein was about 20 mg per liter
of initial culture. SDS—PAGE analysis of the purified proteins indicated a prominent
species of the expected molecular weight, and the molecular weight distribution seen
across the family of expression constructs confirmed the identity of the fusion proteins
(Fig. 16A). A significant amount of other proteins was present in the purification, and
these could not be separated from the intact fusions either by increasing the number of
wash steps or by increasing the salt concentration of the wash buffer. Since the
contaminating proteins primarily spanned the molecular weight range from GST to full
fusion protein, they were likely truncation or degradation variants of the full fusion

protein.

96

GSGS units

Fusion protein

 

012345675*GSGSunits

 

 

.... —— = . Fusion protein

 

 

 

Figure 16: Purification of GST-VP16C-TAND2-linker-TBP fusion proteins. (A) Fusion
proteins encoded by pRSF—based vectors were expressed in E. coli and purified over GSH
resin. Bound proteins were eluted by boiling in SDS-PAGE loading buffer, resolved by
SDS-PAGE, and stained with Coomassie R-250. (B) Fusion proteins encoded by pET-
based vectors were purified as in (A). Lane 5* denotes protein encoded by a pRSF-based
vector, loaded for comparison. GSGS denotes the number of GSGS linker insertions is the
fusion protein.

97

Since the antibiotic kanamycin interferes with protein synthesis, it was possible that
expression under a different selection might improve the quality of the purified proteins.
With this reasoning, the fusion protein constructs were subcloned into a lower copy pET-
based plasmid backbone that conferred ampicillin resistance. After expression under the
same conditions, the quality of the purified protein improved significantly (Fig. 16B),
although the overall yield diminished to about 3 mg purified protein per liter of initial
bacterial culture. Based on the increased quality of the recovered proteins, the pET-based

plasmids were used for all subsequent work.

It was hypothesized that if the engineered linker was too short, VP16C and TAND2
would not be able to reach and bind to their cognate surfaces on TBP. Accordingly, it
was expected that fusion proteins with engineered linkers below this critical threshold
would exist as TBP- linked dimers in solution, whereas fusion proteins with sufficiently
long linkers would exist as monomers in solution, with the TBP dimerization surface
occluded by VP16C (Fig. 17A,B). The apparent solution size of the series of fusion
proteins was resolved by SEC of the purified proteins (Fig. 17C). Interestingly, for
fusion proteins with a linker length of 8 or more residues, the apparent molecular weight
was approximately 33 kDa, consistent with a compactly folded monomeric species,
whereas a shorter linker length led to a larger apparent molecular mass of about 66 kDa,

which is consistent with TBP—linked dimers of fusion protein.

98

 

 

 

 

 

 

 

 

 

 

 

 

35: Peak MW
o 1 670 kDa
8; 2 158 kDa
8 3 44 kDa
E, 4 17 kDa
8 5 1.3 kDa
.D
< f
0 20 40 100 120
g Void Unknown
O
8; ~66 kDa
8
S
.D
.D
< l l
0 20 40 100 120
E Void ~33 kDa
O
00
a
d.)
8
<6
.0
J
.D
< - 1
0 20 40 100 120
Elution volume (mL)

Figure 17: Monomeric and dimeric states of VP16C-TAND2-TBP fusions. (A) Predicted
arrangement when the linker is too short to allow intramolecular binding between VP16C-
TAND2 and TBP. (B) Predicted arrangement when the linker is long enough to allow
intramolecular binding. (C) SEC of purified fusion proteins. Top panel, molecular size
standards. Middle panel, fusion protein with 4 amino acid linker. Bottom panel, fusion

protein with 8 amino acid linker.

99

Size-exclusion chromatography effectively separated the intact fusion proteins from a
large fraction of material of high apparent molecular weight, and thus this step was
included in all subsequent purifications (Fig. 18). The protein in the 33 kDa SEC peak,
when recovered, concentrated, and resolved again over the SEC column, eluted almost
exclusively at the same position (data not shown), which indicated that the folded,
apparently monomeric state of the fusion protein was stable over short times. SDS-
PAGE analysis of proteins recovered from subsequent crystallization experiments
indicated overall primary structure stability over long incubations at room temperature,

although some proteolysis was evident in some cases.

As the size exclusion chromatography fraction was of high purity and consistent with
monomeric fusion protein, it was subjected to crystallization trials, using primarily an
under-oil approach with either paraffin oil (to maintain mixture concentrations), or a 1:1
mixture of silicone oil and paraffin oil to allow slow evaporation of water from the drops.
The length of the engineered linker peptide is a crystallization variable, and thus the GS2,
G83, and GS4 fusion proteins were included in crystallization experiments. Initial trials
using commercially-available sparse screens yielded rapidly- forming precipitates under
most conditions, with the notable exception of low salt, high pH conditions. It was
reasoned that these conditions might be necessary to stabilize the fusion protein, and
directed attempts at crystallization were made under these conditions. A diverse series of
attempts to precipitate the fusion protein under high pH and low salt conditions led only

to clear drops or to phase separation.

100

 

 

 

O m
Eagae E
0 “U :7, 1;; > m
3 5 0 0 o o
2 o --°--0 51° LLl LL]
kDa U in Q Q m w
97 .3
66 WA L
a .. GST-VP16C-TAND2-TBP
56 ~—
43 :3
35 "
--- - —VP16C~TAND2-TBP
27 . - —— —GST
1 2 3 4 5 6

Figure 18: Purification of GST-VP] 6C-TAND2-TBP firsion protein. Protein samples were
resolved by SDS-PAGE and stained with Coomassie R-250. Lane 1, cleared lysate. Lane
2, Proteins bound to GSH resin. Lane 3, products of thrombin digestion. Lane 4, proteins
liberated from resin. Liberated proteins were subjected to SEC. Lane 5, SEC void frac-
tion. Lane 6, proteins eluting at approximately 33 kDa.

101

Based on the high apparent solubility of the fusion proteins in low-salt and high pH
conditions, other crystallization methods were employed in efforts to move gently from
stabilizing to precipitating conditions. The fusion proteins were concentrated to 40
mg/mL and small aliquots were dialyzed stepwise into lower pH and higher salt
conditions. In all cases this method produced slight and then increasing turbidity in the
samples, but no crystallization. To attempt crystallization under high concentrations of
protein and precipitating agents, free interface diffusion techniques were employed.
Fusion protein concentrated to 50 mg/mL was allowed to directly contact a variety of
precipitating conditions in sealed capillary tubes. These methods also led to visibly clear
protein samples, but the protein-precipitant mixing may have been limited by the high

viscosity of the precipitating agents.

102

Discussion

Approximately 2500 attempts to crystallize the G82, G83, and GS4 proteins have failed
to produce any observable crystals, suggesting that crystallization of these fusion proteins
is unlikely to succeed. Protein crystallization can fail for a variety of reasons, including
proteolysis, structural instability, intrinsic disorder, or simply a failure to find (or a
complete absence of) successful crystallization conditions. Among these possibilities,
degradation may be the easiest to diagnose and control. However, during the course of
the fusion protein experiments is the preceding section, the integrity of the proteins was
not frequently monitored, thrombin protease was not explicitly removed from the protein
preparations, and protease inhibitors were not typically included. While an SDS-PAGE
analysis indicated that some conditions preserved the primary structure of the proteins
over long term incubation, other conditions showed signs of proteolysis. Therefore some
of the crystallization attempts may have failed due to proteolysis. Any further attempts at

crystallization of these proteins should incorporate safeguards against proteolysis.

Structural instability can also thwart protein crystallization. While the GSn fusion
proteins remained in a folded, apparently monomeric state after purification and
concentration, this analysis came only after a short incubation at low temperatures. The
long-term structural stability of the fusion proteins remains unexplored. However, a
more thorough analysis of the structural stability of a similar VTzTBP crosslinked
complex (chapter 5) shows that salt concentrations below 10 mM are required to preserve

the compact monomeric state for even a few days at 20 °C. In this respect, most of the

103

crystallization attempts in the preceding sections may have failed due to structural

instability.

Additionally, instrinsic disorder can work against protein crystallization. In the handoff
mechanism of activation, the connection of VP16C to TBP is transient, and the transient
nature of the interaction may be reflected in intrinsic disorder of the activation domain.
Indeed, an NMR analysis of the crosslinked complex (chapter 5) indicated that a
significant portion of the VP16C-TAND2 peptide existed in an unstructured state in the
complex. While the distribution of structured and unstructured elements is unknown, it is
possible that the regions of VP16C and TAND2 included in the fusion protein design
encompass a structured core with unstructured flanking regions. In this case, truncations

of either or both domains may promote crystallization.

The engineered GSGS linker is also a likely source for disorder. The minimal length of
this linker is currently determined as 2 units of 4 residues, but the minimal number of
residues may be anywhere from 5 to 8 residues. It is possible that a minimal length linker
may assist crystallization. Furthermore, the stabilizing conditions of low salt and low
temperature may enable the complex to exist stably without the linker peptide, which is
likely to reduce the intrinsic disorder of the complex by exposing more of the well-
structured TBP surface. While the long-term integrity of the VP16C-TAND22TBP
heterodimeric complex under stabilizing conditions remains to be tested, this complex

may provide the best target for any subsequent attempts at crystallization.

104

Chapter V

N MR analysis of the VP16-TAND2-TBP complex

Introduction

The efforts to crystallize a protein complex containing the VP16C activation domain
complexed with TBP (chapter 4) led to the development of a fusion protein comprising
VP16C-TAND2- linker-TBP. This protein could be purified in an apparently compact,
monomeric form, suggesting that it was a suitable candidate for crystallization. Although
all attempts to crystallize this protein failed, it was noted that this protein could be
concentrated to a degree suitable for protein NMR spectroscopy. This chapter describes
an initial structural characterization of the VP16C-TAND2-linker-TBP fusion protein

using NMR spectroscopy.

NMR can provide a set of through-space distance measures between amide protons in
proteins via Nuclear Overhauser Effect spectroscopy (NOESY). NOESY measures span
short distances (typically 5 A or less), and thus are useful for the detection and definition
of secondary structural elements and tertiary contacts. A full set of NOESY distance
measures can be used to constrain tertiary structural models, providing high-resolution
protein structures. NOESY requires an assignment of amide proton resonances to
particular residues in a protein’s primary structure, both to sort the complex spectra and
to apply the resulting information to a structural model. Resonance assignment typically

involves a sequential identification of bonds in the protein backbone, requiring the use of

105

13 C isotopic labeling and complex spectral analysis. Resonance assignment typically

comprises a large fraction of NMR protein structure determination efforts.

The difficulty of protein NMR increases with the size of the target protein, for a few
reasons. With increasing numbers of resonances, the resulting spectra become
increasingly dense and more prone to spectral overlap. Larger proteins also have longer
rotational correlation times in solution, which leads to inhomogeneous sampling of the
NMR magnetic field and thus broadening of the resonance peaks, compounding problems
of spectral overlap in the already dense spectra. Additionally, spectral assignment
utilizes '3 C resonances, which are intrinsically broad, further complicating resonance
assignment efforts. These effects suggested that resonance assignment of the 28 kDa
VP16C-TAND2-linker-TBP fusion protein would be difficult, although probably

tractable.

One technique for simplifying NMR spectra of large proteins is to label only the
segments of interest. Since the structure of TBP has already been determined, specific
labeling of the VP16C-TAND2 portion of the complex would highlight the unknown
regions and greatly simplify the resulting spectra. A drawback to this approach is that
specific contacts between the labeled segments and the unlabeled segments would not be
determined. The following sections describe the application of NMR techniques to study
a cross- linked mimic of the fusion protein that carries isotopic labeling on only the

VP16C-TAND2 segment.

106

Materials and methods

VP] 6C—TANDZ—linker—Cys plasmid construction. Plasmid pDS42~10, encoding a fusion
of GST to VP16C (452-490) and TAND2 (41-73), was cleaved with EcoRI and XhoI and
the plasmid backbone was isolated. Cohesive oligonucleotides ST1033 and ST1034 were
annealed and phosphorylated with T4 polynucleotide kinase, and directionally cloned
into the EcoRI and XhoI sites of the pDS42~10 plasmid backbone, resulting in plasmid
pDSEl37-3. This plasmid encodes VP16C (452-490) followed by a sequence of EcoRI,
Bglll, and Xhol recognition sites, encoding a fusion of Gly- Ala-Asn—Ser-Cys-stop to Gly

490 of VP16C. This plasmid serves as a destination plasmid for subsequent cloning.

To create a family of plasmids encoding the fusion VP16C-TAND2-linker-Cys, EcoRI-
BamHI segments encoding TAND2 fused to a series of Gly-Ser-Gly-Ser repeats were
extracted from the plasmid family pDSElO6-n (chapter 4), and directionally cloned
between the EcoRI and BglII sites of pDSEl37-3, resulting in plasmid family pDSEl38-n.
This plasmid family encodes GST-VP16C-TAND2-linker-Cys, where n denotes the

number of GSGS tetrapeptide units comprising the linker.

