PE
.. .2.
‘ ...

7n»

3 4.. 1.
flannel}: r

‘ ,
u s ,
“Jain. ,

i.,..;,..,. . .‘..
I. 5.1.1. fi‘ . A .. .A_\ . 1.x.

 

3 0‘33

This is to certify that the
dissertation entitled

X-RAY CRYSTALLOGRAPHIC STUDIES OF RNA
POLYMERASE Ill TRANSCRIPTION FACTOR TFIIIB AND
1L-MYO-INOSITOL 1-PHOSPHATE SYNTHASE

presented by

Xiangshu Jin

has been accepted towards fulfillment
of the requirements for the

 

 

flm ")7 am

Major 7rofessor’ s Signatuue'

(1/7/0;

Date

 

MSU is an Affirmative Action/Equal Opportunity Institution

 

LIBRARY
Michigan State
University

 

 

 

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

 

DATE DUE DATE DUE DATE DUE

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

6/01 c:/ClRC/DateDue.p65-p. 15

X-RAY CRYSTALLOGRAPHIC STUDIES OF RNA POLYMERASE III
TRANSCRIPTION FACTOR TFIIIB AND lL-MYO-INOSITOL l-PHOSPHATE
SYNTHASE
By

Xiangshu J in

A DISSERTATION

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

DOCTOR OF PHILOSOPHY
Department of Chemistry

2002

ABSTRACT
X-RAY CRYSTALLOGRAPHIC STUDIES OF RNA POLYMERASE III
TRANSCRIPTION FACTOR TFIIIB AND lL-MYO-INOSITOL 1-PHOSPHATE
SYNTHASE
By

Xiangshu Jin

TFIIIB is the central initiation factor in the RNA polymerase III transcription
system. Other transcription factors of RNA polymerase 111 such as TFIIIA, TFIIIC, and
SNAPc act as assembly factors to recruit TFIIIB to promoters. TFIIIB then allows the
recruitment Of RNA polymerase III for transcription. In order to investigate the
transcription initiation mechanism of RNA polymerase III, three-dimensional structures
Of TFIIIB-DNA complexes are essential. Toward this goal, recombinant mutants of
individual members Of TFIIIB, namely BRF, B”, and TBP, containing functional domains
were designed, expressed in E.coli cells, and purified to homogeneity for crystallization.
Various promoter DNAs with different lengths were also designed and purified. BRF
functional domains, S. cereviaise BRF435-531, S. cereviaise BRF435-596 (A 538-550),
K. lactis BRF395-501 have been crystallized. Crystallization of BRF/TBP/DNA
complexes was carried out extensively and crystals of 11 different BRF/TBP/DNA
complexes were Obtained.

lL-myo-inositol l-phosphate (MIP) synthase catalyzes the isomerization of D-
glucose 6-phosphate to lL-myo-inositol 1-phosphate, the first committed and rate-

limiting step during the de novo biosynthesis of inositOl-containing compounds important

in signal transduction, cell wall biogenesis, etc. NAD+ serves as a cocatalyst during the
reaction. In order to investigate the mechanism of MIP synthase, structures of MIP
synthase in its apo form, NAD*-bound form, NADH-bound form, and in complex with
NAD+ and a high affinity inhibitor, 2-deoxy-D-glucitol 6-( E )-vinylhomophosphonate
were studied by X-ray crystallography. While the active site residues 351-375 were
missing in the apo structure, the fully occupied NAD“ folds a short helix, (113, that
encompasses 351-361, making interactions with it. When the enzyme is coupled with
NADH, several dramatic structural changes were Observed: all active site residues that
were disordered in the apo and NAD*-bound structures are now completely ordered; the
conformation of the N ADH molecule changed Significantly from its position in the
NAD+-bound structure; two small molecules, phosphate and glycerol, are bound in the
enzyme active site mimicking the substrate binding. The unambiguous position Of the
phosphate in the NADH-bound structure differs significantly from the position of the
phosphate in the previously reported structure Of MIP synthase/NADVZ-deoxy-D-glucitol
6-phosphate. In the structure of MIP synthase in complex with NAD+ and an inhibitor, 2-
deoxy-D-glucitol 6-(E)-vinylhomophosphonate, the inhibitor is fully occupied in the
active site Of one Of two molecules in the asymmetric unit. Based on the new structural

data, a mechanism Of MIP synthase was proposed.

To my parents

iv

ACKNOWLEDGMENTS

First of all, I would like to acknowledge my advisor, Prof. James H. Geiger for his
guidance, encouragement, and support throughout the entire course of my graduate study.
I sincerely thank him for all the invaluable experiences that I have had in his lab, they
helped me become a better researcher and a stronger person. I would like to thank our
collaborator, Dr. Steven Hahn, for providing the plasmids of BRF constructs and B" and
carrying out the Gel Shift assays. I also want to thank Prof. John W. Frost and his group
for providing the inhibitor of MIP synthase that was used in this study.

I would like to acknowledge former and current Geiger group members, Dr. Stacy
Hovde, Dr. Marta Abad, Dr. Adam Stein, Paul Booth, Christopher Geiger, Mike Wolf,
Keith F reel, Erika Mathes, Sara Weaver, and Aimee Brooks for their help, support, and
friendship. Ever since I joined the Geiger group, it has been like a family to me; my life
would never have been the same without having such a nice group of people around. I
want to acknowledge Stacy for all of her help and encouragement, her kind care helped
me to go on with courage and strength during my difficult times. Thanks a lot, Stacy! I
also want to thank Marta for patiently helping me out with the basics and giving me
helpful advice for all the days that we shared in the Geiger lab. I will always remember
your cheerful character. I thank Adam for his grounding work on the MIP synthase
project. I want to thank Paul for his great patience in communicating with me and
teaching me the basics in the early days. I will never forget your encouraging smiles. A
big thanks goes to Mike for his help working on the TFIIIB project as well as the
Valproate inhibition assay. Christopher, Sara, and Keith have also worked on the TFIIIB

project, I am very appreciative of their help! I also want to mention Erika. I remember

vividly those days when you started working in the lab, and I am so happy for what you
have accomplished. All members of the Geiger group from past to present that I had the
chance to meet and work with have my sincere gratitude for their friendship, I wish all of
you good luck on your adventures in Crystallography.

Finally, I want to thank my parents for their unconditional love, support, and
understanding. Without the love fi'om my family, which is inseparable by distance, it
would not have been possible for me to overcome a variety of challenges that I have

encountered in my pursuit of academic goals.

vi

TABLE OF CONTENTS
LIST OF TABLES
LIST OF FIGURES
ABBREVIATIONS

CHAPTER I: INTRODUCTION

1.1 TFIIIB
1.1.1 Transcription
1.1.2 Transcription by RNA polymerase III

1.2 1L-myo-inositol l-phosphate synthase
1.2.1 Cell communication and the importance of inositol
1.2.2 lL-myo-inositol l-phosphate synthase
1.2.3 Previous structural investigations on MIP synthase
1.2.4 Remaining challenges of MIP synthase

1.3 Literature Cited

CHAPTER II: CRYSTALLIZATION AND PRELIMINARY X-RAY
DIFFRACTION ANALYSIS OF THE TFIIIB COMPLEX
2.1 BRF
2.1.1 Over-expression and purification of BRF
2.1.2 Crystallization and data collection of BRF
2.2 B”
2.2.1 Over-expression and purification of B"
2.2.2 Refolding and purification of the BRF/B" complex
2.2.3 Crystallization of the BRF/B" complex
2.3 TBP
2.3.1 Over-expression and purification of wild-type TBP
2.3.2 Proteolysis of wild-type TBP
2.3.3 Over-expression and purification of the deletion mutant TBP 56-240
2.3.4 Proteolysis of the deletion mutant TBP 56-240
2.4 The BRF/TBP/DNA and BRF/B”/TBP/DNA complexes
2.4.1 DNAs
2.4.2 Crystallization of the BRF/TBP/DNA and BRF/ B" /TBP/DNA
complexes
2.5 Discussion
2.6 Literature cited

Chapter III: THE THREE DIMENSIONAL STRUCTURES OF lL-MYO-
INOSITOL l-PHOSPHATE SYNTHASE
3.1 Theory

3.1.1 Structure determination from X-ray diffraction data

3. 1 .2 Molecular replacement

3.1.3 Structure refinement

vii

ix

xi

XX

15
15
21
27
32
41

47
49
53
57
57
60
61
61
61
63
65
66
68
73
75

82
86

88
88
90
92

3.2 Experimental procedures
3.2.1 Protein over-expression and purification
3.2.2 Crystallization
3.3 Data collection and structure refinement
3.3.1 The apo MIP synthase
3.3.2 The MIP synthase/NAD+ complex
3.3.3 The MIP synthase/NADH complex
3.3.4 The MIP synthase/NADH/EDTA complex
3.3.5 The MIP synthase/NAD+(NADH)/2-deoxy-D-glucitol 6-(E)-
vinylhomophosphonate
3.4 Structure of the apo MIP synthase
3.5 Structure of the MIP synthase/NAD+ complex
3.5.1 Structure of the MIP symthase/NAD+ complex tetramer
3.5.2 Structure of the MIP symthase/NAD+ complex monomer
3.5.3 NAD+ binding
3.5.4 Electrostatic charge distribution
3.5.5 The exception of the molecule C in the P21 structure of the MIP
synthase/NAD+ complex
3.6 Structure of the MIP synthase/NADH complex
3.6.1 Overall structure of the MIP synthase/NADH complex
3.6.2 Conformation of NADH
3.6.3 The MIP synthase active site
3.7 Structure of the MIP synthase/NADI/2-deoxy-D-glucitol 6-(E)-
vinylhomophosphonate complex

3.7.1 Overall structure of the MIP synthase/NADVZ—deoxy-D-glucitol 6-(.E)-

vinylhomophosphonate complex
3.7.2 Conformation of NAD+
3.7.3 Structure of the inhibitor 2-deoxy-D-glucitol 6-(E)-
vinylhomophosphonate
3.7.4 Modeling of the substrate and reaction intermediates
3.8 Proposed mechanism Of MIP synthase
3.9 Conclusions
3.10 Literature cited

viii

95
95
96
101
101
104
109
112
115

119
122
122
124
126
134
136

137
I37
140
146
154

155

157
159

165
172
176
I80

LIST OF TABLES

CHAPTER II: CRYSTALLIZATION AND PRELIMINARY X-RAY DIFFRACTION
ANALYSIS OF THE TFIIIB COMPLEX

Table 2.1

Table 2.2

Table 2.3

Table 2.4

Table 2.5

Table 2.6

Table 2.7

BRF crystals

Data collection statistics of S. cerevisiae BRF 435-531 crystal
The BRF/B" complexes used in crystallization attempts

DNAS utilized in crystallization of the BRF/TBP/DNA complexes

DNAS utilized in crystallization of the BRF/ B"/TBP/DNA
complexes

The BRF/TBP/DNA complexes tested for crystallization so far

Crystals of the BRF/TBP/DNA complexes grown so far

55

56

61

74

75

77

79

CHAPTER III: X-RAY CRYSTALLOGRAPHIC STRUCTURAL STUDIES OF 1L-
M YO-INOSITOL l-PHOSPHATE SYNTHASE

Table 3.1

Table 3.2

Table 3.3

Table 3.4

Table 3.5

Table 3.6

Table 3.7

Table 3.8

Crystals of MIP synthase with cofactors and ligands
Data collection and refinement statistics of the apo MIP synthase

Data collection and refinement statistics of the MIP synthase/NADI
complexes

Contents of the MIP synthase/NADI complex structures

Data collection and refinement statistics of the MIP
synthase/NADH complex

Data collection and refinement statistics of the MIP
synthase/NADH/EDTA complex

Data collected on the MIP synthase/NADYNADH)/2-deoxy-D-
glucitol 6-(E)-vinylhomophosphonate complex crystals

Data collection and refinement statistics of the MIP
synthase/NADVZ-deoxy-D-gIucitol 6-(E)-vinylhomophosphonate
complex

98

102

105

106

110

113

116

117

Table 3.9 Interactions between the MIP synthase and NAD+ 128
Table 3.10 Dimerization interactions by the first insertion (93-140) 130

Table 3.1 1 Interactions between MIP synthase and the inhibitor 2-deoxy-D- 162
glucitol 6-(E)-viny1homophosphonate within a 3.5 A cutoff.

LIST OF FIGURES

Images in this dissertation are presented in color.
CHAPTER I: INTRODUCTION

Figure 1.1

Figure 1.2

Figure 1.3

Figure 1.4

Figure 1.5

Figure 1.6

Figure 1.7

Figure 1.8
Figure 1.9

Figure 1.10

Figure 1.11

Diagram of the three types of RNA polymerase 111 genes. +1
indicates the start site of transcription; Tn indicates the termination
site of transcription.

A model of TFIIIC-dependent TFIIIB binding to tRN A promoter.

Sequence alignment of BRF from S. cerevisiae, K. lactis, C.

albican, Arabidopsis thaliana and human. Identical residues are
highlighted in yellow.

Sequence alignment between S. cerevisiae TFIIB and the N-
terrninal 320 amino acid residues of BRF, identical residues are
highlighted in yellow.

Schematic representation of S. cerevisiae BRF. Rose, TFIIB
homology regions; blue, regions that are conserved among other
species.

The amino acid sequence of S. cerevisiae B". Blue, 329-357,
glutamate rich domain; rose, 415-472, SANT domain.

Regions of similarity in the human and S. cerevisiae B" sequences.
The rose boxes correspond to the SANT domain; the hatched
boxes indicate regions of lower but still significant similarity on
either side of the SANT domain. The percentages indicate
identities between the two proteins in the regions delimited by the
dotted lines. The small arrows indicate the repeats in human B”.

CAMP cascade.

Phosphoinositide cascade.

Sequence alignment of MIP synthase from Saccharomyces
cerevisiae, Arabidopsis thaliana, human, Mycobacterium
tuberculosis, and Archaeoglobusfulgidus. Identical residues are

highlighted in yellow.

Proposed mechanism of MIP synthase.

xi

13

13

18

19

23

24

Figure 1.12
Figure 1.13

Figure 1.14

Figure 1.15

Figure 1.16

Figure 1.17

Figure 1.18

Figure 1.19

Figure 1.20

Figure 1.21

Figure 1.22

Figure 1.23

The mechanism of a type I aldolase.
The mechanism of a type II aldolase.

(A) Ribbon model of the MIP synthase monomer. Red, the N-
terminal region; purple, the NADI-binding region; green, the
tetramerization region; blue, the C-terminal domain. (B) Ribbon
model of 2-deoxy-D-glucitol 6-phosphate bound MIP synthase.
Green, the residues that were ordered in the structure with low
occupancy NAD+; red, the newly ordered residues; yellow, NADI;
magenta, 2-deoxy-D-glucitol 6-phosphate.

Proposed mechanism for the transformation catalyzed by MIP
synthase.

The structure of MIP synthase from Mycobacterium tuberculosis.

Cyclic analogues of D-glucose 6—phosphate with no inhibition of
MIP synthase.

Cyclic analogues of D-glucose 6-phosphate that can undergo ring
opening with inhibition of MIP synthase.

The intermediate D, myo-2-inosose l—phosphate, and its analogues
are inhibitors of MIP synthase.

Acyclic analogues of D-glucose 6-phosphate are inhibitors of MIP
synthase.

The mechanism of MIP synthase proposed by Floss.
The mechanism consistent with the substrate binding in a transoid
conformation and the phosphate monoester acting as the base at

the enolization step.

Analogues Of reaction intermediates of the early steps during the
catalysis.

xii

26

26

28

30

31

35

35

36

36

38

39

40

CHAPTER II: CRYSTALLIZATION AND PRELIMINARY X-RAY DIFFRACTION
ANALYSIS OF THE TFIIIB COMPLEX

Figure 2.1

Figure 2.2

Figure 2.3

Figure 2.4
Figure 2.5

Figure 2.6

Figure 2.7

Figure 2.8

Figure 2.9

Figure 2.10

BRF constructs used in this study. 48

SDS-PAGE of Ni-NTA affinity purified K. lactis BRF 395-501. 50
Marker; 2. Flow through; 3. Wash A; 4. Wash B; 5. Wash C; 6-10.
Elutions in 5 fractions in order. M.W. marker: Purple 42,000,

Orange 32,000, Red 17,900, and Blue 7,200.

Over-expressed and purified BRF constructs used in 52
crystallization. (A) 1. MW marker 2. S. cerevisiae BRF311-596

3. S. cerevisiae BRF407-53l 4. S. cerevisiae BRF435-596 (A530-

558) 5. S. cerevisiae BRF435-531 6. S. cerevisiae BRF435-551 7.

S. cerevisiae BRF 420-551. (B) 1. MW marker 2. K. lactis

BRF302-556 3. K. lactis BRF395-501 4. K. lactis BRF302-501 5.

K. lactis BRF395-556.

M.W. marker: Purple 42,000, Orange 32,000, Red 17,900, and

Blue 7,200.

Hanging drop vapor diffusion crystallization method. 54
Batch crystallization method. 54
A crystal of K. lactis BRF 395-501, with dimensions of 0.08 x 55

0.08 x 0.005 mm3.

SDS-PAGE of S. cerevisiae B" 240-520 after Ni-NTA affinity 59

purification; the presence of degradation products and impurities
requires further purification.

SDS-PAGE of the BRF/B" complex purified by F PLC. 60

M.W. marker: Purple 42,000, Orange 32,000, Red 17,900, and
Blue 7,200.

SDS-PAGE of the proteolysis of TBP. 1. Marker; 2. Before the 64
proteolysis (at the time point zero); 3. After 2 hours of reaction at
room temperature; 4. After 12 hours at 4°C (the stop point).

Monitored proteolysis of TBP by Biotinylated Thrombin. 1. M.W. 67
Marker. Purple 42,000, Orange 32,000, Red 17,900, and Blue

7,200; 2. After 12 hours of reaction when 1.5 units thrombin/mg

of TBP was used; 3. After 12 hours of reaction when 0.75 unit

thrombin /mg of TBP was used; 4. After 12 hours when no

thrombin was used; 5. After 18 hours of reaction when 1.5 units
thrombin/mg of TBP was used; 6. After 18 hours of reaction

xiii

Figure 2.11

Figure 2.12

Figure 2.13

Figure 2.14

Figure 2.15

when 0.75 unit thrombin/mg of TBP was used; 7. After 18 hours
when no thrombin was used.

Gel mobility shift assays usingl ng S. cerevisiae TBP (56-240),
and indicated amounts of BRF and B". The probe is the U6
promoter with its TATA box modified to that of the AdMLP
TATA box (U6-MLP promoter). Binding reactions for 45 min. at
room temperature run on 6% TGOE with no MgOAc gel for 60
min. at 4 °C. (A) Titrations of S. cerevisiae BRF311-596 and S.
cerevisiae BRF407-531. (B) Titrations of WT S. cerevisiae B"
with either 2ng S. cerevisiae BRF311-596 or 0.2 ng S. cerevisiae
BRF407-531.