GS T ~ VP] 6C~T AND2~1inker~Cys expression. The pDSEl 38-n plasmid family was
transformed into E. coli strain BL21. Cells containing expression plasmids were grown
in LB medium containing 80 mg/L ampicillin. One liter cultures were grown at 37 °C

with vigorous shaking until an OD600 of about 0.7. Expression of the fusion protein was

107

induced by addition of IPTG to a final concentration of 0.5 mM and growth was
continued for an additional 3-5 h while cultures were allowed to approach ambient
temperature (25 °C). Cells were harvested by centrifugation and resuspended in 20 mL
HEMGT-250 (24 mM HEPES, 0.1 mM EDTA, 12.5 mM MgC12, 10% v/v glycerol, 0.1%
v/v Tween-20, pH 7.5) with 250 mM KCl, supplemented with 1 mM PMSF and 5 mM
DTT. Resuspended cells were lysed by 3 passages through a French Press (Aminco)
operating at 20,000 psi. Cellular lysates were centrifuged at 16,000 x g at 4 °C for 25

min. The supernatant was recovered, flash frozen in liquid nitrogen, and stored at ~80 °C.

To obtain VP16C-TAND2-linker- Cys proteins for NMR spectroscopy, the plasmid
pDSEl38-O was transformed into E. coli strain BL21. Cells containing the plasmid were
grown in LB media with 80 mg/L ampicillin. 5 mL of LB-grown culture was inoculated
into 1 L of M9 medium prepared with [ISN] NH4C1 and containing 80 mg/L ampicillin.
M9 cultures were grown to an OD600 of 0.4, and then shifted to 25 °C. Protein
expression was induced by addition of IPTG to a final concentration of 0.5 mM, and
growth was continued for 20 h at 25 °C. Cells were harvested, lysed, and the soluble

lysate fraction was stored as above.

GS T ~ VP16AD plasmid construction. Plasmid pSJTl 193(CRF 1) was cleaved by BamHI
and Bglll to liberate the VP16 activation domain (413-490). This fragment was cloned
into the BamHI site of plasmid pGEX 4T-2 (Amersham), and a forward oriented
construct was selected, resulting in plasmid pDSE17l~ l , encoding the VP16AD as a

thrombin-cleavable fusion to GST under control of the tac promoter.

108

GS T ~ VP16AD protein expression. Plasmid pDSEl 71-1 was transformed into E. coli
strain BL21. Cells containing the plasmid were grown in LB medium containing 80
mg/L ampicillin. One liter cultures were grown at 37 °C with vigorous shaking to an
OD600 of 1.0. Protein expression was induced by addition of IPTG to a final
concentration of 0.5 mM, and growth was continued 1 h, after which cultures were
chilled in an ice-water bath. Cells were harvested by centrifugation and resuspended in
20 mL buffer HEGK~250 (24 mM HEPES, 0.1 mM EDTA, 10% v/v glycerol, pH 7.5)
with 250 mM KCl and supplemented with 1 mM PMSF and 5 mM DTT. Resuspended
cells were lysed by sonication (Branson Sonifier 450, 80% duty cycle, power level 8, 4
cycles at 30 see with 2 min between cycles). Cellular lysates were centrifuged at 16,000
x g at 4 °C for 25 min. The lysate soluble fraction was collected, flash frozen in liquid

nitrogen, and stored at ~80 °C.

T BP expression plasmids. Plasmid trc-TBP, encoding Saccharomyces cerevisiae TBP
(residues 61-240) (yTBPc) as a 7xHis~tagged thrombin-cleavable fusion protein, was
obtained from Dr. J .H. Geiger. Plasmid TEV-TBP, encoding yTBPc as a 7xHis~tagged
TEV-protease cleavable fusion protein in a pET21 backbone, was obtained from Dr. J .H.

Geiger.

To create a plasmid encoding untagged yTBPc, the TBP coding region was PCR

amplified using an upstream primer encoding an NdeI sequence followed by a match to

TBP Ser 61 and beyond, and a downstream primer matching the plasmid backbone. The

109

resulting DNA fragment was tailored by Ndel and XhoI to isolate the TBPc coding
sequence and directionally cloned between the Ndel and Xhol sites of TEV-TBP to create

plasmid pDSEl48-l in a pET21 backbone.

To create a plasmid encoding yTBPc with an engineered N~terminal cysteine, plasmid
pDSEl 17~5a, encoding GST-VP]6C-TAND2-linker-TBP, was cleaved by NcoI and
BamHI and the plasmid backbone (carrying the TBP sequence) was isolated. Cohesive
oligonucleotides D81 and D82 were annealed, phosphorylated with T4 polynucleotide
kinase, and directionally cloned into the plasmid backbone, resulting in plasmid

pDSE 141 ~ 1 , encoding Met-Gly-Cys-Gly-Ser-Gly-yTBP(6l ~240).

To create plasmids encoding yTBP (49-240) and yTBP (40-240), the corresponding
coding sequences were PCR amplified from a clone of full~ length yTBP (waBP, Dr. J.H.
Geiger) using upstream adaptor-primers D83 or D84 and a downstream T7 terminator
primer matching the vector sequence. The resulting amplified DNA sequences were
tailored by Ndel and HinDIII digestion and directionally cloned into the Ndel and
HinDIII sites of TEV- TBP. The resulting plasmids encode untagged yTBP(49-240)

(pDSE162~ 1) and untagged yTBP(40-240) (pDSEl62-5) in a pET21 vector backbone.

T BP expression. His-tagged TBPc and TBPc with the N-terminal engineered cysteine
were expressed as described (Chapter 4). For untagged TBP versions (residues 40-240,

49-240, and 61-240), E. coli strain BL21(DE3)codon plus was transformed with the

expression vectors, and cells containing the plasmids were grown in 50 mL TB (Terrific

110

Broth) medium with 80 mg/L ampicillin and 40 mg/L chloramphenicol at 37 °C with
vigorous shaking for 4-8 h. 7 mL of the starter culture was inoculated into one liter of
TB medium with 40 mg/L ampicillin and growth was continued at 37 °C with vigorous
shaking until an OD600 of 1.7. Protein expression was induced by addition of IPTG to a
final concentration of 0.1 mM and grth was continued at 20 °C for an additional 10 h.
Cells were harvested by centrifugation and cell pellets were resuspended to 25% w/v in
lysis buffer (50 mM sodium phosphate, 20% v/v glycerol, 500 mM NaCl, 1 mM EDTA,

1 mM PMSF, 5 mM DTT, 0.1% Triton X- 100, pH 7.5). Resuspended cells were lysed by
3 passages through a French Press (Aminco) operating at 20,000 psi and cellular lysates
were centrifuged at 16,000 x g at 4 °C for 35 min. The lysate supernatant was collected,

flash frozen in liquid nitrogen, and stored at ~80 °C.

Purification of protein complexes. Protein complexes comprising GST-activation domain
fusions and TBP versions were purified using a mixed lysate approach (chapter 2 and
chapter 4). Briefly, a saturating amount of soluble lysate containing TBP was added to
soluble lysate fractions containing GST-activation domain variants. Lysate mixtures
were incubated on ice for 1 h, and subsequently purified over GSH resin. Thrombin
protease was used to cleave activatorzTBP complexes from immobilized GST. In certain
cases liberated proteins were further purified by size exclusion chromatography (see

below).

C rosslinking of protein complexes for NMR. l,8-Bis~ maleimidotriethyleneglycol

(BM(PEO)3) (Pierce) was added to VP16C-TAND2:TBP complexes at a final

111

concentration of 0.25 mM and samples were incubated on ice for 30 min. Subsequently
2-mercaptoethanol was added to a final concentration of 2 mM to quench unreacted
BM(PEO)3. Crosslinked proteins were separated from the crosslinking reaction buffer

components by gel filtration.

Reduced IgG. Reduced IgG was prepared from a rabbit IgG preparation by addition of
DTT to a final concentration of 50 mM and incubation at 25 °C for 30 min. Reduced IgG
was dialyzed into a storage buffer (20 mM Tris, 2 mM EDTA, 100 mM KCl, pH 7), and

stored at ~80 °C.

Size exclusion chromatography. Preparative and analytical size exclusion
chromatography (SEC) was performed over Superdex S~200 prep grade media in a
Pharmacia XK-series 75/ 100 column using a Pharmacia F PLC system, operating at 10 °C,
with 1 mL/min flow rate and 280 nm absorbance detection of eluted material. SEC

buffers typically comprised 10 mM Tris, 0.1 mM EDTA, and 10 mM KCl, pH 8.

NMR spectroscopy. 15N-IH heteronuclear single quantum correlation (HSQC) spectra
were acquired at 20 °C using a Varian Unity 600 MHz spectrometer equipped with a
triple resonance probe (Max T. Rogers NMR facility, MSU) with the assistance of Dr.

Aizhou Liu.

112

Results

In an effort to determine structural properties of VP16C bound to TBP we have attempted
to characterize a VP16C-TAND2-TBP fusion protein using NMR spectroscopy. An
initial l-dimensional proton NMR spectra obtained from the VP16C-TAND2-linker-TBP
fusion protein indicated resolvable proton peaks (data not shown), which suggested that
the firsion protein was amenable to NMR structural analysis. However, analysis of this
protein faced two different hurdles. First, expression of the fusion protein in M9 minimal
media (required for isotopic labeling of the protein) was poor. Under optimized
conditions, the overall yield of fusion protein was only 0.5 mg per liter of culture,
implying that about 30 liters of culture would be required to produce a single 0.5 mL, 1
mM protein sample for NMR. The cost of 30 liters of [13C] containing M9 media,
required for backbone resonance assignment, was prohibitive. Second, the 28 kDa fusion
protein is near the upper range for successfirl NMR structure determination, implying that
the spectral analysis would be difficult. To circumvent these weaknesses, it was noted
that the expression of the GST- VP16C-TAND2 fusion protein in M9 media was
sufficient to provide for an NMR sample from less than one liter of M9 minimal media.
Additionally, a separate preparation of VP16C-TAND2 and TBP would allow isotopic
labeling of VP16C-TAND2 only, which would result in 76 labeled residues (instead of

256), greatly simplifying the NMR spectral density.

113

To recapitulate the structure and solubility properties of the single-peptide fusion protein,
it was reasoned that a disulfide bridge might be able to connect an engineered cysteine in
VP16C-TAND2 to TBPc. Yeast TBP has two cysteine residues, and one, Cys 78, is
surface exposed. Cys 78 has been used previously to tether TBP in surface plasmon
resonance assays (120). Cys 78 is near the region of TBP that interacts with TAND2
(102), and thus this residue was a good candidate for disulfide bridging to TAND2.

Since the structure of TAND2 bound to TBP is unknown, a cysteine was engineered onto
the C-terminus of a family of VP16C-TAN D2~ linker proteins, where the linker comprises
a number of GSGS tetrapeptide repeats. With this system, a long linker could be used to
explore the cross linking reaction, and subsequently the shortest linker allowing

intermolecular disulfide formation could be selected for further study.

To promote disulfide formation in the complex between VP16C-TAND2-linker-Cys and
TBP, oxidized glutathione (2 mM) was used as a disulfide shuffling reagent. However,
under conditions that promoted disulfide formation in a reduced IgG control reaction,
very little VP16C-TAN D2 was bridged to TBP. Since little disulfide- linked complex
was formed, it was suspected that TBP Cys 78 might not be available for disulfide
formation in the heterodimeric complex. Cys 78 is near the TAND2 binding site on TBP,
and thus TAND2 binding might occlude Cys 78, preventing access by the C-terminal
cysteine of TAN D2. To work around this possibility, an additional cysteine was
engineered onto the N-terminus of TBPc to provide an alternative crosslinking site.
While this Cys-TBPc could be disulfide bridged to VP16C-TAND2-linker-Cys more

readily, the crosslinking reaction failed to go to completion (Fig. 19, left panels).

114

*5
g- GSSG BM(PEO)3

 

 

 

kDa QNVOOS 2va2 Hours

97— ' 3T:

66—1

56/— ._ m «no... — Crosslinked complex
27—

20—r-I-Oi - Cys-TBP

 

 

 

100
250
500
1000

coo

ooc oo

moo In
—-N

 

 

.- -.-.-—-— -g...‘ - Crosslinked complex

20 pi I ' —Cys~TBP

 

 

Figure 19: Crosslinking of VP16C-TAND2-Cys to Cys-TBP. (A) Purified VP16C-
TAND2~Cys and Cys-TBP were incubated for the specified times at 4°C in the presence of
either 2 mM oxidized glutathione (GSSG) or 50 uM BM(PEO)3 and subsequently resolved
by SDS-PAGE and stained with Coomassie R-250. Free VP16C-TAND2 is not resolved
on the gel. (B) VP16C-TAND2-Cys and Cys-TBP were incubated in the presence of the
indicated concentrations of either GSSG or BM(PEO)3. Samples were incubated at 4°C
for 16 h (GSSG) or 8 h (BM(PEO)3) and subsequently processed as in (A).