Gel mobility shift assays of S. cerevisiae BRF constructs using]
ng S. cerevisiae TBP56-240, and indicated amounts of BRF. The
probe is the U6 promoter with its TATA box modified to that of
the AdMLP TATA box (U6-MLP promoter). Binding reactions
for 45 min. at room temperature run on 6% TGOE with no
MgOAc gel for 90 min. at 4 °C.

Titrations of S. cerevisiae BRF407-531 and S. cerevisiae
BRF435-596 (A 530-558).

Titrations of S. cerevisiae BRF407-531 and S. cerevisiae
BRF435-531.

Gel mobility shift assays of K. lactis BRF constructs using 2 ng
S. cerevisiae TBP 56-240, and indicated amounts of BRF. The
probe is the U6 promoter with its TATA box modified to that of
the AdMLP TATA box (U 6-MLP promoter). Binding reactions
for 45 min. at room temperature run on 6% TGOE + 0.2 mM
MgOAc gel for 90 min. at 4 °C.

Gel mobility shift assays of S. cerevisiae B" usinglng TBP 56-
240, 0.2ng BRF, and indicated amounts of B". The probe is the
U6 promoter with its TATA box modified to that of the AdMLP
TATA box (U6-MLP promoter). Binding reactions for 45 min. at
room temperature run on 6% TGOE with no MgOAc gel for 90
min. at 4 °C. (A) Titrations of WT S. cerevisiae B" and S.
cerevisiae B"240-520 with 0.2ng S. cerevisiae BRF435-596 (A
530-558). (B) Titrations Of WT S. cerevisiae B" and S. cerevisiae
B"240-520 with 0.2 ng S. cerevisiae BRF435-531.

SDS-PAGE of BRF/TBP/DNA complexes. (1) S. cerevisiae BRF
435-531/TBP/DNA1 (2) S. cerevisiae BRF 435-531/TBP/DNA2.
(3) S. cerevisiae BRF 435-531/TBP/DNA3. (4) S. cerevisiae BRF
43 5-53 1/TBP/DNA6.

xiv

69

70

71

72

Figure 2.16

Figure 2.17

Figure 2.18

(A) Crystals of the S. cerevisiae BRF 435-531/TBP/DNA1 comlex
grown in 30 % PEG400, 0.2 M CaClz, 0.1 M Na-HEPES, pH 7.5.
(B) Crystals Of the S. cerevisiae BRF 43 5-53 1/TBP/DNA6
complex grown in 30 % PEG 4000, 0.2 M MgClz, 0.1 M Tris, pH
7.5. (C) Crystals of the S. cerevisiae BRF 435-531/TBP/DNA2
complex grown in 30 % PEG 4000, 0.2 M MgClz, 0.1 M Tris, pH
7.5. (D) Crystals of the S. cerevisiae BRF 435-531/TBP/DNA7
complex grown in 30 % PEG 4000, 0.2M MgClz, 0.1 M Tris, pH
7.5.

Sequence alignment of S. cerevisiae BRF C-terminal domain and
TFIIA small subunit. Identical residues are highlighted in yellow.
Boxed in blue are BRF428-438 and TFIIA 67-72; this region of
TFIIA makes interactions with the TBP/DNA.

Modeled structure of the TFIIA/TFIIB/TBP/DNA complex.

TFIIA small subunit is in yellow, TFIIA large subunit is blue, TBP
is in cyan, TFIIB is in green, and DNA is in silver. The presumed
folding of BRF 1-286 would be somewhat similar to TFIIB; the
presumed protein-protein and protein-DNA contacts of BRF 428-
438 would be similar to those of [31 of TFIIA small subunit; the
presumed folding of BRF 564-5 85 would mimic [33 of TFIIA
small subunit.

81.

83

85

CHAPTER III: X-RAY CRYSTALLOGRAPHIC STRUCTURAL STUDIES OF 1L-
M YO-INOSITOL l-PHOSPHATE SYNTHASE

Figure 3.1

Figure 3.2

Figure 3.3

Figure 3.4

Figure 3.5

Figure 3.6

Ramachandran plot of the apo MIP synthase structure.

Ramachandran plot of the P21crystal form of the MIP
synthase/NAD+ complex structure.

Ramachandran plot of the C2 crystal form of the MIP
synthase/NAD+ complex structure.

Ramachandran plot of the MIP synthase/NADH complex
structure.

Ramachandran plot of the MIP synthase/NADH/EDTA complex
structure.

Ramachandran plot of the MIP synthaselNADVZ-deoxy-D-
glucitol 6-(E)-vinylhomophosphonate structure.

XV

103

107

108

111

114

118

Figure 3.7

Figure 3.8

Figure 3.9

Figure 3.10

Figure 3.11

Figure 3.12

Figure 3.13

Figure 3.14

Figure 3.15

Figure 3.16

Structure of the apo MIP synthase. The region that was disordered
in the previous low occupancy NAD+- bound structure is colored
red.

Superposition of the structures of apo MIP synthase in gold and
the low occupancy NAD+-bound MIP synthase in blue. Boxed in
red are residues 191-198.

The entire tetramer of the P21 structure of the MIP synthase/NAD+
complex. Molecule A is colored in cyan, B is in magenta, C is in
gold, and D is in green.

A monomer of the MIP synthase/NAD+ complex. The helix in red
is the newly ordered region upon NADI binding; NAD+ is in
lavender. The rest Of the or helices in the structure are cyan; all [3
strands are green; the loops are brown.

The GXGGXXG phosphate-binding motif within the MIP
synthase Rossmann fold domain.

Interactions between the adenine portion of NAD+ and MIP
synthase.

The insertion encompassing residues 247-276 is involved in the
interaction with a neighboring tetramer. (A) Two neighboring
tetramers. Lavender, molecule A; green, molecule B; gold,
molecule C; cyan, molecule D. (B) The insertion in molecule C
makes interactions with the loop 464-472 of molecule A in the
neighboring tetramer.

(A) Space-filling model of the NAD+-bound MIP synthase, the
nicotinamide is exposed as shown in atom-color. (B) The tight
hydrogen bond between the nicotinamide nitrogen and the
phosphodiester oxygen stabilizes this conformation.

The electrostatic potential surface of MIP synthase with a view of
looking into the NAD+ binding surface. Blue, EPS> 6kcal/mol;
red, EPS<-6 kcal/mol; white, EPS~0. The dotted circle denotes the
nicotinamide-binding site; the dotted square denotes the potential
substrate-binding site. For clarity, the residues 181-201, 351-361
were removed

(A) An example of the ZFO-Fc electron density of MIP
synthase/NADH structure. (B) Superposition of the structures of
apo MIP synthase and the NADH-bound MIP synthase. The apo
MIP synthase is in gold and the NADH-bound structure is in blue.

xvi

120

121

123

125

129

131

132

133

135

139

Figure 3.17

Figure 3.18

Figure 3.19

Figure 3.20

Figure 3.21

Figure 3.22

Figure 3.23

Figure 3.24

(A) Overlay of cofactors in the NAD+-bound structure in yellow
and the NADH-bound structure in cyan. (B) The 2Fo-Fc electron
density map of the MIP synthase/NADH complex structure
contoured at 2.40 around the NADH.

(A) Simulated annealing omit electron density map contoured at
50 around NADH. (B) Simulated annealing omit electron density
map of MIP synthase/NADH/EDTA contoured at 1.8 O.

(A) The NADH-bound MIP synthase from S. cerevisiae. Cyan, the
modeled putative divalent cation; Orange, the fourth ligand of the
putative divalent; Red, the conserved water molecule. (B) The
NAD+-bound MIP synthase from M. tuberculosis. Zn2+ is in cyan.

7O simulated annealing omit map of MIP synthase/NADH active
site with a phosphate and glycerol modeled in.

Conformational change of the phosphate-binding loop in the
NADH-bound structure in cyan compared with the NAD+-bound
structure in silver. Note: The conformation of this loop in the
structure of the apo MIP synthase, the low occupancy NAD+-
bound MIP synthase, and the MIP synthase/NADVZ-deoxy-D-
glucitol 6-phosphate complex are all identical to that of the NAD+-
bound structure.

Interactions in the enzyme active site observed from the MIP
synthase/NADH complex structure. The phosphate and glycerol
are in gold, the putative divalent cation is in cyan. The fourth
coordination ligand of the putative divalent cation, which was
modeled as a water molecule in the MIP synthase/NADH complex
structure, is in yellow.

Overlay of the S. cerevisiae MIP synthase/NADH complex
structure in cyan and the M. tuberculosis MIP synthase/NAD+
complex structure in yellow. In the dotted box are the phosphate-
binding loops.

Overlay of the MIP synthase/NADH complex structure in silver
and the MIP synthase/NADVZ-deoxy-D-glucitol 6-phosphate
complex structure in lavender. Gold, the phosphate and glycerol
in the NADH-bound structure; blue, 2-deoxy-D-g1ucitol 6-
phosphate in the previous inhibitor-bound structure; yellow, side
chains of the NADH-bound structure; cyan, sides chains of the
previous inhibitor-bound structure.

xvii

142

143

145

147

148

151

152

153

Figure 3.25

Figure 3.26

Figure 3.27

Figure 3.28

Figure 3.29

Figure 3.30

Figure 3.31

Figure 3.32

Overlay of the MIP synthase/NADVZ-deoxy-D-glucitol 6-(E)- 156
vinylhomophosphonate complex structure in blue with the MIP
synthase/NADH complex structure in gold.

Overlay of cofactors in the MIP synthase/NADVZ-deoxy-D- 158
glucitol 6-(E)-vinylhomophosphonate complex structure in cyan

and the MIP synthase/NADH complex structure in gold. The

putative divalent cation is in aqua in the NADH-bound structure

and in cyan in the 2-deoxy-D-glucitol 6-(E)-
vinylhomophosphonate-bound structure. Shown in orange is the

fourth ligand of the putative divalent cation in the NADH-bound

structure. The third ligand water molecule is located at an

identical position in both structures.

(A) Simulated annealing omit map of the MIP synthase/NAD+/2- 160
deoxy-D-glucitol 6-(E)-vinylhomophosphonate structure

contoured at 1.6 O. (B) 2Fo-Fc electron density map of the MIP
synthase/NADI/Z-deoxy-D-glucitol 6-(E)-vinylhomophosphonate
structure contoured at 1.2 O.

(A) Interactions between MIP synthase active site residues and the 161
inhibitor 2-deoxy-D-glucitol 6-(E)-vinylhomophosphonate; the

inhibitor is in lavender. (B) Overlay of 2-deoxy-D-glucitol 6-(E)-
vinylhomophosphonate in gold and phosphate and glycerol in the
NADH-bound structure in cyan. (C) Overlay of 2-deoxy-D-

glucitol 6-(E)-vinylhomophosphonate in gold and 2-deoxy-D-

glucitol 6-phosphate in the previously published inhibitor structure

in blue.

GRASP drawing of the electrostatic potential surface of MIP 164
synthase. Circled dotted is the substrate-binding site centered at

the phosphate-binding pocket. Blue, EPS> 6kcal/mol; red, EPS<-6
kcal/mol; white, EPS~0. For clarity, residues 190-200, 351-366

were removed in this figure.

Modeling of the substrate D-glucose 6-phosphate (yellow) based 166
on the structure of the inhibitor 2-deoxy-D-glucitol 6-(E)-
vinylhomophosphonate (lavender). The putative divalent cation is

in cyan.

Interactions between MIP synthase and the modeled substrate D- 167
glucose 6-phosphate.

Modeling of the reaction intermediate 5-ketO-D-glucose 6- 169

phosphate. The substrate D-glucose 6-phosphate is in yellow, and
the 5-keto-D-glucose 6-phosphate (and the enolate) is in blue.

xviii

Figure 3.33 Interactions between the modeled 5-keto-D-glucose 6-phosphate 170
(enolate) and the active Site residues.

Figure 3.34 Interactions between the modeled myo-2-inosose 1-phosphate and 171
the active site residues.

Figure 3.35 Proposed mechanism of MIP synthase where the phosphate 174
monoester acts as the base at the enolization step

Figure 3.36 An alternative mechanism of MIP synthase where K489 acts as the 175
base at the enolization step.

xix

LIST OF ABBREVIATIONS

A — alanine

AC — adenylate cyclase

ADA2 — adenosine deaminase 2

A. fulgidus — A rchaeglobus fulgidus
AMP - adenosine monophosphate
ATP — adenosine triphosphate

BM — bending magnet
BME - B—mercaptoethanol
bp — base pair

BRF — TFIIB-related factor

C -— cysteine

C/CYT — cytosine

Ca - the alpha carbon

CDP — cytodine diphosphate

CNS - Crystallography & NMR System
C terminal - carboxy terminal

D - aspartic acid

DAG - diacylglycerol

DNA — deoxyribonucleic acid

DPE — downstream promoter element
DSE —distal sequence element

DTT - dithiothreitol

E - glutamic acid

E. coli - Escherichia coli

EDTA —— ethylenediamine tetraacetic acid
EMSA — electrophoretic mobility shift assay
EPS - electrostatic potential surface

F - phenylalanine

F ca. - calculated structure factors

Fob, - observed structure factors

FPLC - fast pressure liquid chromatography

G - glycine

G/GUA — guanine

GDP — guanine diphosphate

GTP - guanine triphosphate

G-protein - guanine nucleotide binding protein
Gu - guanidine

XX

H/His - histidine
HEPES - N-[2-hydroxyethyl] piperazine-N’-[ethane sulfonic acid]
HPLC — high pressure liquid chromatography

I - isoleucine

I — intensity

ID — insertion device

IMCA-CAT — Industrial Macromolecular Crystallography Association Collaborative
Access Team

1 ,4,5-IP3 — myo-inositol-1,4,5-triphosphate

1PTG — isopropyl-B-D-thiogalactoside

K - lysine
K. lactis — Kluyveromyces lactis

L — leucine
LB — Luria-Bertani

M - methionine

MI — myo-inositol

MIP — myo-inositol l-phosphate

MPD - 2-Methyl-2,4-pentanediol

M.W. — Molecular weight

M. tuberculosis — Mycobacterium tuberculosis

N — asparagine

NAD+ - nicotinamide adenine dinucleotide

NADH - nicotinamide adenine dinucleotide, reduced form
N-cor — nuclear receptor co-repressor

O.D. — Optical density

P — proline

PEG - polyethylene glycol

PDB — protein data bank

PKA — protein kinase A

PLC — phospholipase C

PMSF — phenyl methyl sulfonyl fluoride

POL — RNA polymerase

PSE — proximal sequence element

PtdIns — phosphatidyl inositol

PtdIns-(4,5)-P2 - phosphatidyl inositol-4,5-bisphosphate

Q — glutamine

xxi

R - arginine

RNA — ribonucleic acid

rmsd - root mean square deviation
rRNA — ribosomal RNA

S — serine

SANT — SW13, ADA2, N-cor, TFIIIB

S. cerevisiae — Sacchromyces cerevisiae

SDS-PAGE — sodium dodecyl sulfate—polyacrylamide gel electrophoresis
SNAPc — small nuclear RNA activating protein complex

snRNA — small nuclear RNA
SW13 — SWItch gene product 3

T - threonine

T/THY — thymine

TAF — TBP associated factor

TBP - TATA binding protein

TFB — transcription factor B
TFIIA — transcription factor IIA
TFIIB — transcription factor IIB
TFIIIA — transcriptionm factor IIIA
TFIIIB — transcriptionm factor IIIB
TFIIIC - transcriptionm factor IIIC
TRFl — TBP related factor 1

Tris — 2-amino-2-(hydroxymethyl)-l ,3-propanediol

tRNA — transfer RNA

U — uracil

U6-MLP — U6-major late promoter
UBF - upstream binding factor

V - valine

W - tryptophan
WT - wild type

Y - tyrosine

xxii

CHAPTER I
INTRODUCTION
1.1 TFIIIB
1.1.1 TRANSCRIPTION

Transcription is the process by which the genetic message in DNA is transcribed
into RNA. Eukaryotic cells contain three distinct transcription systems, each consisting
of one of three RNA polymerases and a unique set of transcription factors. Accurate and
efficient transcription requires RNA polymerase in association with several transcription
factors to be properly located at a specific start sequence of DNA called a promoter.

RNA polymerase I (P01 1) transcribes ribosomal RNA (rRNA) genes. Accurate
initiation of transcription by RNA polymerase I requires the upstream binding factor
(UBF) and the selectivity factor SL-l complex, which includes the ubiquitous TATA
binding protein (TBP) in addition to several other novel factors unique to this complex
(1-3).

RNA polymerase 11 (Pol II) transcribes the protein-coding messenger RNA
(mRNA) genes and some small nuclear RNA (snRNA) genes. RNA polymerase 11
transcription is the most complex among the three transcription systems, and is by far the
most studied system. All of the RNA polymerase II transcription initiation factors have
been identified, and much is known about how these factors associate at the core
promoter to recruit RNA polymerase II to initiate transcription (4,5). TFIID is the only
basal initiation factor capable of specifically recognizing a core promoter element and it
can do so by binding to one or more of the core promoter DNA sequences (6). TFIID is a

large multi-protein complex consisting of TBP and several ancillary TBP associated

factors (TAFs) (7-10). While TBP specifically recognizes the TATA box core promoter
element, other TAFs can recognize the initiation core element or the downstream
promoter element (DPE). TBP is the central factor in this complex and its recognition of
the TATA element is thought to be the most general and critical of core promoter
recognition events for most RNA polymerase II promoters (11,12).

Though each of the three eukaryotic RNA polymerases uses distinct sets of
transcription initiation factors, all three RNA polymerases share some similarities. First
the RNA polymerases have five shared subunits and several other homologous
subunits(13). Second, all three polymerases require TBP both in vitro and in vivo for
initiation, even though TBP is contained in distinct complexes specific to each
polymerase(14). Third, both RNA polymerase II and RNA polymerase III transcription

systems have a TFIIB-like factor required for transcription initiation (15).

1.1.2 Transcription by RNA polymerase 111
1.1.2.1. Promoters of RNA polymerase III-transcribed genes

The genes transcribed by RNA polymerase 111 (Pol III) encode a variety of small
RNA molecules. Many of these have essential functions in cellular metabolism: tRNA
and SS rRNA are required for protein synthesis; 7SL RNA is involved in intracellular
protein transport; U6, H1, and MRP RNAs are involved in posttranscriptional processing;
VA RNAs of adenovirus serve to divert the translational machinery Of an infected cell
towards the more effective production of viral proteins (16). RNA polymerase III also
transcribes genes with no known function, for example, the 78K genes and the short

interspersed repeat (SINE) (17.18). The distinguishing feature of RNA polymerase III

promoters as a class is the preponderance of gene-internal (transcribed) promoter
elements. The promoters of RNA polymerase 111 genes include discontinuous intragenic
structures that are composed of essential sequence blocks separated by nonessential
nucleotides. It is convenient to distinguish three classes of these promoters (Figure 1.1):
type I genes (5S rRNA genes), type 11 genes (tRNA and 7SL RNA genes), and type 111
genes (U6 small nuclear RNA and 7SK small cytoplasmic RNA genes)(16).