115

As an alternative to disulfide bridging, the homobifunctional crosslinker BM(PEO)3 was
tested for its ability to crosslink VP16C-TAND2- linker-Cys to Cys-TBP. Over a range of
conditions, BM(PEO)3 cross-linking of the complex was essentially complete, resulting
in primarily a single species (Fig. 19, right panels). The presence of the secondary
crosslinked species may reflect a less-favored connection to Cys 78, which is exposed on
the surface of TBP. Further exploration of BM(PEO)3 crosslinking showed that the
engineered N-terminal Cys on TBP was not required for crosslinking (Fig. 20A), which
suggested that Cys 78 of TBP was the point of attachment to the crosslinker. It was also
determined that even the shortest GSGS linker supported crosslinking (Fig. 20A).
Additionally, the molecular weight ladder seen across the cross~ linked species provided
an unambiguous measure of VP16C-TAND2-linker-Cys incorporation into the

crosslinked complex (Fig. 20A).

When the resulting crosslinked complexes were resolved by size-exclusion
chromatography (SEC), they behaved similarly to single-peptide fusion proteins (Chapter
4), with roughly half of the material eluting in the void fraction and the other half eluting
at a low molecular weight position corresponding to a compactly folded monomer (data

not shown). SEC was included as a final. step in purification of the complex (Fig. 20B).

To determine an appropriate buffer system for NMR analysis, the purified and
concentrated cross-linked complexes were dialyzed into a variety of conditions and
visually monitored over several days. Using this approach, it was determined that Tris

buffer was preferred over phosphate or glycine, higher pH values were better, and low

116

A Input Crosslinked
0123456 0123456 Linkerlength

 

 

 

 

kDa
56—
a-ca—II. —VP16C-TAND2-TBP
27“ - _
2 .. 1': .. .. _/6xHrs-TBP
0_ - " TBP
3
‘6 a.)
B E 3 E z 8 '8.
[-4 S '35 m is“ E
>. 3‘ "g a 1;; a .3
S E 8 .839 .“2’0 <3 8..
kD (D E-‘ m C: Q 0 <1:
a ~a

 

—/ GST-VP16C-TAND2
_/ VP 1 6C-TAND2-TBP

 

 

d
— GST
g — TBP
—- VP16C—TAND2
7

 

Figure 20: Crosslinking and purification of VP16C-TAND2-TBP complexes. Protein
samples were resolved by SDS-PAGE and stained with Coomassie R-250. (A) Analysis of
crosslinking reactions. Complexes of VP16C-TAND2-n-Cys and TBP were purified and
treated with 250 uM BM(PEO)3 for 45 min at 0 °C. The 6xHis-tag on TBP was partially
cleaved in this experiment, resulting in a pair of TBP bands. VP16C-TAND2-n-Cys is
present, although it stained poorly with Coomassie R-250. (B) Preparation of the cross-
linked complex. Lanes 1-3, lysates containing both proteins were mixed and purified over
GSH resin. Lanes 4-5, thrombin digestion products and liberated proteins. Lane 6, prod-
ucts of crosslinking with BM(PEO)3. Lane 7, crosslinked complex after size-exclusion
purification and 5 day incubation at 20 °C.

117

salt was required for sample clarity. Using this information, an initial NMR sample

buffer was defined (10 mM Tris, 0.1 mM EDTA, 10 mM KCl, 8 mM NaN3, pH 8.5).

To perform an initial structural analysis of the complex, a sample of cross—linked
complex was prepared with 15 N labeling of the VP16C-TAND2 segment, the complex
was concentrated to 0.64 mM, and 2-dimensional 1H-15 N HSQC spectra were acquired
from a freshly-prepared sample (Fig. 21). The spectra showed low overall dispersion in
the proton dimension, suggesting the presence of some, but not many, structured
elements in VP16C-TAN D2 segment. Non-uniform peak intensities indicated different
peptide backbone dynamics at different positions, suggesting a loose or incompletely
folded structure. The combination of peak broadness and overall low dispersion
indicated that resonance assignment would be difficult. Additionally, about 10 of the
expected 70 peaks were very weak, suggesting that they might not be able to be assigned

to particular residues in VP16C-TAND2.

To asses the stability of the NMR sample, it was incubated at 20 °C for 5 days, and the 2-
dimensional HSQC spectrum was acquired again for comparison (Fig. 22). The spectra
from the aged sample showed a loss of the broad, shifted peaks, and the acquisition of
sharp, poorly dispersed peaks, indicating a loss of the structural elements seen in the fresh
sample and a transition to an unfolded state. An SDS-PAGE analysis of the aged sample
indicated that the primary structure of the complex was largely intact (Fig. 20B, lane 7).
The aged sample was observed to scatter light, suggesting that larger particulates had

formed, but attempts to sediment the sample at 16,000 x g did not diminish the protein

118

 

 

 

 

 

107 - 0
O
(D
0 _ 0
112 - 0 I}. ‘ 03
c
0
117 - A .3
. o. J
(”‘5
122 - \ O i“
(I ,3 9
G
0 pix,
% u
127 -
9 8 7

Figure 21: HSQC spectrum of fresh sample at 10 mM KCl. 15N-labeled VP16C-
TAND2 crosslinked to unlabeled TBP. Horizontal axis: proton chemical shift.

Vertical axis: 15 N chemical shift.

119

 

107 -

112 - ég 3%

117 P-

I)

a
o .

1.

122 - - V... ,
a

127 -

 

l l I

 

 

9 8 7

Figure 22: HSQC spectrum of aged sample at 10 mM KCl. 15N-labeled VP16C-
TAND2 crosslinked to unlabeled TBP. Horizontal axis: proton chemical shift.
Vertical axis: 15N chemical shift.

120

concentration. The aged sample was resolved by size exclusion chromatography, and
about 75% of the protein was found to elute in the void fraction, indicating that most the
apparently compact, monomeric prepared complex had decayed into either an unfolded or
aggregated state (data not shown). The remaining 25% of the protein eluted at a position
corresponding to TBP- linked dimers (Chapter 4), which indicated that the prepared
monomeric complex had evolved into dimeric complexes. These results suggested that
the sample may have followed a decomposition pathway involving a transition from a

monomeric to a dimeric state, followed by subsequent unfolding or aggregation.

In attempts to stabilize the complex, the NMR buffer system was optimized. To assess
sample stability, concentrated cross-linked protein complexes were dialyzed into different
buffers and the protein stability was monitored by visual inspection of precipitation and
turbidity. An NMR buffer successfully used for analysis of a similar complex of TBP
and TANDl-TAND2 (20 mM Tris-HCl, 150 mM KCl, 10 mM MgC12, 5% glycerol, 0.5
mM AEBSF, 10 mM DTT, 0.05 mM NaN3, pH 7.5) (102) was tested and found to not
support stability of the VP16C-TAND2~ TBP cross-linked complex. A family of buffers
rooted on the precedent buffer showed that both lower salt and lower pH significantly
improved sample stability. Additionally, removal of MgCIZ greatly improved the
stability of the sample. As the buffer optimization converged on improved conditions,
the samples became visually indistinguishable, and thus size exclusion chromatography
was employed to resolve the apparent molecular weight distributions of the complexes
after incubation in different buffers. This analysis indicated that lowering the KCl from

40 mM to 10 mM led to significant improvements in maintenance of the compact,

121

monomeric state, and lowering the pH slightly also improved stability (Fig. 23A-C).
However, at this stage of refinement, the complexes were not fully stable over the course

of several days, precluding a practical NMR analysis.

In further attempts to stabilize the complex, adjustments to the protein design were
considered. It was noted that two successful NMR structural analyses of TBP-containing
complexes (99, 102) incorporated more of the N-terminus of TBP. Accordingly, the
previous version of TBP (residues 61-240) was substituted with TBP (residues 40-240)
and TBP (residues 49-240). SEC stability analysis of crosslinked complexes built with
the longer versions of TBP indicated an improved maintenance of the compact,
monomeric state, again with 10 mM KCl preferred over 40 mM (Fig. 23D,E). The
results from TBP (residues 40-240) and TBP (residues 49~240) were indistinguishable,
and thus TBP (residues 49-240) was utilized for further optimization. Since buffers with
10 mM KCl were significantly better than buffers with 40 mM KCl, buffers lacking KCl
were tested, and found to be optimal (Fig. 23F). These results led to a final NMR buffer

definition of 1 mM Tris, 0.1 mM EDTA, 0.02 mM KCl, 0.05 mM NaN3, pH 8.

A 2-dimensional HSQC spectrum was obtained from a cross-linked complex comprising
TBP(residues 49-240)and 15N-labeled VP16C-TAND2 (Fig. 24). A second spectrum
acquired after the sample had been incubated for 2.5 days at 20 °C was essentially the
same, indicating that the sample was structurally stable. A comparison to previous
spectra (Figs. 21,22; Appendix A) showed new peaks dispersed from random-coil

positions, indicating that the new conditions led to more structured elements. However,

122

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

A B C
ii /Void
3 31 $1
3' E e
8 8 8
f3 33 kDa '53 '2
Time Time Time
D E F
8 8 3
$1 51 5
-E 'E '8
8 8 8
in” '<° '2
Time Time Time
Incubation conditions
Panel TBP Buffer pH [KCl]
A 61-240 20 mM glycine 9.3 40 mM
B 61-240 20 mM glycine 9.3 10 mM
C 61-240 20 mM glycine 8.8 10 mM
D 49-240 20 mM glycine 9.0 40 mM
E 49-240 20 mM glycine 9.0 10 mM
F 49-240 1 mM Tris 8.0 0 mM

Figure 23: Optimization of buffer conditions. Samples of VP16C-TAND2-TBP cross-
linked complex were purified, concentrated to approximately 1 mM total protein, and
dialyzed into the indicated buffers. After 2.5 days of incubation at 20 °C, samples were
resolved by size exclusion chromatography. The intact complex eluted at approximately

33 kDa.

123

 

 

 

 

 

G... *>
6
a 90
3%
“ 9
C8
6
9 8 7

Figure 24: HSQC spectrum of crosslinked complex in low salt conditions. 15 N-
labeled VP16C-TAND2 crosslinked to unlabeled TBP. Horizontal axis: proton
chemical shift. Vertical axis: 5N chemical shift.

124

the spectral quality was still low, with overall low dispersion, peaks of varied intensity,
and spectral overlap. Based on the predicted difficulties in resonance assignment and
structure determination, the pursuit of the NMR structure of the complex was not

attempted.

A tabulation of HSQC peaks in the spectra acquired from the different samples indicated
the existence of two distinct classes of structural elements (Appendix A). Among the
shifted peaks in the stable (no salt) sample, a subset was still present in the folded but
unstable sample (10 mM KCl). This result suggested that a core set of structural
elements existed in the VP16C-TAND2 segment at 10 mM KCl, and that an extended set
of structural elements was acquired at the lower-salt condition. The core set comprised
about one third of the total HSQC peaks, and the extended set comprised another third of
the peaks. While the sensitivity of the complex to salt suggested that the extended set of
structural features arose from the lower salt concentration, it must be noted that stable
sample conditions also differed by 0.5 pH units and the inclusion of 12 more N-terminal
residues of TBP. Therefore, the extended set of structural features may also have arisen

from pH dependence or interactions with TBP residues 49-61.

Since the NMR data obtained from the stable VP16C-TAND2-TBP cross- linked complex
indicated few well-structured elements, different alterations to the structure of the
complex were considered. It was recognized that the fusion of VP16C to TAND2 may
have imposed unnatural constraints on VP16C. It was also possible that the fusion to

TAND2 tethered VP16C in an unnatural orientation relative to TBP, which could have

125

led to binding imperfections and structural defects. As VP16C is normally attached to
VP16N, an inclusion of both VP16N and VP16C was considered as an alternative to the
VP16C—TAND2 fusion. VP16N has been shown to interact with TBP, and mutations that
compromise VP16N transcriptional activation function also compromise the association
with TBP (56). Additionally, although VP16N does not carry a TFIID-TFIIA-DNA
assembly function, it does appear to contribute somewhat to this function when coupled
with VP16C (76), suggesting that VP16N doesn’t interfere with VP16C-TBP interactions.
Taken together, these data suggested that VP16N may help VP16C interact with TBP,
and a resulting complex might be more likely to preserve the native VP16C-TBP

interface.

GST-VP16NC was tested as a substitute for GST-VP16C-TAND2 in the co-purification
of TBP. GST-VP16NC retained TBP through binding and washing steps, allowing the
purification of the immobilized complex (Fig. 25A). Thrombin digestion was used to
liberate VP16NC and TBP from immobilized GST and the resulting proteins were further
purified using size exclusion chromatography (Fig. 25B). The SEC peak fraction
comprising VP16NC and TBP was concentrated and subjected to a second round of SEC,
which resolved two peaks (Fig. 26A, first panel). This result indicated either that a
complex of VP16NC and TBP was decomposing into separate proteins or that VP16NC
and TBP existed as separate proteins that were not well resolved in the preparative SEC

step.

126

Cleared lysate
TBP copurification

Purified

-I Uninduced

       

 

 

 

 

 

'U
8
=1
'0
.5
97 =
66
—GST—VP16NC
27
—TBP
20
1 2 3 4 5
B
35 8 ~ N m
S E E 8.5 .5 S
8 a .8 e 8 8 ‘8
E 8 '8 8 “E “b “b
=’ .29 .29 .29 U U 0
63 a Q Q a 5;” 51
u .- —GST-VP16NC
Ila-’0 — -GST
---——-~-- —TBP
H...- —VP16NC
1 2 3 4 5 6 7

Figure 25: Co-purification of VP16NC and TBP. Protein samples were resolved by SDS-
PAGE and stained with Coomassie R-250. (A) Lane 1-3, expression and solubility of
GST—VP16NC. Lane 4, purification of GST—VP16NC over GSH resin. In lane 5, an E. coli
lysate containing TBP was added to the lysate containing GST-VP16NC, and TBP was
copurified. (B) Purification of VP16NC and TBP. Lane 1, lysates containing GST-
VP16NC and TBP were mixed, incubated, and purified over glutathione resin. Lanes 2-3,
thrombin digestion products and liberated proteins. Lanes 5-7, size exclusion chromatog-
raphy fractions.

127

 

   

 

 

3 hours I)? 2 13 hours 33 hours 3.5 days
T Peak 1
E
I:
3
8
8
8
E
8
.D
< Void
Tim—e —>

 

 

 

 

Input
Peak 1
Peak 2

 

..w- —TBP

u ' a —VP16NC

 

 

 

Figure 26: Analysis of the stability of a mixture of VP16NC and TBP. (A) A preparation
of VP16NC and TBP was concentrated and incubated at 20°C for the indicated times, and
then resolved by size exclusion chromatography. The protein mixture resolved into two
peaks, one of which diminished over time. (B) Peak 1 and Peak 2 were isolated from (A),
resolved by SDS-PAGE, and stained with Coomassie R-250.

128

i To analyze the stability of the purified VP16NC and TBP proteins, purified samples were
swapped into different buffers, incubated for 3.5 days at 20 °C, and resolved using SEC.
Interestingly, all samples showed a dramatic loss of one of the two peaks and a gain in
the void fraction, indicating that one of the two proteins was unstable. Additionally, low-
salt conditions could not stabilize the mixture. SDS-PAGE analysis of the incubated
samples showed that the primary structure of both VP16NC and TBP was intact, ruling
out proteolysis and implying that either unfolding or aggregation events were affecting
one of the proteins. An SEC analysis of samples as a function of time indicated that one
of the peaks was stable, and that the other diminished significantly over a period of hours
(Fig. 26A). The peaks were isolated and identified by SDS-PAGE (Fig. 263), indicating

that TBP was the unstable protein.

129

Discussion

Difficulties in the preparation of a fully- labeled VP16C—TAND2- linker-TBP complex,
coupled with potential difficulties in NMR analysis of such a large protein, have led to
the exploration of a segmentally- labeled, cross- linked complex. Conditions of very low
salt concentration were found to stabilize this complex on time scales sufficient for NMR
data acquisition. However, the peaks in the resulting HSQC spectra were poorly
dispersed from random-coil positions, suggesting few well-structured elements in the
VP16C-TAND2 segment. Additionally, variable peak intensities indicated different
peptide backbone dynamics in different regions, suggesting that different regions
underwent conformational changes on different time scales. This result is not surprising,
particularly since TAN D2 and VP16C are separate domains, and as such may fold
independently. Overall, the combination of poor dispersion, broad peaks, and spectral
overlap implied that the sequential assignment of the stabilized, cross-linked complex

using 13c labeling would be rather difficult and likely incomplete.

A comparison of spectra acquired from different states of the complex indicated two
different classes of structure in VP16C-TAND2. Apparently, at low salt a core set of
structural elements was present, and at even lower salt an extended set of structural
elements formed. However, in the absence of resonance assignment, the distribution of

these structural elements in VP16C and TAND2 cannot be determined.

130

The well-resolved spectrum from the unstructured state of VP16C-TAND2 is amenable
to resonance assignment, which suggests an alternative approach to structural analysis. If
the resonance peaks in the unstructured state were assigned to particular residues, then an
accounting of peaks depleted in the structured state would identify which residues were
involved in structural elements. Information obtained in this way could then be
compared to data derived from other studies, to help determine if conserved sets of

residues are involved in structural elements when VP16C interacts with different targets.

Indeed, there are signs that the regions of VP16AD that interact with TBP differ from
those that interact with other targets. Time resolved fluorescence anisotropy and
fluorescence quenching experiments have shown that VP16AD residues Phe 442 and Phe
473 become protected from solvent and conformationally constrained upon TBP binding
(153). However, NMR studies have shown that the chemical shift of these residues
changes little upon TAF 9 binding (166), or TFIIB or PC4 binding (61). Additionally,
surface plasmon resonance studies have identified VP16C residues critical for TBP
interaction (120), but again these residues are only modestly shifted in the TAF 9, PC4,
and TFIIB NMR studies. These comparisons suggest that VP16 may fold in different
ways when interacting with different partners. In this light, a determination of the
VP16C residues involved in TBP interactions, as could be determined by assignment of
the unstructured state of the VP16C-TAND-TBP complex, would provide valuable data

for comparison to the structured regions involved in interactions with other targets.

131

Chapter VI

Discussion and Prospects for Further Study

Crystallization of a complex of VP16C and TBP

The majority of the research efforts discussed in this dissertation were aimed at
determining the structure of the VP16C activation domain when bound to TBP. Since a
crystal structure of the VP16C:TBP complex would provide the highest resolution model
of VP16C, initial efforts were directed at crystallization of the complex. The apparent
crystallization of TBP dimers in initial crystallization attempts led to the reasoning that a
tighter complex was needed, and the GST-VP16C-TAND2 fusion protein was adopted
for subsequent work. The strong association of this fusion protein with TBP allowed for
the co-purification of GVT and TBP directly from mixed bacterial lysates, which
simplified the purification scheme and facilitated a series of crystallization attempts.
However, even this tight complex was found to eventually release TBP. In attempts to
further tighten the complex, a fusion of VP16C-TAND2-linker—TBP was created and

characterized. However, no crystals were obtained from this fusion protein.

The continued failure to obtain crystals of a protein complex representing the
VP16C:TBP interface led us away from attempts at crystallization, and subsequent efforts
were focused on NMR approaches (chapter 5). Importantly, in the efforts to stabilize a

crosslinked variation of the VP16C-TAND2-TBP fusion protein, it was discovered that

132

the protein was unstable under many conditions, and that low salt conditions were
required for the maintenance of a folded and monodisperse state. This finding implied
that most of the crystallization work described in chapter 4 was performed with unstable

protein.

The protein stability characteristics determined in chapter 5 could direct another series of
crystallization experiments, with a focus on conditions supporting protein stability.
Additionally, the NMR results from chapter 5 indicated that the VP16C-TAN D2 peptide
was not uniquely structured, but instead was somewhat disordered. The degree of this
disorder may be temperature dependent. Furthermore, it has been noted that reduced
temperature stabilizes the association of VP16C-TAND2 and TBP (figure C4).
Therefore, crystallization attempts employing low salt and low temperatures would focus

on the conditions most likely to provide overall stability and reduced structural disorder.

Additionally, the VP16C-TAND2- linker-TBP fusion protein may not represent the best
target for continued crystallization trials. The glycine- serine linker peptide probably
exists as a poorly structured loop, and as such may not provide for structured crystal
lattice contacts. In the heterodimeric VP16C-TAND22TBP complex this loop is absent
and additional structured surfaces of TBP are exposed, suggesting that the heterodimeric
complex is a better candidate for subsequent crystallization trails. If this complex is used
for trials, it should first be determined whether the heterodimer can be stabilized on time

scales suitable for crystallization.

133

Truncations of the VP16C and TAN D2 peptide segments may remove unstructured
termini and increase the likelihood of crystallization. It has been reported that a shorter
version of VP16C (residues 470 to 490) was as effective as VP16C (residues 457-490) in
supporting the tight binding of VP16C-TAND2 to TBP (83). Since VP16C residues 450~
470 are not required for the TBP interaction, they may be present as an unstructured
peptide segment, which may work against crystallization of the complex. C-terminal
truncations of TAND2 should also be considered. Although TAND2 residues 42 to 73
have been included in our version of VP16C-TAND2, shorter versions of TAND2 may
still support high affinity interactions with TBP. In this respect, it should be noted that a
shorter TAND2 segment (residues 42-64) is sufficient for high—affinity interactions with

TBP in the context of TANDl-TAND2 (79).

In the event that crystals of VP16C-TAND2zTBP complexes cannot be obtained, the
tightly associated complex may provide for indirect approaches to structure
determination. For example, one might search for pairs of cysteine residues spanning the
VP16C-TBP interface that could participate in disulfide bonds. A panel of cysteine
substitutions in VP16C could be tested with a panel of cysteine substitutions in TBP.
Disulfide formation might indicate specific contacts between VP16C and TBP.
Additionally, the resulting cross- linked species might be suitable targets for subsequent
crystallization trials. A second possible approach is to tether a proteolytic agent such as
FeBABE (111) to specific residues of either VP16C or TBP, and then map the resulting

cleavage patterns to establish proximity of VP16C residues to TBP residues. However,

134

the F eBABE approach might be limited by the diffuse cleavage patterns generated by the

diffiisible hydroxyl radical.

N MR analysis of VP16C-TAND2 bound to TBP

The VP16C-TAND2- linker-TBP fusion protein developed in chapter 4 was found to be
soluble enough for NMR analysis. To reduce the cost of the protein and the complexity
of the spectra, a segmental labeling system was developed, which allowed selective
isotopic labeling of VP16C-TAND2. Initial NMR spectra indicated that the complex had
structural elements but was not stable on the several-day time scale required for NMR
analysis. Optimization of the buffer conditions and protein design resulted in a protein
complex stable enough for NMR analysis. In the resulting NMR spectra, poor dispersion
in the proton dimension indicated a low overall degree of structure, overall broad peaks
indicated conformational variability, and different peak intensities indicated different
backbone dynamics in different regions. Additionally, the broad peaks and variable peak
intensities suggested that assignment of the resonances to particular residues would be

difficult and likely incomplete.

A comparison of the NMR spectra from structured and denatured complexes indicated
that two different classes of structural elements existed, one present both in 10 mM KCl
and in a no-salt buffer, and a second class present only in a no-salt buffer. It has not been

determined how the residues of VP16C-TAND2 are distributed in these two structural

135

classes. However, since the NMR spectrum of the denatured state of VP16C-TAND2 is
well resolved and thus amenable to resonance assignment, it would be possible to assign
the resonances in the unstructured state, and then determine which peaks are absent in the
different structured states. Although this approach would not provide details of
secondary or tertiary structure, it would delineate the regions of VP16C that are involved
in structural elements when bound to TBP. This information could be compared with
data from other NMR analyses (61 , 166) to determine if conserved regions of VP16C are

involved in structural interactions with different binding partners.

TBP orientation

The report that the VP16 activation domain could enforce a productive orientation on an
otherwise poorly oriented TBP:DNA complex was intriguing. We were interested in
whether the TBP-orienting activity was related to transcriptional activation functions. To
address this relationship, we had intended to utilize a panel of VP16AD variants to
measure the correlation between in vitro orienting function and in vivo activation
function. Additionally, the TBP orienting function implied a ternary interaction between
VP16AD, TBP and DNA, but a body of other evidence argued against such a ternary

interaction. We sought to explore this apparent contradiction.

Unexpectedly, and contrary to the precedent reports, we found that TBP could bind to the

TATA sequence with a high degree of orientational specificity. Clear and internally-

136

consistent data from both DNA strands and a series of control experiments using reversed
and symmetrical TATA sequences validated our experimental methods and indicated that
the eight basepair TATA sequence was sufficient to direct highly oriented binding of

TBP to DNA.

The finding that TBP can recognize the TATA sequence with a high degree of
orientational specificity implies that current models of PIC assembly need to be
rethought. In the current models, the ability of TBP to recognize the orientation of the
TATA sequence is assumed to be modest, and other factors, such as TFIIA and TFIIB,
are required to determine transcriptional directionality (50, 156). Our findings show that,
to the contrary, the TBPzTATA connection might define transcriptional directionality in
some cases, and thus the orientation of the TATA sequence may be a more prominent

variable in core promoter structure than currently assumed.

Since other core promoter elements also can direct transcriptional orientation, questions
about the interplay among directional signals arise. For example, if a forward orienting
TFIIB recognition element (BRE) flanks a reverse-orienting TATA sequence, is a TBP-
TFIIB-DNA complex significantly less likely to form? To what extent will it be
oriented? Also, to what extent do the directional signals recognized by TFIID need to
cooperate? For example, given an initiator sequence (Inr) and a downstream promoter
element (DPE), how much will the orientation of the TATA sequence influence TFIID

binding and promoter strength? One approach toward answering these questions is to

137

 

monitor PIC assembly on various core promoters in vitro using reconstituted systems

(16).

In the interpretation of directional signals by the transcriptional machinery the order of
events might be an important factor. For instance, consider TFIID recognition of a
hypothetical promoter comprising forward-orienting Inr and DPE elements and a
reversed TATA sequence. If TFIID were to recognize the Inr and DPE elements of this
promoter before recognizing the TATA sequence, a TFIID:DNA complex might form in
the forward- facing direction. Conversely, if the TBPzTATA interaction mediated the
initial TFIID:DNA contacts, TFIID might bind the promoter in a reversed orientation. In
this case, the reversed orientation of TFIID might prevent the recognition of the Inr and
DPE sequences, leading to several possible outcomes, including stalled PIC assembly,
reduced PIC stability, or even a fully functional PIC with a reverse orientation. Since
core promoter elements may be dynamically masked and exposed by chromatin dynamics
in vivo, it is possible that the ordered recognition of core promoter elements is a variable

and possibly regulated process in transcription.