Type I gene promoters:

The principal feature of type I genes is a binding site for TFIIIA. The mode of
DNA binding by TFIIIA is best understood for the Xenopus protein, which interacts
sequence-specifically with three intragenic sites, encompassing a span of approximately
50 base pairs: the transcription start-proximal box A, the intermediate element (ie), and
box C.

Type II gene promoters:

Type II gene promoters have highly conserved internal promoter elements, box A
and box B, that constitute the noncontiguous binding site of TFIIIC. The box A-box B
separation is variable over a wide range, but separation of 30-60 bp is Optimal for
concurrent occupancy of both sites by TFIIIC.

Type III gene promoters:

The promoters of type 111 genes have exclusively external elements, such as the
entirely upstream-located promoters of the vertebrate U6 snRNA genes. The upstream
promoter elements TATA box, Proximal Sequence Element (PSE), and Distal Sequence
Element (DSE) play a dominant role in type 111 genes. The organization of these

promoters is particularly important because they contain dual function-Pol II/III elements

and they determine distinctive pathways for assembling transcription initiation
complexes.

The intragenic sequences of type 111 genes are highly conserved between different
genes and different species. This reflects a strong conservation of the general
transcription factors employed by RNA polymerase III. The flanking sequences of type I
and II promoters Show little or no conservation, this suggests that these sequences are
more likely to be recognized by gene- or species-specific factors or that their cognate
factors have very flexible DNA binding specificities (16).
1.1.2.2 Transcription factors utilized by RNA polymerase III

Purified RNA polymerase III initiates transcription randomly. Accurate and
specific initiation by RNA polymerase 111 requires the assistance of transcription factors
in order to recruit the polymerase to the appropriate start sites of genes. RNA polymerase
III transcription factors include both general factors and gene-specific factors. General
factors include TFIIIB, TFIIIC, SNAPc etc., and gene-specific factors include TFIIIA,
PSE-binding factor, OCT-1, etc. (19,20). Among all RNA polymerase III transcription
factors, TFIIIB is the central initiation factor since it alone can recruit the polymerase and
specifies the transcription start site. TFIIIC and TFIIIA, the other components of the core
transcription apparatus, bind DNA and serve as assembly factors for TFIIIB; the six-
subunit TFIIIC complex interacts directly with TFIIIB, and TFIIIA serves as a SS rRN A
gene-specific platform for TFIIIC. TFIIIB is composed of three subunits: TBP, BRF,
and B”. TFIIIB alone usually does not show sequence-specific DNA binding, it must be
recruited onto DNA by interaction with other transcription factors. For instance yeast

TFIIIB can bind to tRNA genes only if TFIIIC is bound (Figure 1.2).

TYPE I PROMOTER
SS rRNA genes

 

r" A

 

 

 

 

 

+1

TYPE ll PROMOTER
e.g. tRNA genes

 

 

r—>

+1

 

A B

 

 

 

 

 

 

TYPE III PROMOTER
e.g. Vertebrate U6 RNA genes

 

 

 

 

 

ie C

 

 

 

 

“i-

Tn

 

... DSE ,1 , PSE

 

 

 

 

 

Tn

 

TATA

 

 

 

 

 

 

+1

Figure 1.1 Diagram of the three types of RNA polymerase 111 genes. +1 indicates the

start site of transcription; Tn indicates the termination site of transcription.

Pol III

   
 
  

A

tRNA gene

   

Figure 1.2 A model of TFIIIC-dependent TFIIIB binding to a tRNA promoter.

It is now well established that TATA-binding protein (TBP) is an essential
component of the transcriptional machinery of all three RNA polymerases, and therefore
is required for the expression of all nuclear genes (21,22). The N-terminal region of TBP
is variable in both size and sequence, but the C-terminal domain is highly conserved.
Mutations affecting basal transcription by RNA polymerases I, II and 111 all map to the C-
terminal domain of TBP. This domain alone has been shown to support transcription by
RNA polymerase 111 (23-26). While it has been supposed that the constitution of TFIIIB
is similar in most other eukaryotes, it appears that TBP is replaced by TRFl (TBP-related
factor 1) in Drosophila melanogaster (27).

fl

BRF is the TFIIB-related factor of the TFIIIB complex. Sequence alignment
between S. cerevisiae, K. lactis, C. albican, Arabidopsis and human BRF shows its strong
homology (Figure 1.3). Although the homology is most pronounced in the amino-
terminal half, conserved regions also exist in the carboxy-terminal half that are unique to
BRF. In S. cerevisiae TFIIIB, BRF is a protein of 596 amino acids, with a molecular
mass of ~67 kD and an isoelectric point (pI) of ~6.9. The amino-terminal half of BRF is
homologous to the Pol II transcription factor IIB (TFIIB) (Figure 1.4). The N-terminal
320 amino acids of BRF are 19% identical to TFIIB, with an overall similarity of 44%.
Three regions are conserved between BRF and the various TFIIB proteins. One consists
of the putative Zn finger near the N-terminus, and the other two imperfect direct repeats
(Figure 1.5). The repeat regions Of BRF are 25% and 28% identical to the corresponding

regions of human TFIIB.

s.cerevisiae
k.1actis
c.a1bican
arabidopsis
human

s.cerevisiae
k.1actis
c.a1bican
arabidopsis
human

s.cerevisiae
k.1actis
c.a1bican
arabidopais
human

s.cerevisiae
k.1actis
c.a1bican
arabidopsis
human

a.cerevisiae
k.1actis
c.albican
arabidopsis
human

s.cerevisiae
k.laccis
c.a1bican
arabidopsia
human

s.cerevisiae
k.1actis
c.a1bican
arabidopsis
human

s.cerevisiae
k.1actis
c.albican
arabidopsis
human

Zn fi nger
r 1
MW -------- CKNCHGTBFERDLSNANNDLVCKACGWSEDNPIVSEVTF‘GETSAGMWQGSPIG-AGQSHAAFn-GGSSAL- -Bsn
HASTLQVSSRKCTKNCG s'rnrvanrsm'ruar. rcx vccnvrs ms 2 vssLArtisxsucMi/tocxr-‘vs-ANQAII P’I‘FMSHSGQNAL— -MSR

 

HSKPRK--QQKCKTCGHTQFDVNRYTAAGDVSCLRCGTVLEENPIVSEVQFGESSSGAAMVOGAMVG-ADQARATFAG‘GRQNAM--ESR
NV -------- WCKHCGKN--VPGIRPYDAALSCDLCGRILENFNFSTEVTFVKNAAGQSQASGNILK ------------ SVQSGMSS-SR
MTG ------ RVC RGCGGTD I ELI) - -MRGDAVCTACG SVLEDN I I VSEVQFVESSGGGSSAVGQFV SLWGXTPTIGCB PHVNLGKESR

Direct repeats

 

EATLNNARRKLRAVSYALHIPEYITDA---AFQWYKLALANNFVQGRRSQNVIASCLYVACRKEKTHHMLIDFSSRLQVSVYSIGATFLK
ETTLNNARRKLKAVSYALNIPEYVTDA---AFQWYRLALSNNFVQGRKSQNVIAACLYIACRKERTHHMLIDFSSRLQVSVYSIGATFLR
EQTLSNGKRKIKRIAAALKIPDYIAEA---AGDHFRLALTLNFVQGRRSNNVLATCLYVACRKERTHHMLIDFSSRLQISVYSLGATFLK
ERIIRKATDELMNLRDALGIGDDRDDVIVMASNFFRIALDHNFTKGRSKELVFSSCLYLTCRQFKLAVLLIDFSSYLRVSVYDLGSVYLQ
AQTLQNGRRHIHHLGNQLQLNQHCLDT---AFNFFKMAVSRHLTRGRKMAHVIAACLYLVCRTBGTPHMLLDLSDLLQVNVYVLGKTPLL

Direct repeats

 

HVKXLHITELP—--~LADPSLFIOHFAEKLDLADKKIKVVKDAVKLAQRMSKDWMFEGRRPAGIAGACILLACRMNNLRRTHTEIVAVSH
LAKKLQIVKLP----LADPSLFIQHFAEKLELGDKKIKVIRDAVKLAQTMSRDWMYEGRRPAGIAGACLLLACRMNNLRRTHSEIVAISH
MVKALHITSLP----LADPSLFIQHFVEKLDFKDKATKVAKDAVKLAHRMAADWIHEGRRPAGIAGACVLLAARMNNFRRSHAEIVAVSH
LCDMLYITENHNYEKLVDPSIFIPRFSNMLLKGAHNNKLVLTATHIIASHKRDWHQTGRKPSGICGAALYTAALSHGIKCSKTDIVNIVH
LARELCINA—P----AIDPCLYIPRFAHLLEFGEKNHEVSHTALRLLQRMKRDWMHTGRRPSGLCGAALLVIARHHDPRRIVXBVISUV!

Ihinnihig\-rcginn I
F_ I
VAEETLQQRLNEFKNTKAAKLSVQKPRENDVE ------ DGBARPPSFV--KNRKKERKIKDSLDKBEHFQTSEEALNKNPILTQVLGEOS
VAEETLQQRLNEFKNTTSAKLSVKEFRDDETBVNEGERSAESKPPSPD-~KNRLKEKKIKDSLDTKEHLETSEEAVSRNPILTQVLGAQD
VGEETLQRRLNEFKKTKAGTLSVKSFRE ------- VENLESSNPPSFE--KNRAMELKISKKLQoooe-TDNFEDLSK ------- "TEE!
ICEATLTKRLIEFGDTEAASLTADELSKTEREKBTAALRSKRKPNFYKEG -------- VVLCMHQDCKPVDYGLCESCYDEFHTVSGGLB
VCESTLRKRLTEFEDTPTSQLTIDEFHKIDLB ------- EECDPPSYTAGQRKLRHKQLEQVLSKI--LBEVIGBISSYQDAIBIBL‘NS

 

LSSKEVLF-YLKQPSERRARVVERIKATNGIDGENIYHE-GSENETRKRKLSEVBIQNBHVEGEDK ------------------ BTEGTB
LSSKEVLY- YLKK LS ERRKAEF‘SH IKATHG I IXIEDLHKT- EKD- - -KKRSLDE -------------------------------------
~--KQSVFGKLSKEEAQKQLLHNTILSDITITTENLNDOMDRILKMKXSSLENSLYKTPYELALAN ----------------- ~GSBQDP
GGSDPPAFQRAEK -------------------- ERMEEKASSEENDKQVNLDGHSDESSTLSDVDDRBSDRPTVSQL -------------
RPKAKGGLASLAK -DGSTEDTASSLCGEEDTED£ mamsumLm EIGGAPGSS MSPMGRPPALGSLLDPLPTAMIBIS

IlnnnflugyicghutH

 

I
EKVKK—VKTKTSEEKKENESGHFQDA ------ IDGYSLETDPYCPRNLHL-LPTTDTYLSKVSDDPDNLEDVDDEELNAHLLNEEASKLK

 

------------------------------- IDGYSLEKDPYRPRNLHL-LPTTASLLSKVSDHPENLDDVDDAELDSHLLDEEASKLK
SKIWN-INK ----------------------------------- PKNLVANLPKTDDILQNVSSEVELNSDDDDEIVLESKLTEEEVAIK
~--~DCYFRTPEEVRLVKIFFDHENPGYDEKEAAKKAAGLNACNNASNIFEASKAAAAKSRXEKRQQRAEEEKNIPPPATGIBAVDSMVE
DSIRECISSQSSDPKDASGDGELDLSGIDDLEIDRYILNESEARVKA-ELWMRENABYLREQRBKEARIAKEKELGIYKRHXPKKSCKRR
I
ERIWIGLNADF---LLEOESKRLKOEADIATGNTSVKKKRTRR ----------- RNNTRSDEPTKTVDAAAAIGLMSDI -------- QDK
ERIWIDINGDY---LIEOESKRLKQEADLASGNTSLRKKRSKR ----------- TNRNQSSASIVKVQVD---GL ---------------
ERIWTGLNHDY---LVEQEKKRLKQEADELTGNTS-KSSSGNR ----------- RKRNKSSLPA --------- SLR!!! ------- -GD-
RKKFRDINCDYLEELFDASVEK ------------------------------- SPKRSKTETVHEK ---------------------- KI

EPIQASTAREAIEKMLEQKKISSKINYSVLRGLSSAGGGSPHRBDAQPBHSASARKLSRRRTPASRSGRDPVTSVGKRLRPLVSTQFAKE
Iliuiuihiu\ usuuin HI

I I

 

SGLHAALKAAEESGDFTTADSVKNMLQKASFSKKINYDA ------------- IDOL ------- F ------ R
-PLDVSVDDAD-AVDVVAAGGVKNLLQKTTFSKRINYDA ------------- INGL ------- FG----QK
------ IDLDEDGTPRSAADSAKHYISKTSVSKKINYDS-------------LKGL------~LG----GNMBP
KEEHEIVENEQEEEDYAAPYEQDEEDYAAPYEMNTDKKFYESEVEEEED ----------- GYDFG ------- LY

VATGEALLPSSPTLGAERARPQAVLVESGPVSYHADEEADEEEPDEEDGEPCVSALQHMGSNDYGCDGDEDDGY

Figure 1.3 Sequence alignment of BRF from S. cerevisiae, K. lactis, C. albican,

Arabidopsis thaliana and human. Identical residues are highlighted in yellow.

76
87
84
67
82

163
174
171
157
169

249
260
257
247
254

331
348
329
329
335

401
396
398
386
424

483
453
452
472
513

551
511
509
509
603

596
556
SS3
565
677

R E S I D K R A G R R G P N L N I V L

B M M T
M P V

TFII
BRF

47

N G D D P S R 78

C A L C G L V L S D K L V D T R S E W R T F S N D D H
C K A C G V V S E D N P I V S E V - - ~ T F G E T S -

340

S F V K N R K K E R K 299

P

345
320

TFIIB
BRF

I K D S L D K E E M F O T S E

Figure1.4 Sequence alignment between S. cerevisiae TFIIB and the N-terminal 320

dentical residues are highlighted in yellow.

(1 residues of BRF, i

3.111an acr

I

BRF homology regions

/

I
m
86

\

direct repeats

TFIIB-like domain

 

Zn linger
l__l
4 28

 

 

 

 

 

 

Figure 1.5 Schematic representation of S. cerevisiae BRF. Rose, TFIIB homology

165.

that are conserved among other spec

, regions

blue

0
9

regions

BRF is highly homologous to TFIIB and Archae TFB over the Zn binding domain
and the core domain (two repeats). However, BRF also contains a ~30 kD domain at its
C-terrninus that is conserved only among other BRFs (Figure 1.3) and appears to play a
major role in interaction with TBP/DNA. Both TFIIB and Archea TFB bind the
TBP/DNA complex tightly (28-31). It was quite plausible to anticipate that TBP would
interact with the homologous domain of BRF and TFIIB, but in fact BRF appears to act
differently from these homologous factors. The region of BRF conserved with the TFIIB
family members does not detectably bind TBP (32,33). On the other hand, the C-
terrninal half of BRF, which is not conserved with the TFIIB family members, strongly
interacts with TBP and DNA, and it alone can form a complex with TBP, B”, and DNA
(34,35). Deleting homology regions 11 and III in the BRF C-terminal region abrogates
this interaction (34,3 5). TBP mutations that selectively inhibit RNA polymerase III
transcription in vivo impair interactions between TBP and the BRF C-terminal domain
(36-38).

An important unanswered question in the Pol 111 system is the role of the N-
terminal TFIIB-homologous region of BRF. This region is essential for transcription and
appears to act differently from its relatives in the Archae and RNA polymerase 11
systems. It has been shown that the N-terminal domain of BRF interacts with 1:131, a
subunit of TFIIIC in two-hybrid assays (32), and with C34, a subunit of RNA polymerase
III in affinity chromatography (33). Previous studies have shown that BRF does not
interact on the same face of the TBP/DNA complex as TFIIB. On the other hand, BRF

may overlap in its position with TFIIA for binding to TBP, DNA, or both (37). The N-

10

terminal domain of BRF is also necessary for TFIIIC-dependent transcription in vitro
(34,3 5).

Extensive deletion mutagenesis of the C-terminal domain of BRF has shown that
a small domain encompassing 435-545 is sufficient for the formation of both the
BRF/TBP/DNA complex and the BRF/B”/TBP/DNA complex. When this domain is
reconstituted with the N-terminal domain of BRF, it recovers almost full wild-type BRF
activity in TFIIIC-independent and TFIIIC-dependent transcription initiation complex
formation, as well as TFIIIC-mediated TATA-less transcription and TFIIIC-independent
TATA-dependent transcription in vitro (34,3 5).

_B_’:

B” is the third subunit of TFIIIB. Since TFIIB and the N-terminal half of BRF sit
in corresponding locations in their respective DNA complexes (30,34), and exercise
similar functions, it is surprising that BRF and TBP are not by themselves competent to
direct transcriptional initiation by RNA polymerase III (37). In fact, B” is absolutely
required for transcription by RNA polymerase III in vivo as well as in vitro in duplex
DNA or chromatin, at TATA-containing and TATA-less promoters. It is also B” that
makes the TFIIIB-DNA complex extraordinarily stable (20). In S. cerevisiae, B” consists
of 594 amino acids (Figure 1.6) and has almost the same molecular weight as BRF
(~67kD), although it migrates anomalously in SDS-PAGE (as a ~90kD protein). The
ready renaturability and anomalous electrophoretic mobility of B” suggests that it may
contain a highly stable core structure in the TFIIIB complex (39). S. cerevisiae 3”
contains a SANT domain (Figure 1.7). The SANT domain is related to a Myb repeat and

was originally identified in a number of proteins, including the SW13, ADA2, N-Cor, and

11

yeast TFIIIB B” proteins (40). In S. cerevisiae B”, C-terminal deletions of B” that lack
most of the SANT domain are inactive for in vitro transcription of a TATA-less tRNA
gene, although they are still active for transcription of the TATA-containing U6 snRNA
gene (41).

B” can associate with a preassembled TBP/BRF/DNA but not with a complex
lacking BRF. B” contacts not only BRF but also TBP. At least one mutation in
S. cerevisiae TBP prevents association of B” without affecting association of BRF (37).
B” also makes interactions with DNA. It has been shown that B” interacts upstream of the
TATA box. B” binding to the TBP/BRF/DNA complex extends the DNA footprint ~10
bp upstream of the TATA box, and is stabilized by an additional 15-20 bp DNA stretch
upstream of the TATA box (3 7). Thus, the DNA segment that is additionally covered
when B” adds to the TBP/BRF/DNA complex harbors a site of efficient B” cross-linking
and contributes to the stability of the TFIIIB-DNA complex (42). The binding of B” to
the TBP/BRF/DNA complex induces a bend in the DNA between the TATA box and the
transcription start site, which is in phase with the bend imposed by TBP on the TATA
box (23,25,26). This bending of DNA has been postulated to contribute to the
stabilization of the TFIIIB-DNA complex by helping impede sliding of the DNA out of
the complex (23), a hypothesis consistent with thermodynamic and kinetic data indicating
B”-dependent kinetic trapping of the DNA (43). It has been suggested that BRF may also

undergo a conformational change upon B” binding to the TBP/BRF/DNA complex.