In our studies of TBP orientation we also determined that Gal4~VPl6 did not noticeably
alter the orientation of TPB on the AdMLP promoter sequence. This result is congruent
with the observations that VP16AD, TBP, and DNA do not form the ternary connections
that would be required for VP16AD to orient TBP on DNA. However, it remains
possible that our studies did not detect increased TBP orientation because TBP was

already maximally oriented. In this respect, it would be of value to explore the effects of

138

Gal4~VPl6 on TBP orientation using the cleavage probes with reversed and symmetrical
TATA sequences. This would allow for a larger analytical window in which to detect
changes in TBP orientational specificity. Also, it should be noted that there are two
distinct ways an activator might alter TBP orientation: by orienting TBP relative to the
position of the activator, or by altering the intrinsic TATA-recognition properties of TBP.
1n the latter case, TATA sequences that can only partially direct TBP will be required to

determine whether an activator can alter the intrinsic DNA recognition properties of TBP.

TBP purification

The high affinity interaction between VP16C-TAND2 and TBP has enabled the
development of a novel method of TBP purification (chapter 2). This method is very fast
and easy to perform, requires no extrinsic affinity tags, and the resulting TBP appears
pure, structured, monodisperse, and firnctional. So far, this method has been applied to
the purification of full-length and N-terminally truncated versions of yeast TBP.
However, since both VP16C and TAND2 interact with the highly conserved core region
of TBP, it is likely that the purification method will also work with TBP from other
species. Indeed, preliminary data (not shown) suggests that this method can be used to
purify the conserved core of human TBP. Tests with TBP from other species are

planned.

139

 

Appendix A

HSQC spectra of VP16C-TAND2 complexed with TBP

140

 

Table 2: Intensity of NMR peaks in different spectra

 

 

 

HSQC peak intensitya
Peak A B C Peak A B C Peak A B C Peak A B C
1 3 3 0 26 3 3 1 51 3 3 1 76 3 0 l
2 3 3 1 27 3 3 3 52 0 3 0 77 2 0 1
3 3 3 3 28 3 l 1 53 3 3 2 78 l 0 1
4 3 3 3 29 3 3 1 54 1 3 3 79 2 O 2
5 3 3 0 3O 3 3 1 55 l 3 3 80 0 0 3
6 3 3 1 31 3 3 l. 56 3 2 l 81 0 0 l
7 3 3 1 32 3 3 1 57 3 3 3 82 0 0 3
8 3 2 l 33 3 3 1 58 3 3 3 83 0 0 1
9 3 3 l 34 3 3 l 59 3 3 1 84 0 0 3
10 3 3 .1 35 3 2 2 60 2 2 3 85 O 0 3
11 3 3 1 36 3 3 1 61 0 3 3 86 0 0 3
12 0 3 3 37 3 ? ? 62 0 1 3 87 0 0 3
13 2 2 l 38 3 ? ? 63 1 2 1 88 0 0 3
14 l 2 3 39 3 3 2 64 0 2 1 89 0 0 3
15 3 3 1 40 3 3 l 65 0 2 l 90 0 0 3
16 1 2 3 41 1 3 3 66 0 2 3 91 0 1 2
17 0 1 0 42 3 1 1 67 1 l 3 92 0 O 3
18 3 l 2 43 3 2 1 68 3 2 3 93 0 O 3
19 2 1 0 44 3 3 0 69 3 1 2 94 0 O l
20 1 3 3 45 0 1 3 70 3 l 0 95 0 0 3
21 0 1 0 46 0 3 3 71 3 0 0 96 2 l 2
22 0 3 3 47 3 3 1 72 1 0 0 97 3 0 0
23 0 3 3 48 2 3 3 73 1 0 0 98 3 l 1
24 3 3 1 49 3 3 2 74 2 0 0 99 0 2 3
25 3 3 1 50 3 3 1 75 3 1 1 100 2 1 l

 

A: Intensity of peak in HSQC spectrum of denatured sample at 10 mM KCl (Fig. 27).
B: Intensity of peak in HSQC spectrum of folded sample at 10 mM KCl (Fig. 28).
C: Intensity of peak in HSQC spectrum of folded sample at 0 mM KCl (Fig. 29).

a Intensity estimated by visual inspection of HSQC plots (Figs 27-29) placed on

a scale from 0 (peak not present) to 3 (most intense).

141

 

Table 3: Comparison of NMR spectra

 

. . . . . . a
Comparison of NMR peak mtensrtles ln dlfferent spectra

 

 

 

10 mM KCl structured relative Low salt structured relative
to 10 mM KCl unstructured to 10 mM KCl unstructured
Strong Stronger Weak Strong Stronger Weak
new peaks common peaks new peaks new peaks common peaks new peaks
12 14 17 12 85 14 64
22 16 21 22 86 16 65
23 20 62 23 87 20 81
46 41 64 45 88 41 83
52 48 65 46 89 48 91
61 54 66 61 90 54 94
55 91 62 92 55
63 99 66 93 60
80 95 67
82 99
84

 

a
See Table 2 for tabulation of NMR peak intensities, and Figures 27-29 for peak locations.

142

 

 

107 -

0 e- 79
a 96
59
«978
58 £377

a: l
112 - B
l?

 

 

 

 

 

57
97. 063
976
056
55:11-
117 '-
41
122 -
127 -
9 8 7

Figure 27: HSQC spectrum of aged sample at 10 mM KCl. 15N-labeled VP16C-
TANDZ crosslinked to unlabeled TBP. Horizontal axis: proton chemical shift.
Vertical axis: 15N chemical shift.

143

107

112

117

122

127

 

 

   

6,?
_ f3
91
1’" 153

 

 

4’ 54 64
100 53 52 g t?

47 .
4;_ K5137,”

 

 

 

1 ”’s‘QH 50
‘ _ 49 36
' «4'27 5" 33
:ro.::g; 32°25 35
e 68 921 3 22
.‘w‘tfi “F" O 84
5 .3. '" 19
9.17.6, 189 O 75
0: 4‘. 8 20
7 6
5
3 a 4 r
(:3) 67
2 °\”
70
G 1 069
1 l L
9 8 7

Figure 28: HSQC spectrum of fresh sample at 10 mM KCl. 15N-labeled VP16C-
TAND2 crosslinked to unlabeled TBP. Horizontal axis: proton chemical shift.
Vertical axis: 15 N chemical shift.

144

 

 

 

 

 

107 -
112 -
3 " o
-
117 -
99
45
4146
122 5’3 066
- 27
29 / 022
31 2614
127 L

 

9

 

7

 

 

Figure 29: HSQC spectrum of crosslinked complex in low salt conditions. 15 N-
labeled VP16C- TAND2 crossllisnked to unlabeled TBP. Horizontal axis: proton
chemical shift. Vertical axis: 15N chemical shift.

145

Appendix B

Plasmid maps and oligonucleotide sequence

146

 

Table 4: Oligonucleotide sequences

 

 

Name Sequence Matches Direction
DSl CATGGGTTGCGGCTCTG F
DSZ GATCCAGAGCCGCAACC R
DS3 GCCCCGCCATATGGCTGCCCCAGAATCTG yTBP 49-54 F
DS4 GCGCCGCCATATGTTCCAGAGTGAAGAGG yTBP 40-45 F
ST394 see Triezenberg lab folder pET28a downstream R
ST803 GCGGGATCCCCGGGTCCGG VP16 451-456 F
ST804 ATGAATTCGCACCCCCACCGTACTCGTCAATTCCA VP16 483-490 R
ST805 GGCGGAATTCAAAGGACTATACGGAGCAT yTAFl 43-50 F
ST806 TATCTCGAGCTATCATTCTTCTGGCAAATCGTCATCGTC yTAF 1 66-73 R
ST903 TCATACACATACGATTTAGGTGACA pGSMLT upstream F
ST904 GAAGAGGAGAAGATAATAGGAGGAA pGSMLT downstream R
ST903—6FAM TCATACACATACGAT’I‘I‘AGGTGACA pGSMLT upstream F
ST904-HEX AGAAGAGGAGAAGATAATAGGAGGAA pGSMLT downstream R
ST906 GGATACCATGGGTTCCCCGGGTCCGG VP16 451-456 F
ST907 CGCGCGGATCCTTCTTCTGGCAAATCGTCATC yTAFl 67-73 R
ST908 GATCAGGCTCTG F
ST909 GATCCAGAGCCT R
ST925 GATCTAACCTGCACCCCAAAG yTBP 75-82 R
ST933 GCCCCGCCATATGTCCGGTATTGTTCCA yTBP 61-65 F
ST1033 AA'I‘TCCTGTTAGAGATCTTGCTGATAGC F
ST] 034 TCGAGCTATCAGCAAGATCTCTAACAGG R
STl 130 CCTGAAGGGGGGCTATATATAGGGGTGGGGGCGCG AdMLP TATA F
ST] 131 CGCGCCCCCACCCCTATATATAGCCCCCCT’I‘CAGG AdMLP TATA R
81‘] 132 CCTGAAGGGGGGCCTTTTATAGGGGTGGGGGCGCG AdMLP TATA F
ST! 133 CGCGCCCCCACCCCTATAAAAGGCCCCCCTTCAGG AdMLP TATA R

 

147

 

  
    

30 BamHl (1)
1055 EcoRI (1)
1155 XhoI (1)

AmpR 1776...2435
2119 Pstl (1)

Figure 30: Plasmid pDS42-10.
Parent plasmid: pGEX4T—l (GE Healthcare)

Inserted sequences: VP16C between BamHI-EcoRI
TAND2 between EcoRI and XhoI

BamHI
l
CTGGTTCCGCGTGGATCCCCGGGTCCGGGATTTACCCCCCACGACTCCGCCCCCTACGGC
L V P R G S P G P G F T P H D S A P Y G

GCTCTGGATATGGCCGACTTCGAGTTTGAGCAGATGTTTACCGATGCCCTTGGAATTGAC
A L D M A D F E F E Q M F T D A L G I D
EcoRI
|
GAGTACGGTGGGGGTGCGAATTCAAAGGACTATACGGAGCATTTGCCGGATGCTGTAGAT
E Y G G G A N S K D Y T E H L P D A V D
XhoI
I
TTTGAAGATGAAGATGAACTTGCTGATGACGATGACGATTTGCCAGAAGAATGATAGCTC
F E D E D E L A D D D D D L P E E * *
NotI

GAGCGGCCGC

148

 

T7\promoter 6564...6580
Iacl 6173...5091

  
     
  

pDSE117-0
6580 bp

ColE1,\pBR322\ori 3897...3897
Ap 3136...2276

Figure 31: Plasmid family pDSE117-n.
Parent plasmid: pETDUET—l (Novagen)

Inserted sequences: GST between NcoI and BamHI

VP16C between BamHI-EcoRI
TAND2 between EcoRI and BamHI
TBP between BamHI and XhoI

Note: In this plasmid family, n indicates the number of insertions of the sequence
GGATCAGGCTCT upstream from the BamHI site between TAND2 and TBP.

149

 

10.

11.

List of references

Almlof, T., A. E. Wallberg, J. A. Gustafsson, and A. P. Wright. 1998. Role of
important hydrophobic amino acids in the interaction between the glucocorticoid
receptor tau l-core activation domain and target factors. Biochemistry 37 :9586-94.

Auboeuf, D., A. Honig, S. M. Berget, and B. W. O'Malley. 2002. Coordinate
regulation of transcription and splicing by steroid receptor coregulators. Science
298:416-9.

Bagby, S., T. K. Mal, D. Liu, E. Raddatz, Y. Nakatani, and M. Ikura. 2000.
TFIIA-TAF regulatory interplay: NMR evidence for overlapping binding sites on
TBP. FEBS Lett 468: 149-54.

Bajic, V. B., S. L. Tan, A. Chong, S. Tang, A. Strom, J. A. Gustafsson, C. Y.
Lin, and E. T. Liu. 2003. Dragon ERE Finder version 2: A tool for accurate

detection and analysis of estrogen response elements in vertebrate genomes.
Nucleic Acids Res 31:3605-7.

Barboric, M., R. M. Nissen, S. Kanazawa, N. Jabrane-Ferrat, and B. M.

Peterlin. 2001. NF -kappaB binds P-TEFb to stimulate transcriptional elongation
by RNA polymerase 11. Mol Cell 82327-37.

Beerli, R. R., and C. F. Barbas, 3rd. 2002. Engineering polydactyl zinc-finger
transcription factors. Nat Biotechnol 20: 1 35-41.

Bell, S. D., P. L. Kosa, P. B. Sigler, and S. P. Jackson. 1999. Orientation of the
transcription preinitiation complex in Archaea. Proc Natl Acad Sci U S A
96:13662-13667.

Berger, S. L. 2002. Histone modifications in transcriptional regulation. Curr Opin
Genet Dev 12:142-8.

Bleichenbacher, M., S. Tan, and T. J. Richmond. 2003. Novel interactions
between the components of human and yeast TFIIA/TBP/DNA complexes. J Mol
Biol 332:783-93.