12

1 M83 IVNKSGTRFAPKVRQRRAATGGT PT PKPRT PQLFI PESKEIEEDNSD

51 NDKGVDENETAIVEKPSLVGERSLEGFTLTGTNGHDNEIGDEGPIDASTQ
101 NPKADVIEDNVTLKPAPLQTHRDQKVPRSSRLASLSKDNESRPS FKPS FL
1 5 l DSSSNSNGTARRLST I SNKLPKKI RLGS ITENDMNLKT FKRI—IRVLGKPSS
2 0 l AKKPAGAHRI S IVSKIS PPTAMTDSLDRNEFSSETSTSREADENENYVIS
2 5 l KVKDI PKKVRDGESAKYFI DEEN FTMAELCKPNFPIGQI SEN FEKSKMAK
30 1 KAKLEKRRHLRELRMRARQEFKPLHSLTM
3 5 l HWHTAIQLKLN PDGTMAI DEETMVVDRHKNAS I ENEYKEKVDEN

4 0 1 PFANLYNYGSYGRGSYTDPWTVEEMI

 

4 5 l IRWPILWLRSKLPPNFDEYCCEIKKNIGTVADFNE
5 O 1 KLI ELQNEHKHHMKE I EEAKNTAKEEDQTAQRLN DANLNKKGSGGIMTN D
5 5 l LKVYRKTEVVLGT I DDLKRKKLKERNN DDNEDNEGSEEE PE I DO

Figure 1.6 The amino acid sequence of S. cerevisiae B”. Blue, 329-357, glutamate rich

domain; rose, 415-472, SANT domain.

 

 

SANTdomain
human 8" ‘~ X
167 298 355 470 822 1338
‘W I???lZ?????I I1388
Scarevisiae B” E 21% 543%; 17%;
1 —7///A-'///MI594
281 415 472 578

Figure 1.7 Regions of similarity in the human and S. cerevisiae B” sequences. The rose
boxes correspond to the SANT domain; the hatched boxes indicate regions of lower but
still significant similarity on either side of the SANT domain. The percentages indicate
identities between the two proteins in the regions delimited by the dotted lines. The small

arrows indicate the repeats in human B”.

13

A central ~225 amino acid segment of B” appears to encompass its functional
core. Two domains, one at each end of this core region (in S. cerevisiae B”, amino acids
272-292 and 424-449), are required for TFIIIC-dependent transcription in vitro, and one
or the other of these segments is required for TFIIIC-independent transcription (15). An
additional and partly overlapping ~65 amino acid segment (amino acids 355-421) is
required for transcription of linear DNA (35). The principal target for DNA cross-linking
of B” lies between amino acids 277 and 315, i.e., it encompasses the N-proximal
"essential" amino acids 272-292 segment. This segment of B” faces the upstream part of
the DNA site occupied by TFIIIB. The N-terminal 223 amino acids of B” are not
essential for transcription of bare DNA (41). This segment of B” diminishes DNA-
protein cross-linking upstream of the TATA box at least 2-fold, as though it formed an
obstruction of the corresponding DNA-interacting segments of B”. Deleting B” amino
acids 1-185 has the same effect on cross-linking. A segment of B” extending
approximately from amino acid 190 to amino acid 210 becomes more accessible to
cleavage by hydroxyl radical cleavage upon entry into the TFIIIB-DNA complex (41).
This is consistent with the notion that DNA access of B” might be blocked by its internal

folding, and uncovered upon formation of the TFIIIB-DNA complex.

14

1.2 lL-myo-inositol l-phosphate synthase
1.2.1 Cell communication and the importance of inositol

Cells of multi-cellular organisms communicate with each other using substances
such as ions, hormones and neurotransmitters. Lipophilic agents such as steroid
hormones pass through the lipid bilayer of the cell membrane and bind to specific
intracellular receptors triggering the appropriate cellular responses. Hydrophilic
substances, which cannot cross the cellular lipid bilayer, deliver their message by binding
to specific receptors located on the cell surface. This type of signal transduction, also
called transmembrane signaling, depends on surface receptors. Several classes of
receptors are involved in signal transduction. The first class of cell surface receptor is
linked to an ion channel; stimulation of the receptor can trigger the ion channel to pump
ions into or out of the cell. A change in the given ionic species inside the cell is
perceived as an internal signal, evoking an ultimate response to the external signal. The
second class of cell surface receptors are related to tyrosine kinases (44). Binding of the
extracellular receptor by an agonist activates an intracellular tyrosine kinase and leads to
the phosphorylation of tyrosine residues on the target proteins inside the cell and causes
them to respond. The third class of receptor is coupled via a class of guanine nucleotide
binding proteins (G-protein) to the intracellular enzymes or ion channels, through which
the receptors evoke their responses. The family of G-proteins includes several members
that regulate different intracellular pathways. Upon agonist binding to the receptor, the
activated specific G-protein stimulates or inhibits other membrane bound enzymes that
act as amplifiers. This, in turn, generates second messengers inside the cell. Only a few

second messengers have been identified so far. They include adenosine 3’, 5' cyclic

15

monophosphate (CAMP), guanosine 3’, 5’ cyclic monophosphate (cGMP), diacylglycerol
(DAG), Ca2+, and myo-inositol 1,4,5-triphosphate (1,4,5-IP3). Most important signal

transduction pathways include a CAMP cascade and phosphoinositide cascade.

CAMP cascade

The binding of hormones such as adrenaline, calcitonin, glucagons, thyroid-
stimulating hormone, etc. to a seven-helix receptor in the plasma membrane triggers the
exchange of GTP for GDP bound to the stimulated G protein (Gs). The at subunit of the
G protein (Gsoc-GTP) then dissociates from the By subunits (Gfiy). Gsa-GTP activates
adenylate cyclase, an integral membrane protein. Adenylate cyclase (AC) catalyzes the
cyclization of ATP to produce CAMP. CAMP then activates protein kinase A (PKA) by
binding its regulatory subunit, thus unleashing its catalytic subunits. PKA alters the
activity of many target proteins by phosphorylating serine and threonine residues

(Figurel .8).

Phosphoinositide (PI) cascade

Phosphatidyl inositol (PtdIns) is the most abundant inositol lipid in nature,
making up 5 % of the total membrane phospholipid content in eukaryotes. Subsequent
phosphorylation on the inositol moiety of PtdIns produce a variety of phosphorylated
inositides including phosphatidylinositol-4, 5-bisphosphate (PtdIns (4,5) P2). Agonist
binding to a heterotrimeric G-protein-coupled receptor causes the exchange of GDP for
GTP and subsequent activation of phospholipase C (PLC). Hydrolysis of PtdIns (4,5) P2

by activated PLC produces myo-inositol 1,4,5-triphosphate (1,4,5-IP3), which binds to its

16

receptor buried in the membrane of the endoplasmic reticulum (ER) and stimulates the
release of Ca2+. Ca2+ acts by binding to calmodulin and other calcium sensors. Ca2+ -
bound calmodulin activates target proteins by binding to positively charged amphipathic
helices, causing a diverse set of cellular responses including secretion, excitation,
contraction, growth and proliferation (45-47).

Cell communication in brain

In brain, communication between neurons makes use of the signal transduction
system involving 1,4,5-IP3 and the phosphoinositide cascade and therefore requires a
suitable amount of myo-inositol. In most mammalian cells, myo-inositol is obtained from
diet, however in brain cells, the major portion of myo-inositol is biosynthesized from D-
glucose 6-phosphate since plasma myo-inositol does not effectively cross the blood-brain
barrier (48,49).

The de novo biosynthesis of myo-inositol begins with the conversion of D-
glucose-6-phosphate to 1L-myo-inositol-l-phosphate by MIP synthase, followed by
dephosphorylation of this product by myo-inositol monophosphatase to produce myo-
inositol. Myo-inositol is then combined with cytodine diphosphate-diacylglycerol (CDP-
DAG) in a reaction catalyzed by phosphatidyl inositol (PtdIns) synthase to form

phosphatidyl inositol.

17

hormone
signal

I

outside

    

GTP GDP ATP CAMP + PPi

Figurel .8 CAMP cascade.

    
  
 

cytosol

 

 

 

 

 

Inositol
monophosphatase 1'4‘P
myo- D-I-l-P
inositol LL 1 '1’
Li”
MIP
synthase

D-glucose 6-phosphate

Figure1.9 Phosphoinositide cascade.

Cellular signaling between hyperactive neurons and neighboring cells causes an
increase in the turnover of the Phosphoinositide cascade. In the over stimulated
neighboring cells, the constant resynthesis of PtdIns (4,5)P2 is required in order to
maintain the intracellular level of 1,4,5-IP3 as the primary response to the excess stimuli.
The increased rate of hydrolysis of PtdIns (4,5)P2 to diacylglycerol and 1,4,5-IP3 alters
the electrical activity of the affected neurons. The abnormal electrical activity in the
brain tissue surrounding the pathological cells is thought to result in manic disorders. In
vivo, the inhibition of the phosphoinositide cascade phosphatases by lithium results in the
depletion of cellular myo-inositol. The decrease in the myo-inositol availability to
regenerate the PtdIns (4,5) P2 pool results in signal termination and provides control over
manic disorders. Effects of inositol phosphatase inhibition are only observed in brain
cells since no dietary extra cellular myo-inositol can be pumped into the phosphoinositide
cascade (Figure 1.9) (45,47).

Lithium carbonate is used with variable success to treat manic disorder because of
its ability to inhibit various enzymes of the phosphoinositide cascade, particularly myo-
inositol monophosphatase (50,51). However, the established success of lithium as the
therapeutic treatment for manic disorder is not complete. Suppressing the hydrolysis of
inositol monophosphatases leads to the depletion of myo-inositol and the scarcity of
PtdIns (4,5) P2, the precursor of 1,4,5-IP3. Consequently the hyperactive neuron would
be dampened since it loses a critical pool of PtdIns (4,5)P2 for signal transduction. In
addition, a large amount of lithium carbonate (>2000 mg/day) is required to maintain an
effective drug concentration at the cellular level (0.5-1.2 mM), which is relatively close

to the toxic level (3-5 mM). In addition to myo-inositol monophosphatase, lithium also

20

inhibits other enzymes and causes accumulation of various metabolites. Along with the
side effects of lithium, insensitivity to this drug and a decrease in effectiveness due to its
required continual uptake have also been reported (52).

Inhibition of MIP synthase in brain tissues represents an alternative approach for
treating manic depression. This inhibition is expected to reduce myo-inositol as well as
MIP levels and may present different neurological outcomes than that observed with
lithium treatment and myo-inositol depletion. The accumulated D-glucose 6-phosphate
upon inhibition of MIP synthase could be used as a substrate for other intracellular
enzymes.

Recent in vivo experiments in yeast and Dictyostelium showed that valproate, a
drug used in the treatment of depression, bipolar disorder, and seizure disorder, may act
by inhibition of MIP synthase, thus lowering neuronal inositol pools similar to the action

of lithium (53,54).

1.2.2 lL-myo-inositol l-phosphate synthase

lL- myo-inositol l-phosphate (MIP) synthase (EC 5.5.1.4) is found in all
eukaryotes investigated thus far including protozoa, fungi, algae, plants and animals. The
amino acid sequence of MIP synthase has been remarkably well conserved throughout
evolution over virtually its entire sequence (Figure 1.10).

MIP synthase catalyzes the isomerization of D-glucose 6-phosphate into lL-myo-
inositol l-phosphate (Figure 1.11), B-nicotinamide adenine dinucleotide (NAD+) is
required to catalyze this reaction (55). MIP synthase employs NAD+ as a prosthetic

group, which is neither consumed nor generated during the catalytic cycle.

21

Loewus and Kelly were the first to propose that a cyclization mechanism involving the
generation of a C5 ketose prior to the formation of the C-C bond was the most likely
pathway (56,57). The reaction is proposed to begin with the binding of D-glucose 6-
phosphate to the enzyme-NAD+ complex. The resulting open form of D-glucose 6-
phosphate is oxidized at the C5 position with the simultaneous reduction of NAD+ to
NADH and forms the intermediate B, 5-keto-D-glucose 6-phosphate, in which the acidity
of the or protons at C6 is greatly increased. The subsequent enolization of intermediate B
to intermediate C is followed by the aldol condensation reaction between C6 and C1.
Intermediate D, myo-2-inosiose l-phosphate, is then reduced by the enzyme-bound
NADH fomiing NAD’r and MIP, which is subsequently released from the enzyme active

site.

22

HI‘F‘HI‘

77
74
70
26
27

157
153
150
67
76

236
223
220
128
140

314
301
297
191
219

392
378
374
268
286

472
452
448
336
362

533
511
528
367
391

MTEDNIAPITSVKVVTDKCTYKDNELLTKYSYENAVVTK-—TASGRFD-—VTPTVQDYVFKLDLKKPEKLGIMLIGLGGN

MFIES-PKVESPNVK ----- YTENEIHSVYDYBTTEVVHEKTVNGTYQWIVKPKTVKYDFKTDIRVP-KLGVMLVGLGGN
MEAAAQPFVESPDVV ----- YGPEAIEAQYEYRTTRVSREGGV ----- LKVHPTSTRFTFRTARQVP—RLGVMLVGWGGN
MSEHQ-—SLPAPEASTE -------------------------------------------------- VRVAIVGV-—-GN
MKV --------------------------------------------- WLVGA-—--YGI--—VSTTAMVGARAIERG-—

NGSTLVASVLANKHNVEFQTKEGVKQPNYFGSMTQCSTLKLGIDAEGNDVYAPFNSLLPMVSPNDFVVSGWDINNADLYB
NGSTLTAGVIANKEGISWATKDKVQQANYFGSLTQASSIRVG—SFNGEEIYAPPKSLLPHVNPDDVVFGGWDISDMNLAD
NGSTLTAAVLANRLRLSWPTRSGRKEANYYGSLTQAGTVSLGLDABGQEVFVPPSAVLPMVAPNDLVPDGWDISSLNLAE
CASSLVQGV ------------------ EYYYNADDTSTVP—GLM—--HVRPGPYH ----------------- VRDVKFVA
--------- IAPKIGL-———-—-VSELPHFEGIEKYAPFS--FEFGGHEI-——-—RLLS--NAYEAAKEHWBLN------

AMO-RSQVLEYDLQQRLKAKMSLVKPLPSIYYPDFIAANQDERANNCINLDEKGNVTTRGKWTHLQRIRRDIQNFKEENA
AMA-RARVLDIDLQKQLRPYMENIVPLPGIFDPDFIAANQGSRANH --------- VIKGTKKEQVDHIIKDMREFKEKNK
AMR-RAKVLDWGLQEQLWPHMEALRPRPSVYIPEFIAANQSARADN --------- LIPGSRAQQLEQIRRDIRDFRSSAG

AFDVDAKKVGFDLSDAI ----------------- FASENNTIKIADVAP-—TNVIVQRGPTLDGIGKYYADTIBLSDAEP
------ RHFDREILEAVKSDLEGIVARKG--'TALNCGSGIKELGDIKTLEGEGLSLA---~EMVSRIEEDIKSFAD---

LDKVIVLWTANTERYVEVSPGVNDTMENLLQSIKNDHEEIAPST—IFAAASILEGVPYINGSP-QNTFVPGLVQLABHEG
VDKVVVLWTANTERYSNVVVGMNDTMENLMESVDRDEAEISPST-LYAIACVLEGIPFINGSP-QNTPVPGLIDMAIRNN
LDKVIVLWTANTERFCEVIPGLNDTAENLLRTIELG-LEVSPST-LFAVASILEGCAFLNGSP—QNTLVPGALELAWQHR
VDVVQALKEAKVDVLVSYLPVGSEEADKF ----------------- YAQCAIDAGVAFVNALPVFIASDPVWAKKFTDAR
-DETVVINVASTEPLPNYSEEYHGSLEGFERMIDEDRKEYASASMLYAYAALKLGLPYANFTPSPGSAIPALKELABKKG

TPIAGDDLKS-—GQTKLKSVLAQFLVDAGIKPVSIASYNHLGNNDGYNLSAPKQFRSKEISKSSVIDDIIASNDILYNDK
VLIGGDDFKS--GQTKMKSVLVDFLVGAGIKPTSIVSYNHLGNNDGMNLSAPQTFRSKEISKSNVVDDMVASNGILFEP—
VFVGGDDFKS--GQTKVKSVLVDFLIGSGLKTMSIVSYNHLGNNDGENLSAPLQFRSKEVSKSNVVDDKVQSNPVLYTP-
VPIVGDDIKSQVGATITHRVLAKLFEDRGVQLDRTMQLNVGGNMDFLNMLERERLESKKISKTQAVTSNLKRE—--PKTK
VPHAGNDGKT-—GETLVKTTLAPMFAYRNMEVVGWMSYNILGDYDGKVLSARDNKESKVLSKDKVLBKM -----------

LGKKVDHCIVIKYMKPVGDSKVAMDEYYSELMLGGHNRISIHNVCEDSLLATPLIIDLLVMTEFCTRVSYKKVDPVKBDA
-GEHPDHVVVIKYVPYVADSKRAMDEYTSEIFMGGKNTIVMHNTCEDSLLAAPIILDLVLLABLSTRIQFKS ----- 8GB
-GEEPDHCVVIKYVPYVGDSKRALDEYTSELMLGGTNTLVLHNTCEDSLLAAPIMLDLALLTELCQRVSFCT ----- DMD
----DVHIGPSDHVGWLDDRKWAYVRLEGRAFGDVPLNLEYKLEVWDSPNSAGVIIDAVRAAKIAKDRGIGG --------
LGYSPYSITEIQYFPSLVDNKTAFDFVHFKGFLGKLMKFYFIWDAIDAIVAAPLILDIARFLLFAKKKGVKGV--VKE--

GKPENFYPVLTFLSYWLKAPLTRPGFHPVNGLNKQRTALENPLRLLIGLPSQNELRPEBRL -------------------

GKFHSPHPVATILSYLTKAPLVPPGTPVINALSKQRAMLENIMRACVGLAPENNMIMEFK
PEPQTFHPVLSLLSFLFKAPLVPPGSPVVNALFRQRSCIENILRACVGLPPQNHMLLEHKMERPGPSLKRVGPVAATYPM
------- PVIPASAYLMKSPPEQLPDDIA----——RAQLEEFI~--IG

------------ MAFFFKSPM---DTNVINT---——~-—--—--———-—-HBQFVVLKEw--—--------------YSN
______________________________ L

LNKKGPVPAATNGCTGDANGHLQEEPPMPTT

LK

S. cerevisiae
A. thaliana
human

M. tuberculosis
A. fulgidus

S. cerevisiae
A. thaliana
human

M. tuberculosis
A. fulqidus

S. cerevisiae
A. thaliana
human

M. tuberculosis
A. fulgidus

S. cerevisiae
A. thaliana
human

M. tuberculosis
A. fulgidus

S. cerevisiae
A. thaliana
human

M. tuberculosis
A. fulgidus

S. cerevisiae
A. thaliana
human

M. tuberculosis
A. fulgidus

S. cerevisiae
A. thaliana
human

M. tuberculosis
A. fulgidus

S. cerevisiae
A. thaliana
human

M. tuberculosis
A. fulgidus

Figure 1.10 Sequence alignment Of MIP synthase from Saccharomyces cerevisiae,

Arabidopsis thaliana, human, Mycobacterium tuberculosis, and Archaeoglobusfillgidus.

Identical residues are highlighted in yellow.

23

OH
HO 0
HO

OPOaH 2
D-glucose 6-phosphate

OPOaHz
mflomamw £37230

myo-inositol 1-phosphate (MIP)
(Oxidation) )' H01 (Reduction)
wflOPOaHzi-‘NA HHOfi :POaHz
Omyo-2-inosose 1-phosphate
Enolization H

Aldol cyclization
\ OPOaH 2

HO C

Figure 1.11 Proposed mechanism of MIP synthase.

24

Various experiments support the mechanism of Figure1.11: (a) MIP synthase
requires NAD+ for its activity and NADH is tightly associated with the enzyme.
Incubation of independently synthesized intermediate B with apo MIP synthase
reconstituted with [4-3H] NADH resulted in the formation of D- [5-3 H] glucose 6-
phosphate and [3H]-myo-inositol l-phosphate (58,59). In situ formation of NADH during
substrate turnover has been observed by UV-vis spectroscopy (60). (b) A trapping
experiment using tritiated sodium borohydride ([3H]-NaBH4) indicated the transient
existence of intermediate D (61). (c) Isotope effects were observed at both the C5 and C6
positions of D-glucose 6-phosphate. This observation is consistent with the presence of
intermediate B, which indicates that the enolization is rate limiting (62,63). ((1) The pro-
R hydrogen of the C6 methylene group is preferentially removed during enolization (64).
Sherman, Eisenberg, and Barnett found that the reduction of the MIP synthase reaction
mixture with [3H]-NaBH4 did not furnish any lysyl derivative of glucose, regardless of
the source of MIP synthase used (59,65,66). Another experiment using H2'8O showed
that '80 incorporation into the product was not the result of the hydrolysis of a Schiff
base but came from the nonenzymatic exchange of the glucose aldehyde carbonyl with
the medium. These various observations established that MIP synthase could not be
classified as a type I aldolase, which involves the formation of a Schiff base (Figure
1.12). On the other hand, MIP synthase was not inhibited by EDTA, therefore MIP
synthase does not utilize a type II aldolase mechanism, in which a divalent metal ion such
as Zn”, Mn2+ acts as the Lewis acid (Figure 1.13) (67-69). It has been reported recently
that MIP synthase from Archaeoglobusfulgidus requires divalent cations and therefore is

thought to be a type II aldolase (70).

25

Enzev H+

R1

H+ 00 R2
F12 Enz-B R1
R1

HO R3

Figure 1.12 The mechanism of a type I aldolase.

R1

HO a,

Figure 1.13 The mechanism of a type II aldolase.

26

MIP synthase might be responsible for the type of base-catalyzed cyclization
where the appropriate activation of the substrate is carried out by the enzyme oxido-
reductase function. Loewus showed that only the C6 pro-R hydrogen was lost during
turnover and that the cis-enol C resulted directly from this selective removal (56).
Sherman demonstrated that intermediate B was reduced by a base to produce two major
products identified as epi-inositol 3-phosphate and D-myo-inositol 3-phosphate (71). The
difference in stereo selectivity observed between a base-catalyzed cyclization and MIP
synthase catalyzed cyclization indicated the strong involvement of the enzyme active site
during aldol condensation. Enzymatic base-catalyzed enolization must be stereo
controlled and the orientation of the newly formed myo-inositol l-phosphate must be
under strict enzymatic control.

Although the mechanism proposed above accounts for the available data, a
number of mechanistic questions still remain, especially the nature of the aldol

condensation.

1.2.3 Previous structural investigations on MIP synthase

Crystal structures of MIP synthase from S. cerevisiae with partially occupied
NAD+, and in complex with NAD” and an inhibitor, 2-deoxy-D-glucitol 6-phosphate
have recently been determined (72). The results showed an example of induced fit,
where nearly 60 residues, residues 351-409, in the active site that were disordered in the
structure with a partially occupied NAD+ become ordered upon the binding of the
inhibitor (Figure 1.14). It appeared that the complete folding of the enzyme active site

required either complete NADI occupancy or inhibitor/substrate binding, or both. The

27

 

Figure 1.14 (A) Ribbon model of the MIP synthase monomer. Red, the N-terminal
region; purple, the NADI-binding region; green, the tetramerization region; blue, the C-
terminal domain. (B) Ribbon model of the MIP synthase/NADVZ-deoxy-D-glucitol 6-
phosphate complex. Green, the residues that were ordered in the structure with low
occupancy NAD”; red, the newly ordered residues; yellow, NAD”; magenta, 2-deoxy-D-
glucitol 6-phosphate.

28

inhibitor, 2-deoxy-D-glucitol 6-phosphate was bound in its extended conformation
inconsistent with the substrate during the reaction. Based on the structural data, the
substrate modeling in the conformation necessary for cyclization was performed and a
mechanism was proposed (Figure 1.15). It was proposed that the first step involves an
oxidation at C5. Subsequently the substrate is reoriented to a conformation where the
phosphate can act as a base at the enolization step. The developing negative charge is
stabilized by a nearby NH4+. Nucleophilic attack by C6 on C1 is promoted by K369
stabilization of the developing negative charge on 01. Subsequent protonation at 01 by
K369 and reduction of C5 by NADH yields the product MIP.

The crystal structure of MIP synthase from Mycobacterium tuberculosis with
NAD+ has also been published recently (73). In this structure, residues 241-267, which
are conserved as 362-391 in S. cerevisiae MIP synthase, are missing (Figure 1.16). The
structure has an NAD+ moiety and a zinc ion in the active site. The nicotinamide moiety
of the NAD” is in the syn conformation about the N-glycosidic bond and is held there by
coordination with the Zn2+. The Zn” ion was identified by its tetrahedral coordination
and from bond lengths of close to 2.2 A for each of its four ligands. It lies adjacent to the
NAD+ nicotinamide moiety, bridging the nicotinamide oxygen atom and phosphodiester
oxygen N02 and coordinating the S311 OG atom and a water molecule. The presence of
Zn2+ may influence the pK. of the acidic side chains of D235 and D310, which coordinate
this water molecule. The Zn2+ ion is more buried than that present in similar enzymes
that employ Zn2+ directly in the mechanism, e. g., horse liver alcohol dehydrogenase

(PDB code 3bto). A detailed analysis of the catalytic mechanism of MIP synthase was

29

3+ ‘ 1 [(39
Ht? :0" ‘. :"'N-b/\/

sari ..- ‘~ -

\/°H:. k P o-

DB8 N354
o/xkq CAME
K489 ‘~ "
+ 'NI'U K369
- "a s‘ N
m. ‘I. °-’°¢"§W~Z"~'f‘
vow: \ O\P.o‘:“.‘o
II
0
j: Nasa
0/'_\0.‘ ”02m,
K489 + ‘~ '1
. K369
We: . + N
\FD 00 0"::an
m .a‘ ‘.‘ t -
who o\E""“oi-i
NAD
NADH
HO OH OH
HO pfiKEH-MO'
MIP

Figure 1.15 Proposed mechanism for the transformation catalyzed by MIP synthase (72).

30

not performed owing to the lack of a structure complexed with the substrate or suitable

inhibitors.

 

Figure 1.16. The structure of MIP synthase from Mycobacterium tuberculosis (73).

31

1.2.4 Remaining challenges of MIP synthase
1) The effect of NAD+

NAD+ is one of the most commonly used cofactors in living cells. As first noted
by Rossmann et al. (74), most NAD+ binding proteins are similar in tertiary structure in
the region where NAD+ binds. This core topology region consists of a BorBaB unit with
at least one additional parallel B strand. Within the BOLBOLB unit is a sequence of 30-35
amino acids forming the finger print region, which can be used to identify the location of
NAD+ binding. This finger print region includes a glycine-rich phosphate binding
consensus sequence (GXGXXG), where the first strictly conserved glycine allows for a
tight turn of the main chain, which is important for positioning the second glycine. The
second glycine, because of its missing side chain, allows for close contact of the main
chain to the pyrophosphate of NAD+. It is thought that any side chain in this position
would protrude into the binding site of NAD+ and disrupt binding. The third glycine is
important for the close packing of the secondary structure elements of the first [3 strand
and the first or helix. Instead of the common GXGXXG motif, MIP synthase from
eukaryotes has a GXGGXXG motif. Interestingly, it appears that MIP synthase from
Archaeglobusfulgidus does not have a similar GXG(G)XXG motif. In addition, several
other NAD+-interacting residues are not conserved in A. fizlgidus enzyme (Figure 1.10).
Therefore, it can be presumed that the NAD+ binding of A. fidgidus MIP synthase could
be quite different from that of MIP synthase from eukaryotes and Mycobateria.

Enzymes that bind NAD+ catalyze reactions central to energy production, storage,
and transfer by exploiting the ability of the nicotinamide group to transfer hydride ions or

electrons. In most NAD+-dependent catalysis, NAD+ is a cosubstrate. However, there is

32

also a family of NAD+-requiring enzymes where NAD+ plays the role of a catalytic
prosthetic group (75). These enzymes catalyze reactions such as epirnerization, aldol
condensation, cyclization, a,B-elimination, and decarboxylation, in which strong bonds
with no obvious lability are cleaved. A transient oxidation of the appropriate substrate
carbon along with the simultaneous reduction of NAD’( labilizes the scissile bond by
introducing a carbonyl or an irnine. NADH then reduces the oxidized center at a later
step subsequent to the enzymatic transformation of the oxidized activated substrate. MIP
synthase uses NAD+ as a prosthetic group, which is neither consumed nor generated at
the end of each catalytic cycle as shown in Figure 1.11. The three-dimensional structure
of the MIP synthase/NAD+ complex must represent the enzyme at its state when the
substrate is not yet bound or the product is already produced and released from the
enzyme active site. This structure should also provide insights into any conformational
changes that might occur in both the Rossmann fold domain and the active site of MIP
synthase due to the binding of NADL One big question to be answered is whether the
active site disorder observed previously in the structure of MIP synthase with partially

occupied NAD+ (72) could be due to not fully occupied NAD+.

2) The structure of the NADH-bound MIP synthase

Radio labeling experiments have shown that the hydride transferred from the C5
position of the substrate to NAD+ during the oxidation step was returned to myo-2-
inosose l-phosphate by NADH at the reduction step (76). The nicotinamide ring of
NAD+ must be located adjacent to the pyranose ring since there is a direct hydride

transfer, and the same face of the nicotinamide ring must participate in the oxidation and

33

reduction steps. After the substrate is oxidized with the formation of NADH from NAD”,
the substrate or nicotinamide ring may move in order to reposition the carbonyl oxygen at
C5, and consequently the side chains of active site residues may need to experience
conformational changes as well. Furthermore, all non-redox steps of the reaction
catalyzed by MIP synthase occur when the enzyme is coupled with NADH after the
oxidation of the substrate, implicating that the three-dimensional structure of the MIP
synthase/NADH complex may provide more information on the enzyme in its active

state.

3) The structure of MIP synthase in complex with structural analogues of the substrate
and reaction intermediates

Even with the previously published structural data on the MIP synthase/NADVZ-
deoxy-D-glucitol 6-phosphate complex, some questions still remain unanswered. More
structures complexed with various structural analogues of the substrate and reaction
intermediates should produce a more complete picture of the MIP synthase mechanism.
Previous inhibition studies have provided clues on the substrate binding to MIP synthase.
First, in order to investigate whether the cyclic or acyclic form of the substrate binds MIP
synthase, a series of cyclic and acyclic substrate analogues were synthesized. None of the
molecules that are covalently locked in the cyclic form or that undergo ring opening
slowly led to any detectable inhibition of MIP synthase (Figure 1.17) (77).
Several cyclic analogues of the substrate D-glucose 6-phosphate that can undergo ring
opening such as 2-deoxy-D-glucose 6-phosphate and 2-deoxy-D-glucose 6-

homophosphonate are inhibitors of MIP synthase (Figure 1.18) (78). One of the

34

intermediates in the proposed reaction pathway, myo-2-inosose l-phosphate has been
synthesized and found to be a potent inhibitor (Ki= 3.6 x 10'6 M), and several analogues
of this intermediate were also found to be inhibitors of MIP synthase (Figure 1.19) (79).
On the other hand, several acyclic analogues of the substrate were found to be very potent
inhibitors of MIP synthase (Figure 1.20) (60). Structures of MIP synthase complexed

with these acyclic and cyclic inhibitors will present the conformation of MIP synthase at

its different steps of the catalytic cycle.

    

0"""00-

0P03H2 5H OPO3H2 OH 0P03H2

Olllllnn
I

OPO3H2

Figure 1.17 Cyclic analogues of D-glucose 6-phosphate with no inhibition of MIP

   

synthase (77).
OPO3H2 OH CH2P03H2 C=)H
2-deoxy-D- glucose 6-phosphate 2-deoxy-D- glucose 6-homophosphonate
Ki=9.lxlO*’M Ic=7.1xro"M

Figure 1.18 Cyclic analogues of D-glucose 6-phosphate that can undergo ring opening

with inhibition of MIP synthase (78).

35

0 OH OH 0 OH
P03H2 . P03H2 H2P03H2

myo-Z-inosose l-phosphate 2-deoxy-myo-inositol l-phosphate l—deoxy-l-(phosphonomethyl)-
myo-2-inosose
K,=3.6x10‘M KI=170><106M Ki=37xlofM

Figure 1.19 The intermediate D, myo-Z-inosose l-phosphate, and its analogues are

inhibitors of MIP synthase (79).

        

0P03H2 OH CH2P03H2 OH

0P03H2 OH CH2P03H2 OH
D-glucitol 6-phosphate D-g]ucjto] 6— 2-deoxy-D-glucitol 6- 2-deoxy-D-glucitol 6—
homophosphonate phosphate homophosphonate
Ki = 1.5 x 10" M

K,=l.lxlO“‘M K.=2.3x10-6M Ki=5.8x10'°M

   

OPOaHz 8“ 0P03H2 C=)H
D-arabinitol 5-phosphate D-thhfiIOI 4-phosphate
Ki=l.7X10'5M K;=4.7X10'5M

Figure 1.20 Acyclic analogues of D-glucose 6-phosphate are inhibitors of MIP synthase

(60).

36

Second, in order to investigate the mechanism proposed by Floss (80) that a
single, active site base and its conjugate acid mediate all of the non-redox steps (Figure
1.21), conformationally restricted (Z)- and (E)- vinylhomophosphonate analogues of the
substrate were synthesized (77). While none of the (Z)-vinylhomophosphonates were
inhibitors of MIP synthase, all (E)-vinylhomophosphonates were competitive inhibitors.
These results are consistent with the reaction pathway where transoid conformations of
the substrate D-glucose 6-phosphate and the intermediate 5-keto-D-glucose 6-phosphate
may be found. A transoid conformation would position the dibasic phosphate monoester
of 5-keto-D-glucose 6-phosphate for the removal of the pro-R hydrogen of C6 (Figure
1.22). Stereo selective proton removal would result in the monobasic phosphate
monoester, which could then deliver the abstracted proton back to the C1 carbonyl
oxygen to complete its catalytic role in the formation of the bond between C1 and C6.
Intramolecular removal of a proton by a phosphate monoester is precedented for 3-
dehydroquinate synthase, an enzyme mechanistically related to MIP synthase in its
catalytic use of NAD+ and aldol cyclization (81). Though these inhibition studies are
consistent with the catalytic involvement of the phosphate monoester, they do not prove
it. The structures of MIP synthase complexed with (E)-vinylhomophosphonate analogues

of the substrate will answer this question.

37

HO
B=CH20 or CHQOH
no inhibition 9 /

HO +
OH NAD NADH
0‘ HR
0

OH
OH
OH
OH
13 o\® OH
3..
.. I,
HO OH HO OH
H/o \j -—> m
\ - @— °" c
.8 O 13 @—
3.
01

O

 

 

NAD’ NADH
HO
OH #-
HO
H203PO

OH

Figure 1.21 The mechanism of MIP synthase proposed by Floss (80).

38

OH OH

HO HO
OH NAD+ NADH OH
0‘ w 0\ “OHR

O
®'-O OH _ \/p/_.O O
O

OH OH
OH NAD‘ NADH OH
HO / ‘ HO /
OH ® ' 0

2-deoxy-D-glucitol 6-(E)-vinylhomophosphonate
Ic=0.67xrot M

Figure 1.22 The mechanism consistent with the substrate binding in a transoid

conformation and the phosphate monoester acting as the base in the enolization step (77).

39

Third, non-charged species such as D-glucose or 2-deoxy-D-glucose are not
inhibitors of MIP synthase (82), indicating the importance of the phosphate in the
substrate. In order to establish whether the phosphate moiety of the substrate is
interacting with one or two binding pockets during each turnover, a diphosphate analogue
of the substrate was synthesized and found to be an inhibitor with K; = 0.06 mM (Figure
1.23) (78). The structure of MIP synthase complexed with this inhibitor will put to rest

the question of whether the phosphate would move during the enzyme catalysis.

 

0P03H2 OPO3H2
o on 0P03H2
HO _‘ HO
HO —-. *—
HO OH OH OH
OH OH
Km=1.2 [TIM \ /
0P03H2
0P03H2
HO
HO OH
OH
Kj=0.06 mM

 

 

 

Figure 1.23 Analogues of reaction intermediates of the early steps during the catalysis by

MIP synthase(78).

40

1.3 Literature cited

1.

2.

10.

11.

12.

13.

14.

15.

Jacob, S. T., and Ghosh, A. K. (1999) J.Cell.Biochem.Suppl, 41-50

Paule, M. R. (1990) Nature 344, 819-820

Paule, M. R., and White, R. J. (2000) Nucleic Acids Res 28, 1283-1298
Woychik, N. A., and Hampsey, M. (2002) Cell 108, 453-463

Orphanides, G., Lagrange, T., and Reinberg, D. (1996) Genes Dev 10, 2657-2683
Kaufrnann, J., and Smale, S. T. (1994) Genes Dev 8, 821-829

Verrijzer, C. P., Chen, J. L., Yokomori, K., and Tjian, R. (1995) Cell 81, 1115-
1125

Gill, 6., and Tjian, R. (1992) Curr Opin Genet Dev 2, 236-242

Dynlacht, B. D., Hoey, T., and Tjian, R. (1991) Cell 66, 563-576

Tanese, N., Pugh, B. F., and Tjian, R. (1991) Genes Dev 5, 2212-2224

Zawel, L., and Reinberg, D. (1995) Annu Rev Biochem 64, 533-561

Carey, M., and Smale, S. T. (2000) Transcriptional regulation in eukaryotes.
Concepts, strategies, and techniques, lst Ed., Cold Spring Harbor Laboratory
Press, Cold Spring Harbor

Chedin, S., Ferri, M. L., Peyroche, G., Andrau, J. C., Jourdain, S., Lefebvre, 0.,
Werner, M., Carles, C., and Sentenac, A. (1998) Cold Spring Harbor Symp.
Quant. Biol. 63, 381-389

Hernandez, N. (1993) Genes Dev 7B, 1291-1308

Kumar, A., Grove, A., Kassavetis, G. A., and Geiduschek, E. P. (1998) Cold

Spring Harb Symp Quant Biol 63, 121-129

41

16.