Boardman, P. E., S. G. Oliver, and S. J. Hubbard. 2003. SiteSeer:
Visualisation and analysis of transcription factor binding sites in nucleotide
sequences. Nucleic Acids Res 31:3572-5.

Boube, M., L. Joulia, D. L. Cribbs, and H. M. Bourbon. 2002. Evidence for a
mediator of RNA polymerase II transcriptional regulation conserved from yeast to
man. Cell 110: 143-51.

150

 

12.

l3.

14.

15.

16.

17.

18.

19.

20.

21.

22.

23.

24.

Brooks, C. L., and W. On. 2003. Ubiquitination, phosphorylation and
acetylation: the molecular basis for p53 regulation. Curr Opin Cell Biol 15:164-71.

Brown, S. A., A. N. lmbalzano, and R. E. Kingston. 1996. Activator-dependent
regulation of transcriptional pausing on nucleosomal templates. Genes Dev
10: 1479-90.

Brown, S. A., C. S. Weirich, E. M. Newton, and R. E. Kingston. 1998.
Transcriptional activation domains stimulate initiation and elongation at different
times and via different residues. Embo I 17 :3 146-54.

Bucher, P. 1990. Weight matrix descriptions of four eukaryotic RNA polymerase
11 promoter elements derived from 502 unrelated promoter sequences. J Mol Biol
212:563-78.

Buratowski, S., S. Hahn, L. Guarente, and P. A. Sharp. 1989. Five
intermediate complexes in transcription initiation by RNA polymerase 11. Cell
56:549-61.

Burley, S. K., and R. G. Roeder. 1996. Biochemistry and structural biology of
transcription factor IID (TFIID). Annu Rev Biochem 65:769-99.

Butler, J. E., and J. T. Kadonaga. 2001. Enhancer-promoter specificity
mediated by DPE or TATA core promoter motifs. Genes Dev 15:2515-9.

Candau, R., D. M. Scolnick, P. Darpino, C. Y. Ying, T. D. Halazonetis, and S.
L. Berger. 1997. Two tandem and independent sub-activation domains in the

amino terminus of p53 require the adaptor complex for activity. Oncogene
15:807-16.

Chasman, D. 1., K. M. Flaherty, P. A. Sharp, and R. D. Kornberg. 1993.
Crystal structure of yeast TATA-binding protein and model for interaction with
DNA. Proc Natl Acad Sci U S A 90:8174-8.

Chiang, C. M., and R. G. Roeder. 1995. Cloning of an intrinsic human TFIID
subunit that interacts with multiple transcriptional activators. Science 267:531-6.

Cho, E. J., M. S. Kobor, M. Kim, J. Greenblatt, and S. Buratowski. 2001.
Opposing effects of Ctkl kinase and F cpl phosphatase at Ser 2 of the RNA
polymerase II C-terminal domain. Genes Dev 15:3319-29.

Coleman, R. A., A. K. Taggart, L. R. Benjamin, and B. F. Pugh. 1995.
Dimerization of the TATA binding protein. J Biol Chem 270: 13842-9.

Corbi, N ., M. Perez, R. Maione, and C. Passananti. 1997. Synthesis of a new

zinc finger peptide; comparison of its 'code' deduced and 'CASTing' derived
binding sites. F EBS Lett 417:71-4.

151

25.

26.

27.

28.

29.

30.

31.

32.

33.

34.

35.

36.

Cormack, B. P., and K. Struhl. 1992. The TATA-binding protein is required for
transcription by all three nuclear RNA polymerases in yeast cells. Cell 69:685-96.

Cox, J. M., M. M. Hayward, J. F. Sanchez, L. D. Gegnas, S. van der Zee, J. H.
Dennis, P. B. Sigler, and A. Schepartz. 1997. Bidirectional binding of the
TATA box binding protein to the TATA box. Proc Natl Acad Sci U S A
94:13475-80.

Cramer, P., C. G. Pesce, F. E. Baralle, and A. R. Kornblihtt. 1997. Functional
association between promoter structure and transcript alternative splicing. Proc
Natl Acad Sci U S A 94:11456-60.

Cress, W. D., and S. J. Triezenberg. 1991. Critical structural elements of the
VP16 transcriptional activation domain. Science 251:87-90.

Dahlman-Wright, K., H. Baumann, I. J. McEwan, T. Almlof, A. P. Wright, J.
A. Gustafsson, and T. Hard. 1995. Structural characterization of a minimal

functional transactivation domain from the human glucocorticoid receptor. Proc
Natl Acad Sci U S A 92: 1699-703.

 

Donaldson, L., and J. P. Capone. 1992. Purification and characterization of the
carboxyl-terminal transactivation domain of Vmw65 from herpes simplex virus
type 1. J Biol Chem 267:1411-4.

Dynlacht, B. D., T. Hoey, and R. Tjian. 1991. Isolation of coactivators
associated with the TATA-binding protein that mediate transcriptional activation.
Cell 66:563-76.

Emami, K. H., W. W. Navarre, and S. T. Smale. 1995. Core promoter
specificities of the Spl and VP16 transcriptional activation domains. Mol Cell
Biol 15:5906-16.

Evans, R., J. A. Fairley, and S. G. Roberts. 2001. Activator- mediated
disruption of sequence-specific DNA contacts by the general transcription factor
TFIIB. Genes Dev 15:2945-9.

Falke, D., and R. L. Juliano. 2003. Selective gene regulation with designed
transcription factors: implications for therapy. Curr Opin Mol Ther 5:161-6.

Felinski, E. A., and P. G. Quinn. 1999. The CREB constitutive activation
domain interacts with TATA-binding protein-associated factor 1 10 (TAF 1 10)

through specific hydrophobic residues in one of the three subdomains required for
both activation and TAF l 10 binding. J Biol Chem 274:11672-8.

Ferdous, A., T. Kodadek, and S. A. Johnston. 2002. A nonproteolytic function
of the 198 regulatory subunit of the 26S proteasome is required for efficient
activated transcription by human RNA polymerase II. Biochemistry 41:12798-
805.

152

37.

38.

39.

40.

41.

42.

43.

44.

45.

46.

47.

48.

49.

Garvie, C. W., and C. Wolberger. 2001. Recognition of specific DNA
sequences. Mol Cell 8:937-46.

Ge, K., M. Guermah, C. X. Yuan, M. Ito, A. E. Wallberg, B. M. Spiegelman,
and R. G. Roeder. 2002. Transcription coactivator TRAP220 is required for
PPAR gamma 2-stimulated adipogenesis. Nature 417:563-7.

Geiger, J. H., S. Hahn, S. Lee, and P. B. Sigler. 1996. Crystal structure of the
yeast TFIIA/TBP/DNA complex. Science 272:830-6.

Ghosh, S., and M. Karin. 2002. Missing pieces in the NF-kappaB puzzle. Cell
109 Suppl:S8l-96.

Gill, G., E. Pascal, Z. H. Tseng, and R. Tjian. 1994. A glutamine-rich
hydrophobic patch in transcription factor Spl contacts the dTAFIIl 10 component
of the Drosophila TFIID complex and mediates transcriptional activation. Proc
Natl Acad Sci U S A 91: 192-6.

Gonzalez, F., A. Delahodde, T. Kodadek, and S. A. Johnston. 2002.
Recruitment of a 198 proteasome subcomplex to an activated promoter. Science
296:548-50.

Goodrich, J. A., T. Hoey, C. J. Thut, A. Admon, and R. Tjian. 1993.
Drosophila TAF 1140 interacts with both a VP16 activation domain and the basal
transcription factor TFIIB. Cell 75:519-30.

Gossen, M., and H. Bujard. 2002. Studying gene function in eukaryotes by
conditional gene inactivation. Annu Rev Genet 36: 153-73.

Gossen, M., and H. Bujard. 1992. Tight control of gene expression in
mammalian cells by tetracycline-responsive promoters. Proc Natl Acad Sci U S A
89:5547-51.

Gossen, M., S. Freundlieb, G. Bender, G. Muller, W. Hillen, and H. Bujard.
1995. Transcriptional activation by tetracyclines in mammalian cells. Science
268: 1766-9.

Grabe, N. 2002. AliBaba2: context specific identification of transcription factor
binding sites. In Silico Biol 2:81-15.

Grant, P. A., D. Schieltz, M. G. Pray-Grant, D. J. Steger, J. C. Reese, J. R.
Yates, 3rd, and J. L. Workman. 1998. A subset of TAF(II)s are integral
components of the SAGA complex required for nucleosome acetylation and
transcriptional stimulation. Cell 94:45-53.

Greaves, R., and P. O'Hare. 1989. Separation of requirements for protein-DNA

complex assembly from those for functional activity in the herpes simplex virus
regulatory protein Vmw65. J Virol 63: 1641-50.

153

 

50.

51.

52.

53.

54.

55.

56.

57.

58.

59.

60.

61.

62.

Hahn, S. 2004. Structure and mechanism of the RNA polymerase II transcription
machinery. Nat Struct Mol Biol 11:394-403.

Hampsey, M., and D. Reinberg. 1999. RNA polymerase II as a control panel for
multiple coactivator complexes. Curr Opin Genet Dev 9: 132-9.

Hayashi, F., R. Ishima, D. Liu, K. I. Tong, S. Kim, D. Reinberg, S. Bagby,
and M. Ikura. 1998. Human general transcription factor TFIIB: conformational
variability and interaction with VP16 activation domain. Biochemistry 37:7941-
51.

Heery, D. M., E. Kalkhoven, S. Hoare, and M. G. Parker. 1997. A signature
motif in transcriptional co-activators mediates binding to nuclear receptors.
Nature 387:733-6.

Hengartner, C. J., C. M. Thompson, J. Zhang, D. M. Chao, S. M. Liao, A. J.
Koleske, S. Okamura, and R. A. Young. 1995. Association of an activator with
an RNA polymerase 11 holoenzyme. Genes Dev 9:897-910.

Iizuka, M., and M. M. Smith. 2003. Functional consequences of histone
modifications. Curr Opin Genet Dev 13: 154-60.

Ingles, C. J., M. Shales, W. D. Cress, S. J. Triezenberg, and J. Greenblatt.
1991. Reduced binding of TFIID to transcriptionally compromised mutants of
VP16. Nature 351:588-90.

Ito, M., C. X. Yuan, S. Malik, W. Gu, J. D. Fondell, S. Yamamura, Z. Y. Fu,
X. Zhang, J. Qin, and R. G. Roeder. 1999. Identity between TRAP and SMCC
complexes indicates novel pathways for the function of nuclear receptors and
diverse mammalian activators. Mol Cell 3:361-70.

Jackson, B. M., C. M. Drysdale, K. Natarajan, and A. G. Hinnebusch. 1996.
Identification of seven hydrophobic clusters in GCN4 making redundant
contributions to transcriptional activation. Mol Cell Biol 16:5557-71.

Jenuwein, T., and C. D. Allis. 2001. Translating the histone code. Science
293: 1074-80.

Joliot, V., M. Demma, and R. Prywes. 1995. Interaction with RAP74 subunit of
TFIIF is required for transcriptional activation by serum response factor. Nature
373:632-5.

Jonker, H. R., R. W. Wechselberger, R. Boelens, G. E. Folkers, and R.
Kaptein. 2005. Structural properties of the promiscuous VP16 activation domain.
Biochemistry 44:827-39.

J uo, Z. S., T. K. Chiu, P. M. Leiberman, 1. Baikalov, A. J. Berk, and R. E.
Dickerson. 1996. How proteins recognize the TATA box. J Mol Biol 261:239-54.

154

 

63.

64.

65.

66.

67.

68.

69.

70.

71.

72.

73.

74.

75.

Kadonaga, J. T. 2002. The DPE, a core promoter element for transcription by
RNA polymerase 11. Exp Mol Med 34:259-64.

Kamiuchi, T., E. Abe, M. lmanishi, T. Kaji, M. N agaoka, and Y. Sugiura.
1998. Artificial nine zinc- finger peptide with 30 base pair binding sites.
Biochemistry 37: 13827-34.

Kanazawa, S., T. Okamoto, and B. M. Peterlin. 2000. Tat competes with
CIITA for the binding to P-TEFb and blocks the expression of MHC class 11

genes in HIV infection. Immunity 12:61-70.

Kanazawa, S., L. Soucek, G. Evan, T. Okamoto, and B. M. Peterlin. 2003. c-
Myc recruits P-TEF b for transcription, cellular proliferation and apoptosis.
Oncogene 22:5707-11.

Kays, A. R., and A. Schepartz. 2002. Gal4-VP16 and Gal4-AH increase the
orientational and axial specificity of TATA box recognition by TATA box
binding protein. Biochemistry 41:3147-55.

Kays, A. R., and A. Schepartz. 2000. Virtually unidirectional binding of TBP to
the AdMLP TATA box within the quaternary complex with TFIIA and TFIIB.
Chem Biol 7:601-610.

Kel, A. E., E. Gossling, l. Reuter, E. Cheremushkin, O. V. Kel-Margoulis,
and E. Wingender. 2003. MATCH: A tool for searching transcription factor
binding sites in DNA sequences. Nucleic Acids Res 31:3576-9.

Kim, J. L., D. B. N ikolov, and S. K. Burley. 1993. Co-crystal structure of TBP
recognizing the minor groove of a TATA element. Nature 365:520-7.