17.

18.

19.

20.

21.

22.

23.

24.

25.

26.

27.

28.

29.

White, R. J. (1998) RNA Polymerase III Transcription, Springer-Verlag, New
York

Geiduschek, E. P., and Kassavetis, G. A. (2001) J Mol Biol 310, 1-26

Huang, Y., and Maraia, R. J. (2001) Nucleic Acids Res 29, 2675-2690

White, R. J. (1998) IntJ Oncol 12, 741-748

Kassavetis, G. A., Braun, B. R., Nguyen, L. H., and Geiduschek, E. P. (1990) Cell
60, 235-245

Apone, L. M., and Green, M. R. (1998) Nature 393, 114-115

Geiduschek, E. P., and Kassavetis, G. A. (1995) Current Opinion in Cell Biology
7, 344-351

Grove, A., Kassavetis, G. A., Johnson, T. E., and Geiduschek, E. P. (1999) J Mol
Biol 285, 1429-1440

Grove, A., Galeone, A., Yu, E., Mayol, L., and Geiduschek, E. P. (1998) J Mol
Biol 282, 731-739

Braun, B. R., Kassavetis, G. A., and Geiduschek, E. P. (1992) J Biol Chem 267,
22562-22569

Leveillard, T., Kassavetis, G. A., and Geiduschek, E. P. (1991) J Biol Chem 266,
5162-5168

Takada, S., Lis, J. T., Zhou, S., and Tjian, R. (2000) Cell 101, 459-469

Kuras, L., Kosa, P., Mencia, M., and Struhl, K. (2000) Science 288, 1244-1248
Kosa, P. F., Ghosh, G., DeDecker, B. S., and Sigler, P. B. (1997) Proc Natl Acad

Sci U S A 94, 6042-6047

42

30.

31.

32.

33.

34.

35.

36.

37.

38.

39.

40.

41.

42.

Nikolov, D. B., Chen, H., Halay, E. D., Usheva, A. A., Hisatake, K., Lee, D. K.,
Roeder, R. G., and Burley, S. K. (1995) Nature 377, 119-128

Bagby, S., Kim, S., Maldonado, E., Tong, K. I., Reinberg, D., and Ikura, M.
(1995) Cell 82, 857-867

Chaussivert, N., Conesa, C., Shaaban, S., and Sentenac, A. (1995) J Biol Chem
270, 15353-15358

Khoo, B., Brophy, B., and Jackson, S. P. (1994) Genes Dev 8, 2879-2890
Kassavetis, G. A., Kumar, A., Ramirez, E., and Geiduschek, E. P. (1998) Mol Cell
Biol 18, 5587-5599

Kassavetis, G. A., Bardeleben, C., Kumar, A., Ramirez, E., and Geiduschek, E. P.
(1997) Mol Cell Biol 17, 5299-5306

Buratowski, S., and Zhou, H. (1992) Cell 71, 221-230

Colbert, T., Lee, S., Schirnmack, G., and Hahn, S. (1998) Mol Cell Biol 18, 1682-
1691

Cormack, B. P., and Struhl, K. (1993) Science 262, 244-248

Huet, J ., Conesa, C., Manaud, N., Chaussivert, N., and Sentenac, A. (1994)
Nucleic Acids Res 22, 3433-3439

Aasland, R., Stewart, A. F., and Gibson, T. (1996) Trends Biochem Sci 21, 87-88
Kumar, A., Kassavetis, G. A., Geiduschek, E. P., Hambalko, M., and Brent, C. J.
(1997) Mol Cell Biol 17, 1868-1880

Shah, S. M., Kumar, A., Geiduschek, E. P., and Kassavetis, G. A. (1999) J Biol

Chem 274, 28736-28744

43

43.

44.

45.

46.

47.

48.

49.

50.

51.

52.

53.

54.

55.

56.

57.

Cloutier, T. E., Librizzi, M. D., Mollah, A. K., Brenowitz, M., and Willis, 1. M.
(2001) Proc Natl Acad Sci U S A 98, 9581-9586

Goutsouliak, V., and Rabkin, S. W. (1997) Cell Signal 9, 505-512

Lenox, R. H., and Watson, D. G. (1994) Clin Chem 40, 309-314

Rana, R. S., and Hokin, L. E. (1990) Physiol Rev 70, 115-164

Duman, R. S., Heninger, G. R., and Nesteler, E. J. (1997) Arch. Gen. Psychiatry
54, 597

Margolis, R. U., Press, R., Altszuler, N., and Stewart, M. A. (1971) Brain Res 28,
535

Sherman, W. R. (1989) Inositol Lipids in Cells Signaling (Downes, C. P., Ed.),
Academic, New York

Berridge, M. J ., Downes, C. P., and Hanley, M. R. (1982) Biochem J 206, 587-
595

Sherman, W. R., Gish, B. G., Honchar, M. P., and Munsell, L. Y. (1986) Fed
Proc 45, 2639-2646

Peet, M., and Pratt, J. P. (1993) Drugs 46, 7

Vaden, D. L., Ding, D., Peterson, 8., and Greenberg, M. L. (2001) J Biol Chem
276, 15466-15471

Williams, R., S. 3., Cheng, L., Mudge, A. W., and Harwood, A., J. (2002) Nature
417, 292-295

Barnett, J. E., and Corina, D. L. (1968) Biochem J 108, 125-129

Loewus, M. W., and Loewus, F. A. (1980) Carbohydr Res 82, 333-342

Loewus, F. A., and Kelly, S. (1962) Biochem Biophys Res Commun 7, 204-208

58.

59.

60.

61.

62.

63.

64.

65.

66.

67.

68.

69.

70.

71.

72.

Barnett, J. E., Rasheed, A., and Corina, D. L. (1973) Biochem Soc Trans 1, 1267-
1269

Barnett, J. E., Rasheed, A., and Corina, D. L. (1973) Biochem J 131, 21-30

Peng, J. (1996) Ph.D. Dissertation, Michigan State University

Chen, C. H., and Eisenberg, F ., Jr. (1975) J Biol Chem 250, 2963-2967

Chen, 1. W., and Charalampous, F. C. (1967) Biochim Biophys Acta 136, 568-570
Loewus, M. W. (1977) J Biol Chem 252, 7221-7223

Loewus, M. W., Loewus, F. A., Brillinger, G. U., Otsuka, H., and Floss, H. G.
(1980) JBiol Chem 255,11710-11712

Sherman, W. R., Hipps, P. P., Mauck, L. A., and Rasheed, A. (1978) In Cyclitols
and phosphoinositides (Eisenberg, F ., Ed.), Academic Press, New York

Maeda, T., and Eisenberg, F., Jr. (1980) J Biol Chem 255, 8458-8464

Mauck, L. A., Wong, Y. H., and Sherman, W. R. (1980) Biochemistry 19, 3623-
3629 '

Pittner, F ., and Hoffmann—Ostenhof, O. (1976) Hoppe Seylers Z Physiol Chem
357, 1667-1671

Pittner, F., and Hoffrnann-Ostenhof, O. (1978) Hoppe Seylers Z Physio] Chem
359, 1395-1400

Chen, L., Zhou, C., Yang, H., and Roberts, M. F. (2000) Biochemistry 39, 12415-
12423

Kiely, D. E., and Sherman, W. R. (1975) J Am Chem Soc 97, 6810-6814

Stein, A. J ., and Geiger, J. H. (2002) J Biol Chem 277, 9484-9491

45

73.

74.

75.

76.

77.

78.

79.

80.

81.

82.

Norman, R. A., McAlister, M. S., Murray-Rust, J., Movahedzadeh, F., Stoker, N.
G., and McDonald, N. Q. (2002) Structure (Comb) 10, 393-402

Rossmann, M. G. (1976) Fed Proc 35, 2112-2114

Frey, P. A. (1987) In Pyridine Nucleotide Coenzyes: Chemical, Biochemical and
Medicinal Aspects. Part B. (Avramovic, 0., Ed.), Vol.2, John Wiley & Sons, Inc.,
New York

Byun, S. M., and Jenness, R. (1981) Biochemistry 20, 5174-5177

Tian, F., Migaud, M. E., and Frost, J. W. (1999).]. Am. Chem. Soc. 121, 5795-
5796

Migaud, M. E. (1996) Ph. D. Dissertation, Michigan State University

Migaud, M. E., and Frost, J. W. (1995) J. Am. Chem. Soc. 117

Floss, H. G. (1997) Nat Prod Rep 14, 433-452

Carpenter, E. P., Hawkins, A. R., Frost, J. W., and Brown, K. A. (1998) Nature
394, 299-302

Gumber, S. C., Loewus, M. W., and Loewus, F. A. (1984) Arch Biochem Biophys

231, 372-377

46

CHAPTER II
CRYSTALLIZATION AND PRELIMINARY X-RAY DIFFRACTION ANALYSIS OF
THE TFIIIB COMPLEX

In order to determine the structure of TFIIIB, the central initiation factor of Pol III
transcription, recombinant mutants of individual members of TFIIIB, namely BRF, B”,
TBP, and different promoter DNAS were designed based on published and unpublished
data (1-4). This chapter will describe the experimental procedures of purification of
BRF, B”, TBP, and DNAS, crystallization, as well as preliminary X-ray diffraction

analysis.

2.1 BRF

Crystals of a full-length member of a TFIIB-like protein have not been obtained.
In fact, it has been necessary to remove the Zn ribbon domain in order to obtain crystals
in all crystal structures of TFIIB-like proteins published so far (5-8). An NMR structure
of the C-terminal repeat domains of TFIIB indicates that there are only weak interactions
between the two cyclin domains in the absence of TBP binding (9). Since the TFIIB-
homology N-terminal region of BRF does not detectably interact with TBP/DNA, it may
be difficult to elucidate the structure of this region. Therefore, we decided to focus on the
C-terrninal region of BRF. Figure 2.1 shows the deletion mutants of BRF that were
constructed for the structural investigations; these include six S. cerevisiae BRF mutants
and four K. lactis BRF mutants. All BRF mutant plasmids were generated by subcloning
the coding sequence for BRF with a six-His tag at the C-terrnini into pET vectors. Our

collaborator, Dr. Steven Hahn, prepared and provided all plasmids.

47

S. cerevisiae

(Full length):
311-596:
407-531:

435-596 (A530-558)‘

435-531:
420-531:
420-551

K. lactis
(Full length):
302-501:
302-556:
395-501:
395-556:

11

Figure 2.1 BRF constructs used in this study.

48

homology regions
./ I \.

2863.04 439 515 570595
W596

311
I\
407 531
.\\\\\\\\\\§

435 559

 

VO-
M
77

51
W

homology regions

I l I
302320 409 485 529554
W

556
302 301

W.

39
W

~.\

WW

2.1.1 Over-expression and purification of BRF
2.1.1.1 Over-expression of BRF

All S. cerevisiae BRF plasmids were transformed into E. coli BL21 (DE3) strains.
The LB medium containing 100mg/L ampicillin was inoculated by a single colony of
transformed E. coli BL21 (DE3) cells. Growth conditions of 37 °C and rapid shaking at
250 rpm were maintained for 18 hours or until the medium was saturated with cells. One
liter of LB/ampicillin was then inoculated with the saturated culture and growth
continued until an CD. at 600 nm of 0.6. The inducer, isothiopropyl-B-galactoside
(1PTG), was then added to a final concentration of 0.5 mM, and shaking continued for
another 3 hours. Cells were then harvested by centrifiigation at 5000 rpm.
All K lactis BRF plasmids were transformed into E. coli BL21 (DE3) codon plus cells,
all growth media contained 34 mg/L chloramphenicol in addition to 100 mg/L arnpicillin,

but the rest of the protocol remained the same.

2.1.1.2 Purification of BRF

Cells were re-suspended in the lysis buffer (6M Guanidine-HCI, 0.1M NaH2PO4, 0.01M
Tris-Cl, pH 8.0), and sonicated using 2 or 3 one-minute pulses on ice. The lysate was
then centrifuged for 20-30 minutes at 15000 rpm to spin the debris down; the proteins

were in the supernatant.

Ni-NTA affinig chromatography
Ni-NTA agarose (Qiagen) was used for 6x His-tagged purification. Buffers used for

purification of BRF proteins under denaturing conditions were:

49

o Lysis buffers
Buffer A: 6 M Gu-HCl; 0.1M NaH2PO4; 0.01M Tris-Cl, pH 8.0
Buffer B: 6 M Urea; 0.1M NaH2PO4; 0.01M Tris 'Cl, pH 8.0

0 Wash buffer
Buffer C: 6 M urea; 0.1M NaH2PO4; 0.01M Tris °Cl, pH 6.8

o Elution buffers
Buffer D: 6 M urea; 0.1M NaH2PO4; 0.01M Tris -Cl, pH 4.5
Buffer E: 6M Gu-HCl; 0.2M Acetic acid

Ni-NTA superflow slurry was resuspended completely and poured into the

column to allow the resin to settle. The column was then equilibrated with 5 bed
volumes of the lysis buffer. The above lysate was loaded onto the column and washed
with the lysis buffers and wash buffer, and the BRF was eluted with the elution buffers.
Fraction samples were analyzed by the Bradford method and SDS-PAGE. The purity of
the protein can be seen in the SDS-PAGE shown in Figure 2.2. Fractions containing

BRF of similar purity were pooled together for further preparation.

12345678910

 

Figure 2.2 SDS-PAGE ofNi-NTA affinity purified K. lactis BRF 395-501.
Marker; 2. Flow through; 3. Wash A; 4. Wash B; 5. Wash C; 6-10. Elutions in 5 fractions
in order. M.W. marker: Purple 42,000, Orange 32,000, Red 17,900, and Blue 7,200.

50

Protein re oldin

Proteins were diluted in a dilution buffer (6M Urea, 10% Glycerol, 20mM Tris-Cl,
500mM KCl, 5mM DTT) to about 0.5mg/ml. They were then dialyzed against the
refolding buffer (10% Glycerol, 20mM Tris-C1, 500mM KCl, 2mM EDTA, 5mM DTT,
lmM PMSF) for 6 hours. The refolding buffer was changed and the dialysis was

repeated three more times.

FPLC ion exchange chromatogmphy

Pharmacia FPLC equipment was used to purify refolded proteins by ion exchange
chromatography. Source 15Q matrix (anion exchanger with an average particle size of
15pm) was used as the stationary phase. Refolded BRF proteins were diluted 3 fold with
a buffer containing 10% Glycerol, 20mM Tris~Cl, pH 8.5, 5mM DTT, then loaded onto a
Source-Q column, and a gradient was run from 50 mM KCl (buffer A: 10% Glycerol,
20mM Tris.Cl, pH 8.5, 50mM KCl, 5mM DTT) to IM KCl (buffer B: 10% Glycerol,
20mM Tris.Cl, pH 8.5, 1M KCl, 5mM DTT). BRF proteins elute at the salt
concentration of about ZOOmM. The fractions were analde using the Bradford assay
and SDS-PAGE. The fractions containing pure BRF proteins were buffer-exchanged in
10% Glycerol, 20mM Tris'Cl, pH 8.5, 100mM KCl, 5mM DTT, and concentrated using
Centriprep-IO (Amicon) by centrifugation at 6000 rpm. The final concentration of BRF
was from 5mg/ml to 35 mg/ml for the purpose of crystallization. Figure 2.3 shows the

purified BRF constructs on SDS-PAGE.

51

A. Over-expressed and purified S. cerevisiae BRF constructs

1234567

B. Over-expressed and purified K. lactis BRF constructs

Figure 2.3 Over-expressed and purified BRF constructs used in crystallization.

(A) 1. M.W. marker 2. S. cerevisiae BRF311-596 3. S. cerevisiae BRF407-531

4. S. cerevisiae BRF435-596 (A530-558) 5. S. cerevisiae BRF435-531 6. S. cerevisiae
BRF435-551 7. S. cerevisiae BRF 420-551.

(B) 1. M.W. marker 2. K. lactis BRF302-501 3. K. lactis BRF395-501 4. K. lactis
BRF302-556 5. K. lactis BRF395-556.

M.W. marker: Purple 42,000, Orange 32,000, Red 17,900, and Blue 7,200.

52

2.1.2 Crystallization and data collection of BRF

In order to obtain well diffracting crystals of BRF functional domains, extensive
screening of conditions was performed on all 10 constructs of BRF. Both the hanging
drop vapor diffusion method (Figure 2.4) and the batch crystallization method (Figure
2.5) were used. In the hanging drop vapor diffusion method, a drop containing a mixture
of protein and reservoir solution is equilibrated against the reservoir. In the batch
crystallization method, the protein and precipitant are mixed directly under oil to reach
super saturation. Nearly 3000 conditions were screened to produce the first crystal of
BRF.

Crystals of S. cerevisiae BRF435-531 grew over 10 months in a 1:1 mixture of
protein solution (10mg/m1) and reservoir solution (8 % PEG 8000, 0.1 M Tris, pH 8.5) at
room temperature. The crystals of BRF are exquisitely sensitive to small changes in
protein concentration, precipitant concentration, and pH. Despite the difficulty of
obtaining crystals, crystals of S. cerevisiae BRF 43 5-531 were reproduced and improved.
Crystals of S. cerevisiae BRF435-596 (A530-558) and K lactis BRF395-501 were also
grown with similar difficulty (Table 2.1). An example of a K. lactis BRF395-501 crystal
is depicted in Figure 2.6. The crystals were cryoprotected using a cryoprotectant solution
containing 30% glycerol, 8 % PEG 8000, 0.1 M Tris, pH 8.5 for data collection. Despite
the difficulty of obtaining crystals and the small crystal sizes, data were collected on an

S. cerevisiae BRF435-531 crystal. Data collection statistics are listed in Table 2.2.

53

_ . coverslip
Protein + reservorr solution $

\

1 U !\

i H O grease
2

 

 

reservoir

 

 

 

Figure 2.4 Hanging drop vapor diffusion crystallization method.

Protein + precipitant oil

 

 

 

 

Figure 2.5 Batch crystallization method.

54

Table 2.1 BRF crystals

 

 

 

 

 

 

 

 

 

 

Protein Buffer Reservoir Resolution
concentration conditions
S. cerevisiae 8.0 mg/ml 20 mM Tris, pH 8 % PEG 8000, 3.5 A
BRF 435-531 8.0, 100 mM KCl, 0.1 M Tris,
10 % Glycerol, 5 pH 8.5
mM DTT
S. cerevisiae 6.0 mg/ml 20 mM Tris, pH 8 % PEG 8000, 4.0 A
BRF 435-596 8.0, 100 mM KCl, 0.1 M Tris,
(A530-558) 10 % Glycerol, 5 pH 8.5
mM DTT
K. lactis BRF 20.0 mg/ml 20 mM phosphate, 9.5 ~10.5 % 3.0 A
395-501 pH 8.0 PEG 8000, 0.1M
Tris, pH 8.5
0.08 mm

Figure 2.6 A crystal of K. lactis BRF 395-501, with dimensions of 0.08 X 0.08 x 0.005

1111113.

55

 

Table 2.2 Data collection statistics of S. cerevisiae BRF 435-531 crystal*I

Wavelength (A) 0.938

Resolution range (A) 30.0-3.5 (362-3.5)

Space group P2

Unit cell parameters a=l3l.087 A, b=90.773 A, c=143.59 A, B=lOS.314°

Matthew’s coefficient (AI/Dalton) 3.94
Solvent content (%) 69

Molecules per asymmetric unit ~ 20

Number of observations 139555
Unique reflections 37517
Completeness (%) 88.4 (90.1)
1/0 4.47 (2.73)
R,,.,,(I)l (%) 32.4 (62.0)

* Data were collected at the Advanced Photon Source, BIOCARS BM14 beamline.

I Values in parenthesis refer to the last resolution shell.

, where I is an individual intensity measurement and <I> is the

average intensity for this reflection, with summation over all data.

56

The crystals of S. cerevisiae BRF 435-531 belong to the space group P2 with the
unit cell parameters a=l3l.087 A, b=90.773A, c=143.59 A, B=103.314°. Given the
relatively small size of the protein, this unit cell is huge with about 20 molecules in the
asymmetric unit, and a very high solvent content of 69 %. The length of time for crystal
growth, difficulty of reproduction, low diffraction quality, and complex asymmetric unit
have hindered progress towards structure determination of this functional domain of

BRF.

2.2 B”

For structural investigation, a deletion mutant of S. cerevisiae B” was generated
by subcloning the coding sequence for B” 240-520 with a six-His tag at the C-termini
into a pET vector by our collaborator, Dr. Steven Hahn. This mutant B” contains the
SANT domain and was found to have full activity in binding TBP-DNA.

2.2.1 Over-expression and purification of B”
2.2.1.1 Over-expression of B”

The plasmid encoding B”240-520 was transformed into E. coli BL21 (DE3) cells.
The concentrations of ampicillin and kanamycin in all growth media were 50 mg/L and
30 mg/L respectively. An overnight culture was grown from a single colony of
transformed E. coli BL21 (DE3) cells in the LB media containing ampicillin and
kanamycin, and then diluted 250 fold into fresh LB media containing ampicillin and
kanamycin. The culture was incubated at 37°C with shaking at 250 rpm until the CD. at

600 nm reached 0.5-0.6. The inducer IPTG was then added to a final concentration of

57

0.5mM, and shaking continued at 30°C for 3 more hours. The cells were then spun down
by centrifugation at 5000 rpm and stored at —80°C.

2.2.1.2 Purification of B”

Ni NTA affmity purification

Ni-NTA agarose (Qiagen) was used for 6x His-tagged purification, buffers used

for purification of B” under native conditions are:

0 Lysis Buffer
20mM HEPES, 300mM NaCl, 10% Glycerol (pH 8.0), add 5mM B— mercaptoethanol
(BME) and protease inhibitors (PMSF, Benzamidine, Leupeptin, Pepstatin, Chymostatin)
before use.

- Binding/W ash Buffer
Lysis Buffer + 20mM imidazole (pH 8.0), add BME and PMSF before use.

0 Elution Buffer
Lysis Buffer + 100mM imidazole (pH 8.0), add BME and PMSF before use.

Ni-NTA agarose (Qiagen) was spun at 4°C to remove the storage buffer and
equilibrated in 30ml binding buffer, again spun and the buffer was removed. Cells were
resuspended in the lysis buffer (50ml per 2L of cells) on ice and sonicated. The crude
extract was spun down by centrifugation at 15000 rpm. The supernatant containing B”
was added to the pre-equilibrated agarose and incubated at 4°C for 60 minutes. Protein-
bound agarose was spun for the supernatant to be removed and washed twice with 40ml
of the wash buffer. Protein-bound agarose was then packed into an HR 10/ 10 column,
and washed with 5 bed volumes of the wash buffer at a flow rate 2ml/min, protein was

eluted with 7 bed volumes of the elution buffer at a flow ratelml/min. Upon analysis by

58

the Bradford method and SDS-PAGE (Figure 2.7), fractions containing B” were pooled

together for further purification.

 

Figure 2.7 SDS-PAGE of S. cerevisiae B” 240-520 after Ni-NTA affinity purification;

the presence of degradation products and impurities requires further purification.

FPLC ion exchange chromatography

Pharmacia F PLC equipment was used to purify Ni-NTA purified B” by ion
exchange chromatography using Source 15S matrix (cation exchanger with an average
particle size of lSum) as the stationary phase. Ni-NTA affinity chromatography-purified
proteins were buffer-exchanged into buffer A (10% Glycerol, 20mM HEPES, pH 8.0, 75
mM NaCl, 5mM BME, 1 mM PMSF). A gradient was run against buffer B (10%
Glycerol, 20mM HEPES, pH 8.0, 500 mM NaCl, 5mM BME, 1 mM PMSF). B” elutes
over most of the gradient. This indicates the unusual conformational and/or aggregation
heterogeneity of B”. Therefore, it was thought that refolding BRF and B” together might

lead to the production of a more homogeneous material.

59

2.2.2 Refolding and purification of the BRF/B” complex
Refolding of BRF and B”

Ni-NTA purified BRF and B” 240-520 were combined together in 1:1 molar ratio,
and this mixture was diluted with a denaturing buffer (6M Urea, 10% Glycerol, 20 mM
Tris-Cl, pH 8.0, 500mM KCl, 5mM DTT) to about 0.4mg/ml. The mixture was dialyzed
against the refolding buffer (10% Glycerol, 20 mM Tris-Cl, pH 8.0, 500mM KCl, 2mM
EDTA, 5mM DTT, lmM PMSF) for 6 hours. The refolding buffer was replaced fresh
and the dialysis was repeated three times.
Purification of the BRF/ B” complex

Refolded BRF/ B” complex was buffer-exchanged into buffer A (10 % Glycerol,
20 mM Tris-Cl, pH 8.0, 50 mM KCI, 2mM EDTA, 5mM DTT, lmM PMSF), and loaded
onto an FPLC Source-Q column. A salt gradient was run against buffer B (10 %
Glycerol, 20 mM Tris-Cl, pH 8.0, 2mM EDTA, 5mM DTT, lmM PMSF). The BRF/B”
complex eluted at about 130 mM KCl in a nice sharp peak, in contrast to the broad range
of peaks of B” alone. SDS-PAGE (Figure 2.8) confirmed that the peak fractions indeed

contain the purified BRF/B” complex.

 

Figure 2.8 SDS-PAGE of the BRF/B” complex purified by FPLC.

M.W. marker: Purple 42,000, Orange 32,000, Red 17,900, and Blue 7,200.

2.2.3 Crystallization of the BRF/B” complex

Two BRF/B” complexes listed in Table 2.3 were extensively screened for
crystallization using the hanging drop vapor diffusion method (Figure 2.4).
Unfortunately, neither of them produced crystals. Efforts are being made towards

crystallization of the BRF/B”/I‘BP/DNA complexes as will be described in section 2.4.

Table 2.3 The BRF/B” complexes used in crystallization attempts.

 

 

 

 

 

BRF B”

S. cerevisiae BRF 435-531 S. cerevisiae B” 240-520
K. lactis BRF 395-501 S. cerevisiae B” 240-520
2.3 TBP

The C-terminal conserved domain of S. cerevisiae TBP was used in attempts to
produce crystals of BRF/TBP/DNA complexes. This domain was chosen based on the
previous success on structure determination of the TBP/TATA and the TFIIA/TBP/DNA
complexes (10,11). We have worked on both the wild-type TBP and a deletion mutant,

TBP 56-240 with an N-terminal 6x His tag.

2.3.1 Over-expression and purification of wild-type TBP

Over-expression of wild-gpe TBP

In the process of optimizing the over-expression level, the plasmid encoding wild-

type S. cerevisiae TBP was transformed into several different strains of E. coli cells,

61

 

including BL21 (DE3), BL21 (DE3) codon plus, BL21 (DE3) pLys S, Tuner, and Tuner
pLys S. It has been found that wild-type S. cerevisiae TBP is over-expressed in BL21
(DE3) codon plus and Tuner pLys S cells equally well and better than all the rest of the
E. coli strains so far tested.

An LB medium containing carbenicillin and chloramphenicol was inoculated with
a single colony of transformed E. coli BL21 (DE3) codon plus or Tuner pLys S cells.
The concentrations of carbenicillin and chloramphenicol in all growth media were 100
mg/L and 34 mg/L respectively. Growth conditions of 37 °C and rapid shaking at 250
rpm were maintained for 18 hours or until saturated. The saturated culture was then
diluted 40 fold into fresh LB media containing carbenicillin and chloramphenicol, and
growth continued until an CD. at 600 nm of 1.0 was reached. The inducer IPTG was
then added to a final concentration of 0.4 mM, and shaking continued for another 3 hours

at 30 °C. Cells were then harvested by centrifugation at 5000 rpm.

Pm’ficatjon of wild-type TBP
Cells were resuspended in the lysis buffer (30 mM Tris, pH 7.5, 10 % Glycerol,

50 mM KCl, 1 mM DTT, 1 mM PMSF, 2 mM EDTA), and sonicated in a dry ice/Ethanol
bath. The crude extract was spun down by centrifugation at 15000 rpm. The supernatant
containing TBP was loaded onto a Phast-Q (Pharmacia) in tandem with a Phast-S
(Pharmacia) column that were pre-equilibrated with the lysis buffer. While most E. coli
proteins bind to the Phast-Q column, TBP binds to the Phast-S column. The columns
were washed with the lysis buffer until the CD. at 280 nm of the flow through reached

lower than 0.1. The Phast-Q column was then disconnected, and a gradient was run on

62

the Phast-S column from 0.2 M KCl to 0.6 M KCl in the buffers containing 30 mM Tris,
pH 7.5, 10 % Glycerol, 1 mM DTT, 1 mM PMSF, 2 mM EDTA. Fractions containing
TBP detected by UV-vis and evaluated by SDS-PAGE were pooled together and
concentrated to about lmg/ml for proteolysis as will be described in section 2.3.2.
Source 15S matrix (cation exchanger with an average particle size of lSum) was
used for ion-exchange chromatography. Proteolyzed TBP was buffer-exchanged into
buffer A (10% Glycerol, 30mM Tris, pH 7.5, 50 mM KCl, 1 mM DTT, 1 mM PMSF, 2
mM EDTA), and loaded onto an F PLC Source-S column. A salt gradient was run against
buffer B (10% Glycerol, 30mM Tris, pH 7.5, l M KCl, 1 mM DTT, 1 mM PMSF, 2 mM
EDTA). The eluted fiactions were analyzed by the Bradford assay and SDS-PAGE. The
fractions containing pure TBP were concentrated for BRF/TBP/DNA complex formation

and crystallization.

2.3.2 Proteolysis of wild-type TBP

In order to remove the unconserved N-terminal segment that is dispensable for
basal transcription and yeast cell viability (12-14), proteolysis of wild-type S. cerevisiae
TBP was performed using two enzymes, endoproteinase Lys-C and Trypsin.
Proteolysis with endoproteinase Lys-C

The enzyme endoproteinase Lys-C purchased from Roche was dissolved in the
storage buffer (50 mM Tricine, pH 8.0, 10 mM EDTA) in 15 units/ml. Phast-S purified
and concentrated TBP was introduced into the proteolysis buffer (50 mM Tricine, pH 8.0,

20 % Glycerol, 200 mM KCl, 5 mM DTT); one unit of the endoproteinase Lys-C was

63

added to each 20 mg of TBP to carry out the proteolysis. Reaction was done at room
temperature for 4 hours followed by additional 12 hours at 4 °C.
Many trials of proteolysis with endoproteinase Lys-C have been unsuccessful; a

different approach to obtain proteolyzed TBP became necessary.

Proteolysis with Tgpsin

Immobilized Trypsin purchased from PIERCE was used in proteolysis attempts.
The immobilized Trypsin gel was washed with 3 bed volmnes of the digestion buffer
(0.1M NH4HCO3, pH 8.0) for three times, then added to TBP in the proteolysis buffer (30
mM Tris, pH 7.5, 10 % Glycerol, 250 mM KCl). The reaction mixture was incubated at
room temperature with gentle shaking at 100 rpm for 2 hours. The reaction then
continued at 4 °C for additional 12 hours. The immobilized Trypsin was then separated
from the reaction mixture by centrifugation. This method successfully proteolyzed TBP

at the desired site. Figure 2.9 shows an example of SDS-PAGE of the proteolysis result.

p—

2 3 4

‘23 III

.4 .. ,9 —-Full length TBP
--Truncated TBP

Figure 2.9 SDS-PAGE of the proteolysis of TBP.
1. M.W. marker; 2. Before the proteolysis (at the time point zero); 3. After 2 hours of
reaction at room temperature; 4. After 12 hours at 4°C (the stop point).

M.W. marker: Purple 42,000, Orange 32,000, Red 17,900, and Blue 7,200.

64

2.3.3 Over-expression and purification of TBP 56-240
Over-expression of TBP 56-240

The plasmid encoding S. cerevisiae TBP 56-240 with an N-terminal 6x His tag
was transformed into one of several different strains of E. coli cells, including BL21
(DE3), BL21 (DE3) codon plus, BL21 (DE3) pLys S, Tuner, and Tuner pLys S with
Tuner pLys S cells producing the optimal over-expression. The cell growth condition

was the same as that of wild-type TBP.

Purification of TBP 56-240

Ni-NTA affinity purification under native conditions was performed. Cells were
re-suspended in the lysis buffer (50mM Tris, pH 7.5, 20% Glycerol, 500 mM KCl, lmM
PMSF, 5mM BME, and 3 tablets/L protease inhibitor cocktail, EDTA fi'ee). Cells were
lysed by sonication and centrifuged. The supernatant containing TBP was loaded on a
Ni-NTA column that was pre-equilibrated with the lysis buffer, and the column was
washed with the lysis buffer, wash bufiers (lysis buffer plus 20mM imidazole and lysis
buffer plus 100mM imidazole). Subsequently, TBP was eluted with elution buffers (lysis
buffer plus 250mM imidazole and lysis buffer plus 400mM imidazole). Emotions
containing TBP identified by the Bradford method and SDS-PAGE were pooled together
for further purification. The 6x His tag at its N-terminus may affect the formation of
protein-TBP-DNA complex and prevent the formation of crystals, so the Ni-NTA
purified TBP was proteolyzed using Thrombin before purification by ion exchange

chromatography as will be described in section 2.3.4.

65

For FPLC ion exchange purification, protein was diluted with no salt buffer (30
mM Tris, pH 7.5, 10 % Glycerol, 1 mM DTT, 1 mM PMSF, 3 tablets/L protease inhibitor
cocktail, EDTA-free), and loaded on a Source-S (cation exchange) column. A salt
gradient was run from 30mM KCl to 1000mM KCl. The fractions containing TBP were
identified, pooled together, and concentrated for the BRF/TBP/DNA complex formation

and crystallization setups.

2.3.4 Proteolysis of TBP 56—240

Purified TBP with its N-terminal 6x His tag was successfully complexed with
BRF and DNAS, and these complexes were crystallized. However, compared to the
crystals of complexes formed of TBP with its His tag removed, they were evaluated to be
of lower quality based on their appearances and small sizes. Thrombin was used for the
removal of the N-terminal 6x His tag, Biotinylated Thrombin was much more effective
than free Thrombin. Biotinylated Thrombin was buffer-exchanged to the proteolysis
buffer (30 mM Tris, pH 7.5, 10 % Glycerol, 250 mM KCl) before the reaction. For each
milligram of TBP, 1.5 units of Biotinylated Thrombin were sufficient for a successful
proteolysis. Figure 2.10 shows an example of the proteolysis being monitored at
different time points and with varied amount of the enzyme used. After the proteolysis is
completed, Biotinylated Thrombin was removed by adding streptavadin into reaction
mixture followed by filter centrifugation, only TBP but not thrombin remained in the

filtrate by this method.

66

4..., _, -—-- - ‘ " —— His-tagged TBP
.. \'. M ”'x TBP with no His tag

Figure 2.10 Monitored proteolysis of TBP by Biotinylated Thrombin.

1. M.W. Marker. Purple 42,000, Orange 32,000, Red 17,900, and Blue 7,200;
2. After 12 hours of reaction when 1.5 units thrombin/mg of TBP was used;

3. After 12 hours of reaction when 0.75 unit thrombin /mg of TBP was used;
4. After 12 hours when no thrombin was used;

5. After 18 hours of reaction when 1.5 units thrombin/mg of TBP was used;

6. After 18 hours of reaction when 0.75 unit thrombin/mg of TBP was used;

7. After 18 hours when no thrombin was used.

67

2.4 The BRF/TBP/DNA and BRF/B”/TBP/DNA complexes

The primary goal of this research is to understand the transcription initiation
mechanism of the Pol III transcription system; this goal can be achieved through
elucidation of the structural details of the TFIIIB complex bound to a promoter DNA
sequence. All purified BRF constructs and B” described above were evaluated for
TBP/DNA binding activities by electrophoretic mobility shift assay (EMSA) by our
collaborator, Dr. Steven Hahn. From the EMSA results shown in Figures 2.11-2.14, all
BRF constructs and the B” construct that we have over-expressed and purified are active
in BRF/TBP/DNA and BRF/B”/I‘BP/DNA complex formation, thus all can be used for

the crystallization of BRF/TBP/DNA and BRF/B”/TBP/DNA complexes.

68

s.eerevlslae BRF311-595 S.oerovlslae BRF407.531

 

 

n9 BRF — -- 0.25 0.5 1.0 1.5 2.0 0.02 0.04 0.06 0.03 0.1
ng TBPc -- 1.0 >
_BRF-TBP-DNA
, .- ——BRF-TBP-DNA
“’1 M
""4”“ *" ' ' . \TBP-DNA

1 2 3 4 5 6 7 8 9 10 11 12

 

 

 

A
S. cerevislae BRF311-596 $.0erevlslae BRF407.531
ng WT B" - - -- 1 2 4 8 - 1 2 4 8
n9 BRF -- - 2 > 0.2 %
ng TBPc - 1 5

_B"-BRF-TBP-DNA
_BRF-TBP-DNA
j, —BRF-TBP~DNA

. \TaP-DNA

 

Figure 2.11 Gel mobility shift assays using] ng S. cerevisiae TBP56-240, and indicated
amounts of BRF and B". The probe is the U6 promoter with its TATA box modified to
that of the AdMLP TATA box (U6-MLP promoter). Binding reactions for 45 min. at
room temperature run on 6% TGOE (25 mM Tris, 190 mM glycine, pH 8.3) with no
MgOAc gel for 60 min. at 4 °C.