Kim, L. J., A. G. Seto, T. N. Nguyen, and J. A. Goodrich. 2001. Human
Taf(II)130 is a coactivator for NF ATp. Mol Cell Biol 21:3503-13.

Kim, T. K., S. Hashimoto, R. J. Kelleher, 3rd, P. M. Flanagan, R. D.
Kornberg, M. Horikoshi, and R. G. Roeder. 1994. Effects of activation-
defective TBP mutations on transcription initiation in yeast. Nature 369:252-5.

Kim, Y., J. H. Geiger, S. Hahn, and P. B. Sigler. 1993. Crystal structure of a
yeast TBP/TATA-box complex. Nature 365:512-20.

Kim, Y. J ., S. Bjorklund, Y. Li, M. H. Sayre, and R. D. Kornberg. 1994. A
multiprotein mediator of transcriptional activation and its interaction with the C-
terminal repeat domain of RNA polymerase 11. Cell 77:599-608.

Klemm, R. D., J. A. Goodrich, S. Zhou, and R. Tjian. 1995. Molecular cloning
and expression of the 32-kDa subunit of human TFIID reveals interactions with
VP16 and TFIIB that mediate transcriptional activation. Proc Natl Acad Sci U S
A 92:5788-92.

155

76.

77.

78.

79.

80.

81.

82.

83.

84.

85.

86.

Kobayashi, N ., T. G. Boyer, and A. J. Berk. 1995. A class of activation
domains interacts directly with TFIIA and stimulates TFIIA-TFIID-promoter
complex assembly. Mol Cell Biol 15:6465-73.

Kobayashi, N ., P. J. Horn, S. M. Sullivan, S. J. Triezenberg, T. G. Boyer, and
A. J. Berk. 1998. DA-complex assembly activity required for VP16C
transcriptional activation. Mol Cell Biol 18:4023-31.

Kokubo, T., D. W. Gong, S. Yamashita, M. Horikoshi, R. G. Roeder, and Y.
N akatani. 1993. Drosophila 230-kD TFIID subunit, a functional homolog of the

human cell cycle gene product, negatively regulates DNA binding of the TATA
box-binding subunit of TFIID. Genes Dev 7: 1033-46.

Kokubo, T., M. J. Swanson, J. I. Nishikawa, A. G. Hinnebusch, and Y.
N akatani. 1998. The yeast TAF145 inhibitory domain and TFIIA competitively
bind to TATA-binding protein. Mol Cell Biol 18: 1003- 12.

Kokubo, T., S. Yamashita, M. Horikoshi, R. G. Roeder, and Y. N akatani.
1994. Interaction between the N-terminal domain of the 230-kDa subunit and the
TATA box-binding subunit of TFIID negatively regulates TATA-box binding.
Proc Natl Acad Sci U S A 91:3520-4.

Koleske, A. J., and R. A. Young. 1994. An RNA polymerase 11 holoenzyme
responsive to activators. Nature 368:466-9.

Kosa, P. F., G. Ghosh, B. S. DeDecker, and P. B. Sigler. 1997. The 2.1-A
crystal structure of an archaeal preinitiation complex: TATA-box-binding
protein/transcription factor (II)B core/TATA-box. Proc Natl Acad Sci U S A
94:6042-7.

Kotani, T., K. Banno, M. Ikura, A. G. Hinnebusch, Y. N akatani, M.
Kawaichi, and T. Kokubo. 2000. A role of transcriptional activators as

antirepressors for the autoinhibitory activity of TATA box binding of
transcription factor IID. Proc Natl Acad Sci U S A 97 :7178-83.

Kotani, T., T. Miyake, Y. Tsukihashi, A. G. Hinnebusch, Y. N akatani, M.
Kawaichi, and T. Kokubo. 1998. Identification of highly conserved amino-
terminal segments of dTAF 11230 and yTAF 11145 that are functionally

interchangeable for inhibiting TBP-DNA interactions in vitro and in promoting
yeast cell growth in vivo. J Biol Chem 273:32254-64.

Kou, H., J. D. Irvin, K. L. Huisinga, M. Mitra, and B. F. Pugh. 2003.
Structural and functional analysis of mutations along the crystallographic dimer
interface of the yeast TATA binding protein. Mol Cell Biol 23:3186-201.

Kulkarni, M. M., and D. N. Arnosti. 2003. Information display by
transcriptional enhancers. Development 130:6569-75.

156

87.

88.

89.

90.

91.

92.

93.

94.

95.

96.

97.

98.

99.

Kumar, K. P., S. Akoulitchev, and D. Reinberg. 1998. Promoter-proximal
stalling results from the inability to recruit transcription factor IIH to the

transcription complex and is a regulated event. Proc Natl Acad Sci U S A
95:9767-72.

Kussie, P. H., S. Gorina, V. Marechal, B. Elenbaas, J. Moreau, A. J. Levine,
and N. P. Pavletich. 1996. Structure of the MDM2 oncoprotein bound to the p53
tumor suppressor transactivation domain. Science 274:948-53.

Lau, J. F., I. Nusinzon, D. Burakov, L. P. Freedman, and C. M. Horvath.
2003. Role of metazoan mediator proteins in interferon-responsive transcription.
Mol Cell Biol 23:620-8.

Lee, C., J. H. Chang, H. S. Lee, and Y. Cho. 2002. Structural basis for the
recognition of the E2F transactivation domain by the retinoblastoma tumor
suppressor. Genes Dev 16:3199-212.

Lee, T. I., and R. A. Young. 2000. Transcription of eukaryotic protein-coding
genes. Annu Rev Genet 34:77-137.

Lewis, B. A., and D. Reinberg. 2003. The mediator coactivator complex:
functional and physical roles in transcriptional regulation. J Cell Sci 116:3667-75.

Lieberman, P. M., and A. J. Berk. 1994. A mechanism for TAFs in
transcriptional activation: activation domain enhancement of TFIID-TFIIA--
promoter DNA complex formation. Genes Dev 8:995-1006.

Liljelund, P., C. J. Ingles, and J. Greenblatt. 1993. Altered promoter binding of
the TATA box-binding factor induced by the transcriptional activation domain of
VP16 and suppressed by TFIIA. Mol Gen Genet 241:694-9.

Lin, Q., C. F. Barbas, 3rd, and P. G. Schultz. 2003. Small-molecule switches
for zinc finger transcription factors. J Am Chem Soc 125:612-3.

Lin, Y. S., 1. Ha, E. Maldonado, D. Reinberg, and M. R. Green. 1991. Binding
of general transcription factor TFIIB to an acidic activating region. Nature
353:569-71.

Lipford, J. R., and R. J. Deshaies. 2003. Diverse roles for ubiquitin-dependent
proteolysis in transcriptional activation. Nat Cell Biol 5:845-50.

Littlefield, O., Y. Korkhin, and P. B. Sigler. 1999. The structural basis for the
oriented assembly of a TBP/T F B/promoter complex. Proc Natl Acad Sci U S A
96:13668-13673.

Liu, D., R. Ishima, K. I. Tong, S. Bagby, T. Kokubo, D. R. Muhandiram, L. E.
Kay, Y. N akatani, and M. Ikura. 1998. Solution structure of a TBP-TAF(II)230

157

100.

101.

102.

103.

104.

105.

106.

107.

108.

109.

complex: protein mimicry of the minor groove surface of the TATA box unwound
by TBP. Cell 94:573-83.

Liu, P. Q., E. J. Rebar, L. Zhang, Q. Liu, A. C. Jamieson, Y. Liang, H. Qi, P.
X. Li, B. Chen, M. C. Mendel, X. Zhang, Y. L. Lee, S. P. Eisenberg, S. K.
Spratt, C. C. Case, and A. P. Wolffe. 2001. Regulation of an endogenous locus
using a panel of designed zinc finger proteins targeted to accessible chromatin
regions. Activation of vascular endothelial growth factor A. J Biol Chem

276: 1 1323-34.

Luscombe, N. M., and J. M. Thornton. 2002. Protein-DNA interactions: amino
acid conservation and the effects of mutations on binding specificity. J Mol Biol
320:991- 1009.

Mal, T. K., Y. Masutomi, L. Zheng, Y. N akata, H. Ohta, Y. N akatani, T.
Kokubo, and M. Ikura. 2004. Structural and functional characterization on the
interaction of yeast TFIID subunit TAF 1 with TATA-binding protein. J Mol Biol
339:681-93.

Malik, S., and R. G. Roeder. 2000. Transcriptional regulation through Mediator-
like coactivators in yeast and metazoan cells. Trends Biochem Sci 25:277-83.

Maniatis, T., and R. Reed. 2002. An extensive network of coupling among gene
expression machines. Nature 416:499-506.

Marmorstein, R., and M. X. Fitzgerald. 2003. Modulation of DNA-binding
domains for sequence-specific DNA recognition. Gene 304:1-12.

Martel, L. S., H. J. Brown, and A. J. Berk. 2002. Evidence that TAF-TATA
box-binding protein interactions are required for activated transcription in
mammalian cells. Mol Cell Biol 22:2788-98.

Martin, M. L., P. M. Lieberman, and T. Curran. 1996. F os-Jun dimerization
promotes interaction of the basic region with TFIIB-34 and TFIIF. Mol Cell Biol
16:2110-8.

Matys, V., E. Fricke, R. Geffers, E. Gossling, M. Haubrock, R. Hehl, K.
Hornischer, D. Karas, A. E. Kel, O. V. Kel-Margoulis, D. U. Kloos, S. Land,
B. Lewicki-Potapov, H. Michael, R. Munch, I. Reuter, S. Rotert, H. Saxel, M.
Scheer, S. Thiele, and E. Wingender. 2003. TRANSF AC: transcriptional
regulation, from patterns to profiles. Nucleic Acids Res 31:374-8.

McEwan, I. J., K. Dahlman-Wright, J. Ford, and A. P. Wright. 1996.
Functional interaction of the c-Myc transactivation domain with the TATA

binding protein: evidence for an induced fit model of transactivation domain
folding. Biochemistry 35:9584-93.

158

 

110.

111.

112.

113.

114.

115.

116.

117.

118.

119.

120.

121.

122.

McKenna, N. J., and B. W. O'Malley. 2002. Combinatorial control of gene
expression by nuclear receptors and coregulators. Cell 108:465-74.

Meares, C. F., S. A. Datwyler, B. D. Schmidt, J. Owens, and A. Ishihama.
2003. Principles and methods of affinity cleavage in studying transcription.
Methods Enzymol 371 :82- 106.

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

Mitchell, P. J., and R. Tjian. 1989. Transcriptional regulation in mammalian
cells by sequence-specific DNA binding proteins. Science 245:371-8.

Mittler, G., T. Stuhler, L. Santolin, T. Uhlmann, E. Kremmer, F. Lottspeich,
L. Berti, and M. Meisterernst. 2003. A novel docking site on Mediator is critical
for activation by VP16 in mammalian cells. Embo J 22:6494-6504.

Muratani, M., and W. P. Tansey. 2003. How the ubiquitin-proteasome system
controls transcription. Nat Rev Mol Cell Biol 4: 192-201.

Myers, L. C., and R. D. Kornberg. 2000. Mediator of transcriptional regulation.
Annu Rev Biochem 69:729-49.

N akajima, N ., M. Horikoshi, and R. G. Roeder. 1988. Factors involved in
specific transcription by mammalian RNA polymerase II: purification, genetic
specificity, and TATA box-promoter interactions of TFIID. Mol Cell Biol
8:4028-40.

N arlikar, G. J., H. Y. Fan, and R. E. Kingston. 2002. Cooperation between
complexes that regulate chromatin structure and transcription. Cell 108:475-87.

N edialkov, Y. A., D. D. Shooltz, and S. J. Triezenberg. 2003. Purification and
protein interaction assays of the VP16C transcription activation domain. Methods
Enzymol 370:522-35.

N edialkov, Y. A., and S. J. Triezenberg. 2004. Quantitative assessment of in
vitro interactions implicates TATA-binding protein as a target of the VP16C
transcriptional activation region. Arch Biochem Biophys 425:77-86.

N ikolov, D. B., H. Chen, E. D. Halay, A. Hoffman, R. G. Roeder, and S. K.
Burley. 1996. Crystal structure of a human TATA box-binding protein/TATA
element complex. Proc Natl Acad Sci U S A 93:4862-7.

Nikolov, D. B., H. Chen, E. D. Halay, A. A. Usheva, K. Hisatake, D. K. Lee, R.
G. Roeder, and S. K. Burley. 1995. Crystal structure of a TFIIB-TBP-TATA-
element ternary complex. Nature 377:1 19-28.

159

 

123.

124.

125.

126.

127.

128.

129.

130.

131.

132.

133.

134.

135.

Nikolov, D. B., S. H. Hu, J. Lin, A. Gasch, A. Hoffmann, M. Horikoshi, N. H.
Chua, R. G. Roeder, and S. K. Burley. 1992. Crystal structure of TFIID TATA-
box binding protein. Nature 360:40-6.

N ishikawa, J ., T. Kokubo, M. Horikoshi, R. G. Roeder, and Y. N akatani.
1997. Drosophila TAF(II)230 and the transcriptional activator VP16 bind
competitively to the TATA box-binding domain of the TATA box-binding protein.
Proc Natl Acad Sci U S A 94:85-90.

N ogues, G., S. Kadener, P. Cramer, D. Bentley, and A. R. Kornblihtt. 2002.
Transcriptional activators differ in their abilities to control alternative splicing. J
Biol Chem 277:43110-4.

O'Hare, P., and G. Williams. 1992. Structural studies of the acidic
transactivation domain of the Vmw65 protein of herpes simplex virus using 1H
NMR. Biochemistry 31:4150-6.

O'Shea-Greenfield, A., and S. T. Smale. 1992. Roles of TATA and initiator
elements in determining the start site location and direction of RNA polymerase II
transcription. J Biol Chem 267: 1391-402.

Ottosen, S., F. J. Herrera, and S. J. Triezenberg. 2002. Transcription.
Proteasome parts at gene promoters. Science 296:479—8 1.

Pan, C. Q., R. Landgraf, and D. S. Sigman. 1994. DNA-binding proteins as
site-specific nucleases. Mol Microbiol 12:335-42.

Pearson, A., and J. Greenblatt. 1997. Modular organization of the E2F1
activation domain and its interaction with general transcription factors TBP and
TFIIH. Oncogene 15:2643-58.

Poon, D., R. A. Knittle, K. A. Sabelko, T. Yamamoto, M. Horikoshi, R. G.
Roeder, and P. A. Wei]. 1993. Genetic and biochemical analyses of yeast
T ATA-binding protein mutants. J Biol Chem 268:5005- 13.

Praz, V., R. Perier, C. Bonnard, and P. Bucher. 2002. The Eukaryotic
Promoter Database, EPD: new entry types and links to gene expression data.
Nucleic Acids Res 30:322-4.

Proudfoot, N. J., A. Furger, and M. J. Dye. 2002. Integrating mRNA
processing with transcription. Cell 108:501-12.

Ptashne, M., and A. Gann. 1997. Transcriptional activation by recruitment.
Nature 386:569-77.

Pugh, B. F. 2000. Control of gene expression through regulation of the TATA-
binding protein. Gene 255: 1-14.

160

136.

137.

138.

139.

140.

141.

142.

143.

144.

145.

146.

147.

Rachez, C., and L. P. Freedman. 2001. Mediator complexes and transcription.
Curr Opin Cell Biol 13:274-80.

Radhakrishnan, I., G. C. Perez-Alvarado, D. Parker, H. J. Dyson, M. R.
Montminy, and P. E. Wright. 1997. Solution structure of the KIX domain of
CBP bound to the transactivation domain of CREB: a model for
activatorzcoactivator interactions. Cell 91 :741- 52.

Rawson, R. B. 2003. The SREBP pathway--insights from Insigs and insects. Nat
Rev Mol Cell Biol 4:631-40.

Regier, J. L., F. Shen, and S. J. Triezenberg. 1993. Pattern of aromatic and
hydrophobic amino acids critical for one of two subdomains of the VP16
transcriptional activator. Proc Natl Acad Sci U S A 90:883-7.

Ren, B., H. Cam, Y. Takahashi, T. Volkert, J. Terragni, R. A. Young, and B.
D. Dynlacht. 2002. E2F integrates cell cycle progression with DNA repair,
replication, and G(2)/M checkpoints. Genes Dev 16:245-56.

Ren, B., F. Robert, J. J. Wyrick, O. Aparicio, E. G. Jennings, 1. Simon, J.
Zeitlinger, J. Schreiber, N. Hannett, E. Kanin, T. L. Volkert, C. J. Wilson, S.
P. Bell, and R. A. Young. 2000. Genome-wide location and fimction of DNA
binding proteins. Science 290:2306-9.

Roberts, S. G., and M. R. Green. 1994. Activator- induced conformational
change in general transcription factor TFIIB. Nature 371:717-20.

Roberts, S. G., 1. Ha, E. Maldonado, D. Reinberg, and M. R. Green. 1993.
Interaction between an acidic activator and transcription factor TFIIB is required
for transcriptional activation. Nature 363:741-4.

Rojo-Niersbach, E., T. Furukawa, and N. Tanese. 1999. Genetic dissection of
hTAF(II)130 defines a hydrophobic surface required for interaction with
glutamine-rich activators. J Biol Chem 274:33778-84.

Rougvie, A. E., and J. T. Lis. 1988. The RNA polymerase II molecule at the 5'
end of the uninduced hsp70 gene of D. melanogaster is transcriptionally engaged.
Cell 54:795-804.

Sadowski, 1., J. Ma, S. Triezenberg, and M. Ptashne. 1988. GAL4-VP16 is an
unusually potent transcriptional activator. Nature 335:563-4.

Salghetti, S. E., A. A. Caudy, J. G. Chenoweth, and W. P. Tansey. 2001.
Regulation of transcriptional activation domain function by ubiquitin. Science
293: 1651-3.

161

148.

149.

150.

151.

152.

153.

154.

155.

156.

157.

158.

159.

160.

Salghetti, S. E., M. Muratani, H. Wijnen, B. Futcher, and W. P. Tansey. 2000.
Functional overlap of sequences that activate transcription and signal ubiquitin-
mediated proteolysis. Proc Natl Acad Sci U S A 97:3118-23.

Sauer, F., S. K. Hansen, and R. T jian. 1995. Multiple TAFIIs directing
synergistic activation of transcription. Science 270: 1783-8.

Schmitz, M. L., M. A. dos Santos Silva, H. Altmann, M. Czisch, T. A. Holak,
and P. A. Baeuerle. 1994. Structural and functional analysis of the NF -kappa B
p65 C terminus. An acidic and modular transactivation domain with the potential
to adopt an alpha-helical conformation. J Biol Chem 269:25613-20.

Sera, T., and C. Uranga. 2002. Rational design of artificial zinc- finger proteins
using a nondegenerate recognition code table. Biochemistry 41:7074—8 l.

Shen, F., S. J. Triezenberg, P. Hensley, D. Porter, and J. R. Knutson. 1996.
Critical amino acids in the transcriptional activation domain of the herpesvirus

protein VP16 are solvent-exposed in highly mobile protein segments. An intrinsic
fluorescence study. J Biol Chem 271:4819-26.

Shen, F., S. J. Triezenberg, P. Hensley, D. Porter, and J. R. Knutson. 1996.
Transcriptional activation domain of the herpesvirus protein VP16 becomes

conformationally constrained upon interaction with basal transcription factors. J
Biol Chem 271:4827-37.

Sigler, P. B. 1988. Transcriptional activation. Acid blobs and negative noodles.
Nature 333:210-2.

Sigman, D. S., M. D. Kuwabara, C. H. Chen, and T. W. Bruice. 1991.
Nuclease activity of 1,10-phenanthroline-copper in study of protein-DNA
interactions. Methods Enzymol 208:414-33.

Smale, S. T., and J. T. Kadonaga. 2003. The RNA polymerase 11 core promoter.
Annu Rev Biochem 72:449-79.

Spencer, J. V., and K. M. Arndt. 2002. A TATA binding protein mutant with
increased affinity for DNA directs transcription from a reversed TATA sequence
in vivo. Mol Cell Biol 22:8744-8755.

Strahl, B. D., and C. D. Allis. 2000. The language of covalent histone
modifications. Nature 403:41-5.

Stringer, K. F., C. J. Ingles, and J. Greenblatt. 1990. Direct and selective
binding of an acidic transcriptional activation domain to the TATA-box factor
TFIID. Nature 345:783-6.

Sullivan, S. M., P. J. Horn, V. A. Olson, A. H. Koop, W. Niu, R. H. Ebright,
and S. J. Triezenberg. 1998. Mutational analysis of a transcriptional activation

162

161.

162.

163.

164.

165.

166.

167.

168.

169.

170.

171.

172.

region of the VP16 protein of herpes simplex virus. Nucleic Acids Res 26:4487-
96.

Tan, 8., Y. Hunziker, D. F. Sargent, and T. J. Richmond. 1996. Crystal
structure of a yeast TFIIA/TBP/DNA complex. Nature 381: 127-5 1.

Tora, L. 2002. A unified nomenclature for TATA box binding protein (TBP)-
associated factors (TAPS) involved in RNA polymerase II transcription. Genes
Dev 16:673-5.

Triezenberg, S. J. 1995. Structure and function of transcriptional activation
domains. Curr Opin Genet Dev 5:190-6.

Triezenberg, S. J., R. C. Kingsbury, and S. L. McKnight. 1988. Functional
dissection of VP16, the trans-activator of herpes simplex virus immediate early
gene expression. Genes Dev 2:718-29.

Tsai, F. T. F., and P. B. Sigler. 2000. Structural basis of preinitiation complex
assembly on human Pol II promoters. Embo J 19:25-36.

Uesugi, M., O. Nyanguile, H. Lu, A. J. Levine, and G. L. Verdine. 1997.
Induced alpha helix in the VP16 activation domain upon binding to a human TAF.
Science 277: 1310-3.

Van Dyke, M. W., R. G. Roeder, and M. Sawadogo. 1988. Physical analysis of
transcription preinitiation complex assembly on a class II gene promoter. Science
241:1335-8.

Van Hoy, M., K. K. Leuther, T. Kodadek, and S. A. Johnston. 1993. The
acidic activation domains of the GCN4 and GAL4 proteins are not alpha helical
but form beta sheets. Cell 72:587-94.

Walker, 8., R. Greaves, and P. O'Hare. 1993. Transcriptional activation by the
acidic domain of Vmw65 requires the integrity of the domain and involves

additional determinants distinct from those necessary for TFIIB binding. Mol Cell
Biol 13:5233-44.

Warnmark, A., E. T reuter, A. P. Wright, and J. A. Gustafsson. 2003.
Activation functions 1 and 2 of nuclear receptors: molecular strategies for
transcriptional activation. Mol Endocrinol 17: 1901-9.

Warnmark, A., A. Wikstrom, A. P. Wright, J. A. Gustafsson, and T. Hard.
2001. The N-terminal regions of estrogen receptor alpha and beta are unstructured
in vitro and show different TBP binding properties. J Biol Chem 276:45939-44.

Weinmann, A. S., S. M. Bartley, T. Zhang, M. Q. Zhang, and P. J. Farnham.

2001. Use of chromatin immunoprecipitation to clone novel E2F target promoters.
Mol Cell Biol 21:6820-32.

163

 

173.

174.

175.

176.

177.

178.

179.

180.

I81.

182.

183.

184.

Weinmann, P., M. Gossen, W. Hillen, H. Bujard, and C. Gatz. 1994. A
chimeric transactivator allows tetracycline-responsive gene expression in whole
plants. Plant J 5:559-69.

Werstuck, G., and J. P. Capone. 1989. Mutational analysis of the herpes
simplex virus trans- inducing factor Vmw65. Gene 75:213-24.

Wieczorek, E., M. Brand, X. Jacq, and L. Tora. 1998. Function of TAF(II)-
containing complex without TBP in transcription by RNA polymerase 11. Nature
393: 187-91.

Willy, P. J ., R. Kobayashi, and J. T. Kadonaga. 2000. A basal transcription
factor that activates or represses transcription. Science 290:982-5.

Wolfe, S. A., E. I. Ramm, and C. O. Pabo. 2000. Combining structure-based
design with phage display to create new Cys(2)His(2) zinc finger dimers.
Structure Fold Des 8:739-50.

Wolstein, 0., A. Silkov, M. Revach, and R. Dikstein. 2000. Specific interaction
of TAF 11105 with OCA-B is involved in activation of octamer-dependent
transcription. 1 Biol Chem 275:16459-65.

Xiao, H., A. Pearson, B. Coulombe, R. Truant, S. Zhang, J. L. Regier, S. J.
Triezenberg, D. Reinberg, O. Flores, C. J. Ingles, and et a1. 1994. Binding of

basal transcription factor TFIIH to the acidic activation domains of VP16 and p53.
Mol Cell Biol 14:7013-24.

Xie, Y., C. Denison, S. H. Yang, D. A. Fancy, and T. Kodadek. 2000.
Biochemical characterization of the TATA-binding protein-Gal4 activation
domain complex. J Biol Chem 275:31914-20.

Xie, Y., L. Sun, and T. Kodadek. 2000. TATA-binding protein and the 0314
transactivator do not bind to promoters cooperatively. J Biol Chem 275:40797-
803.

Yang, F., R. DeBeaumont, S. Zhou, and A. M. Naar. 2004. The activator-
recruited cofactor/Mediator coactivator subunit ARC92 is a functionally
important target of the VP16 transcriptional activator. Proc Natl Acad Sci U S A
101:2339-44.

Yankulov, K., J. Blau, T. Purton, S. Roberts, and D. L. Bentley. 1994.
Transcriptional elongation by RNA polymerase II is stimulated by transactivators.
Cell 77:749-59.

Zhang, L., S. K. Spratt, Q. Lin, B. Johnstone, H. Qi, E. E. Raschke, A. C.
Jamieson, E. J. Rebar, A. P. Wolffe, and C. C. Case. 2000. Synthetic zinc
finger transcription factor action at an endogenous chromosomal site. Activation
of the human erythropoietin gene. .1 Biol Chem 275:33850-60.

164

 

_‘

   

1111111111111111111111111191111111