(A) Titrations of S. cerevisiae BRF311-596 and S. cerevisiae BRF407-53 l.

(B) Titrations of WT S. cerevisiae B" with either 2ng S. cerevisiae BRF31 1—596 or 0.2
ng S. cerevisiae BRF407-531.

Experiment done by Dr. Steven Hahn.

69

vaislae BRF407.531 S.cerevlslao BRF435.595 (A 530.553)

119 BRF - - 0.01 0.02 0.04 0.06 0.00 0.01 0.02 0.04 0.06 0.08
ng TBPc - 1 T

V

" ' —BRF-TBP-DNA

 

 

 

.. — TBP-DNA
S.0erevlslae Bane-[.531 S.0erevlsiae BRF435531
ngBRF - - 0.01 0.02 0.04 0.00 0.00 0.01 0.02 0.04 0.00 0.00
ng TBPc .. 1 Ar
—-BRF-TBP DNA
_TBP-DNA

      

13°14". ‘5” 0 7 3 9 10 11 12

Figure 2.12 Gel mobility shift assays of S. cerevisiae BRF constructs usingl ng

S. cerevisiae TBP56-240, and indicated amounts of BRF. The probe is the U6 promoter

with its TATA box modified to that of the AdMLP TATA box (U6-MLP promoter).

Binding reactions for 45 min. at room temperature run on 6% TGOE (25 mM Tris, 190

mM glycine, pH 8.3) with no MgOAc gel for 90 min. at 4 °C.

(A) Titrations of S. cerevisiae BRF407-531 and S. cerevisiae BRF435-596 (A 530-558).

(B) Titrations of S. cerevisiae BRF407-531 and S. cerevisiae BRF435-531.

Experiment done by Dr. Steven Hahn.

70

S.cerevlsiae K. lactis K. lactis K. lactis K. lactis
311-596 302-501 395-501 395-556 302-556

I 11 11 fit II I
ng BRF -. .. 0.2 2 0.2 2 0.2 2 0.2 2 0.2 2

ng TBPc -- 2 >

 

 

I, _ S.cerevlslaa BRF311-596
W P- DNA

a... W ~ ; . ~m 2eilril\ — K. lactis BRF-TBP-DNA

\TB P-DNA

‘
hmwfihnmmmw‘

123456789101112

Figure 2.13 Gel mobility shift assays of K. lactis BRF constructs using 2 ng S. cerevisiae
TBP 56-240, and indicated amounts of BRF. The probe is the U6 promoter with its
TATA box modified to that of the AdMLP TATA box (U6-MLP promoter). Binding
reactions for 45 min. at room temperature run on 6% TGOE (25 mM Tris, 190 mM
glycine, pH 8.3) + 0.2 mM MgOAc gel for 90 min. at 4 °C. Experiment done by Dr.

Steven Hahn.

71

wr S.cerevisiae B" S-cerevlslae 3" 240-520

 

 

"9 B" - - - 0.25 0.5 1.0 2.0 0.06 0.12 0.25 0.5 1.0
ng BRF .. 0.2 >
"4935-5I’6c (A530-558) A

1 r

 

_ B"-BRF-TBP-DNA

N“ A‘if‘l‘ W ”W— B"- BRF-TBP- DNA
— BRF- TBP- DNA

1* ”m ' —TBP-DNA

   
   
      

11‘1” 12

 

10

 

 

 

 

   

wr S.cerevlsiae B“ S.cerevisiae B" 240-520
“9 B" — - 0.25 0.5 1.0 2.0 0.06 0.12 0.25 0.5 1.0
ng BRF __ _ 0.2 >
435-531

ng TBPc >

_ 2.... m w — B"-BRF-TBP- DNA

. " 2. , .2WMM_B'-BRF-TBP-DNA

, — BRF-TBP-DNA
—T8P-DNA

12 3 4 5 s 7""s"”9 10 1112

Figure 2.14 Gel mobility shift assays of S. cerevisiae B" usinglng TBP 56-240, 0.2ng
BRF, and indicated amounts of B". The probe is the U6 promoter with its TATA box
modified to that of the AdMLP TATA box (U6-MLP promoter). Binding reactions for
45 min. at room temperature run on 6% TGOE (25 mM Tris, 190 mM glycine, pH 8.3)
with no MgOAc gel for 90 min. at 4 °C. (A) Titrations of WT S. cerevisiae B" and

S. cerevisiae B"240-520 with 0.2ng S. cerevisiae BRF435-596 (A 530-558).

(B) Titrations of WT S. cerevisiae B" and S. cerevisiae B"240-520 with 0.2 ng

S. cerevisiae BRF435-531. Experiment done by Dr. Steven Hahn.

72

2.4.1 DNAs

Previous mutagenesis, DNA footprinting, and photocrosslinking results have
provided information on the architecture of TFIIIB on the promoter DNA. In addition to
being responsible for most of the TBP binding by interacting predominantly with the N-
terrninal top and stirrup of TBP, BRF makes most of its DNA contacts within a 15 bp
region immediately upstream of the TBP binding site. B”, on the other hand, interacts
with DNA within 10 bp immediately downstream of the TBP binding site(2,3,15-17).

According to these data, DNAs were designed for complex formation and crystallization.

Purification of DNAS

Oligonucleotides were ordered from the Yale W.M. Keck Facility.
Oligonucleotides were purified using Perkin Elmer HPLC equipment on a Source-Q
column; UV absorbance at 260 nm was monitored. A strand of DNA was loaded onto
the Source-Q column with buffer A (10 mM NaOH, 0.2 M NaCl), and eluted with a
gradient against buffer B (10 mM NaOH, 1.0 M NaCl). Collected fractions were
neutralized with 1 M Tris, pH 7.5, then diluted with10 mM Tris, pH 7.5 and loaded on a
DEAE column. DNAs were then eluted with a buffer of 10 mM Tris, pH 7.5, l M NaCl.
Eluted DNAs were concentrated and annealed in equimolar amounts. Table 2.4 and
Table 2.5 list all of the Oligonucleotides used in TFIIIB complex crystallization attempts.
For DNAs (Table 2.5) that were used for the BRF/B”/TBP/DNA complex formation, four

strands with overhangs were annealed together in equimolar amounts.

73

Table 2.4 DNAs utilized in crystallization of the BRF/TBP/DNA complexes

 

DNA sequence

 

DNA]

DNA2

DNA3

DNA4

DNAS

DNA6

DNA7

DNA8

DNA9

DNAlO

DNAll

DNA12

DNA13

DNA14

 

 

 

TACTATAAAAQAATGI I I I I I ICGCATA
ATGATAI I I ICTTACAAAAAAAGCGTAT

AATACTATAAAAGAATGI I l I I I ICGCA
TTATGATAI I I ICTTACAAAAAAAGCGT

TACTATAAAAGAATGI I I I I I ICGCAT
ATGATAI I I ICTTACAAAAAAAGCGTA

ATACTATAAAAGAATGI I I I I I ICGCA
TATGATAI I I ICTTACAAAAAAAGCGT

TACTATAAAAGAATGTTAATTTCGCA
ATGATAI I I ICTTACAATTAAAGCGT

TACTATAAAAGAATGI I I I I I ICGC
ATGATAI I I ICTTACAAAAAAAGCG

TACTATAAAAGAATGI I I I I I ICG
ATGATAI I I ICTTACAAAAAAAGC

CTATAAAAGAATGI I I I I I ICGCA
GATATT'TTCTTACAAAAAAAGCGT

TTACTATAAAAGAATGI I I I I I ICGC
AATGATAI I I ICTTACAAAAAAAGCG

TTACTATAAAAGAATGI I I I I I ICGC
ATGATAI I I ICTTACAAAAAAAGCGT

ATACTATAAAAGAATGI I I I I I ICGC
ATGATAI I I ICTTACAAAAAAAGCGA

GTACTATAAAAGAATGI I I I I I ICGC
ATGATAI I I ICTTACAAAAAAAGCGC

T‘TACIATflAGAATGI I I I I I ICGC
ATGATAI I I ICTTACAAAAAAAGCGA

ATACTATAAAAGAATGTTTITITCGC
ATGATATTITCTTACAAAAAAAGCGT

 

74

 

Table 2.5 DNAs utilized in crystallization of the BRF/B”/TBP/DNA complexes

 

DNA sequence

 

DNA ACTAI I I ICGGCTA CTATAAAAGAATGI I I I I I ICGCA
15 TGATAAAAGC CGATGATAI I I ICTTACAAAAAAAGCGT

DNA CGTCCACTAI I I I CGGCTACTATAAAAGAATGI I I I I I ICGCA
16 GCAGGTGATAAAAGCCGAT GATAI I I ICTTACAAAAAAAGCGT

DNA CGTCCACTAI l I IC GGCTACTATAAAAGAATGI I I I I I ICGCAAC
17 GCAGGTGATAAAAGCCGAT GATAI I I ICTTACAAAAAAAGCGTTG

 

 

 

 

2.4.2 Crystallization of the BRF/TBP/DNA and BRF/ B”/TBP/DNA complexes

BRF, TBP, and DNA were diluted and combined in the ratio of l:l:l.5 in a buffer
containing 50 mM Tris, pH 8.0, 10 % Glycerol, 300 mM KC], 1 mM DTT, then left on
ice for approximately 20 minutes to get complete binding. Very often, aggregation of the
formed complexes was observed even with a concentration as low as 0.08 mg/ml. In
these cases, a buffer containing higher concentration of salt was added to the mixture.
Any precipitates were removed by centrifugation before crystallization setups. So far 26
different BRF/TBP/DNA complexes have been made and used in crystallization as listed
in Table 2.6, of which 11 comlexes crystallized. Table 2.7 lists the conditions where the
BRF/TBP/DNA complex crystals were grown. In order to confirm the presence of
individual members of BRF/TBP/DNA complexes in the grown crystals, SDS-PAGE was
performed on dissolved crystals. The results shown in Figure 2.15 verified the presence
of BRF and TBP. Since BRF was known to interact with TBP poorly in the absence of
DNA, although the presence of DNA was not tested we are sure that the DNAs are

present in those complexes. A few examples of BRF/TBP/DNA complex crystals are

75

shown in Figure 2.16. Currently, the attempts to crystallize the BRF/B”/TBP/DNA

complexes are being made.

1234

 

Figure 2.15 SDS-PAGE of the BRF/TBP/DNA complexes. (1) S. cerevisiae BRF 435-
53l/TBP/DNA1 (2) S. cerevisiae BRF 435-53l/TBP/DNA2. (3) S. cerevisiae BRF 435-

53 l/TBP/DNA3. (4) S. cerevisiae BRF 435-531/TBP/DNA6.

76

Table 2.6 The BRF/TBP/DNA complexes tested for crystallization so far

 

 

 

 

S. cerevisiae BRF 435-531

 

 

Complexes BRF DNA* Crystals
1 S. cerevisiae BRF 311-596 DNAl Yes
2 S. cerevisiae BRF 435-596(A530-558) DNA] No
3 S. cerevisiae BRF 435-596(A530-558) DNA2 No
4 s. cerevisiae BRF 435-596(A530-558) DNA3 Yes
5 s. cerevisiae BRF 43 5-596(A530-558) DNA6 No
6 s. cerevisiae BRF 407-531 DNAI No
7 s. cerevisiae BRF 407-531 DNA2 No
8 S. cerevisiae BRF 407-531 DNA3 N0
9 s. cerevisiae BRF 407-531 DNA4 No
10 S. cerevisiae BRF 407-531 DNAS N0
11 s. cerevisiae BRF 407-531 DNA6 No
12 s. cerevisiae BRF 407-531 DNA7 No
13 S. cerevisiae BRF 407-531 DNA8 N0
14 S. cerevisiae BRF 435-531 DNAl Yes
15 S. cerevisiae BRF 435-531 DNA2 Yes
16 S. cerevisiae BRF 435-531 DNA3 Yes
17 S. cerevisiae BRF 435-531 DNA4 N0
13 s. cerevisiae BRF 435-531 DNAS No
19 S. cerevisiae BRF 435-531 DNA6 Yes
20 DNA7 Yes

 

77

 

Table 2.6 (cont’d)

 

 

 

 

 

 

BRF DNA Crystals
21 S. cerevisiae BRF 420-531 DNA9 No
22 K. lactis BRF 395-501 DNA] Yes
23 K. lactis BRF 395-501 DNA2 Yes
24 K. lactis BRF 395-501 DNAS Yes
25 K. lactis BRF 395-501 DNA6 Yes
26 K. lactis BRF 395-501 DNA7 No

 

* DNA numberings are consistent with Table 2.5.

78

 

Table 2.7 Crystals of the BRF/TBP/DNA complexes grown so far.

 

Complexes

Size (at the
longest

direction)

Conditions

 

I S. cerevisiae BRF 31 l-596/TBP/DNA1

0.05 mm

30 % PEG 8000,
0.2 M NaAc, 0.1 M
Na Cacodylate , pH
6.5

 

2 S. cerevisiae BRF 435-596 (A530-
558)/TBP/DNA4

0.05 mm

30 % PEG 4000,
0.2 M MgCl2, 0.1
M Tris, pH 7.5

 

0.2 mm

2.0 M NH4H2PO4
2.0 M, 0.1 M Tris,
pH 8.5

 

3 S. cerevisiae BRF 435-531/TBP/DNA1

0.1mm

30 % PEG400, 0.2
M CaCl2, 0.1 M
Na-HEPES, pH 7.5

 

0.08 mm

28 % PEG 4000,
0.2 M MgCl2, 0.1
M Tris, pH 7.5

 

4 S. cerevisiae BRF 435-531/TBP/DNA2

0.15m

30 % PEG 4000,
0.2 M MgCl2, 0.1
M Tris, pH 7.5

 

5 S. cerevisiae BRF 435-53l/TBP/DNA3

0.08 mm

30 % PEG 4000,
0.2 M MgCl2, 0.1
M Tris, pH 7.5

 

6 S. cerevisiae BRF 435-53 l/TBP/DNA6

0.05 mm

35 % PEG 4000,
0.2M MgCl2, 0.1 M
Tris, pH 7.5

 

7 S. cerevisiae BRF 435-531/TBP/DNA7

0.05 mm

30 % PEG 4000,
0.2M MgCl2, 0.1 M
Tris, pH 7.5

 

0.05 mm

18 % PEG 8000,
0.2 M
Ca(CH3COO)2,

0.1 M Na-
Cacodylate, pH 6.5

 

8 K lactis BRF 395-501/TBP/DNA1

 

 

 

0.2 mm

 

35 % PEG400, 0.2
M CaCl2, 0.1 M
Na-HEPES, pH 7.5

 

79

 

Table 2.7 (cont’d)

 

 

 

 

 

 

 

 

Complexes Size (at the Conditions
longest
directiory
9 K. lactis BRF 395-501/TBP/DNA2 0.1 mm 30 % PEG 4000,
0.2 M MgCl2, 0.1M
Tris, pH 9.5
10 K. lactis BRF 395-501/TBP/DNA5 0.05 mm 28 % PEG 4000,
0.2 M MgCl2, 0.1M
Tris, pH 7.5
11 K. lactis BRF 395-501/TBP/DNA6 0.05 mm 30 % PEG 400, 0.2

M CaCl2, 0.1 M
Tris, pH 7.5

 

80

 

 

Figure 2.16 (A) Crystals of the S. cerevisiae BRF 435-531/TBP/DNA1 complex grown
in 30 % PEG400, 0.2 M CaCl2, 0.] M Na—HEPES, pH 7.5. (B) Crystals of the S.
cerevisiae BRF 435-531/TBP/DNA6 complex grown in 30 % PEG 4000, 0.2 M MgCl2,
0.1 M Tris, pH 7.5. (C) Crystals of the S. cerevisiae BRF 435-531/TBP/DNA2 complex
grown in 30 % PEG 4000, 0.2 M MgCl2, 0.1 M Tris, pH 7.5. (D) Crystals of the

S. cerevisiae BRF 435-531/TBP/DNA7 complex grown in 30 % PEG 4000, 0.2M MgCl2,

0.1 M Tris, pH 7.5.

81

2.5 Discussion

Regardless of all the efforts that we have made on this project and the above-
mentioned preliminary data, regrettably we had to give up our pursuit of this project,
since the structure of the BRF/TBP/DNA complex has recently been determined by
another research group (18).

Our final thoughts go to the homology between the C-terrninal domain and
TFIIA. As shown in the sequence alignment between BRF and the TFIIA small subunit
(Figure 2.17), few but significant residues are conserved among two transcription
initiation factors. Especially, the key region of TFIIA that interacts with TBP, residues
67-72, shares a good homology with BRF 428—438. The unpublished structure (18) of
BRF/TBP/DNA contains residues 43 5-507, with the last 4 residues in this homology
region at the N-terminal. It is interesting that in this structure, the N-terrninus including
these 4 residues begins with a helix, while in the TFIIA structure this region is a strand.
However, It can still be assumed that interactions between this region of BRF and
TBP/DNA would be similar to those of TFIIA. It was also noted that the C-terrnini of
two proteins, BRF 564-585 within the homology region 111, and TFIIA 99-120, are

remarkably well conserved (Figure 2.17).

82

311 UTSEEALNKNPILTOVLGEUELSSIEVLFYLKQFSERRARWERIKIITIIG S. cerevisiae BRF
1 1111va w 5. camwsi'aeTFllA ssu

361 IDGEIIIIYHE GSENETRKRKLSEVSI QNEHVE GE DKET‘EGTEEKVKKVKTK S. CRIBW'NIQB BRF
ELYRRSTIGNSLVDALDTL ISDGRIEASLAHRVLETFDKWAETLK S. CHOWDER TFIIA $90

411 TSEEKKENESGHFQDA DGYS LETDPYC 1' RNLHLLPTTDTYL SKVSDD PD 5. cerew'a‘ae BRF
54 DNTQSKLTVKGNL DTYG FC 5. CBfEW'Sl'SBTFIIA ssu

461 NLEDVDDEELNAHLLNEEASKLKERIUI GLNAD FL LE DESKRLKUEAD Ill; 5. CQIPW'R'BR BRF
73

DDWT FIV K NCUVE S. cemw'a'aeTFllA ssu
511 TGNTSVKKKRTRRRNHTRSDEPTKTVDMAAIGLHSDLODKSGLHAALKA S. cemwsfee BRF
89 D5 ASNU S. cerevisiaeTFllA ssu
561 AEEsc-D F'I'I‘ADSV KNHLOKASFSKKINYDAIDGLFR s. cerevisiae BRF
99 SGDSUSVISVDKLRIV ACNSKKSE 5; 09,312,992er ssu

Figure 2.17 Sequence alignment of S. cerevisiae BRF C-terrninal domain and TFIIA
small subunit. Identical residues are highlighted in yellow. Boxed in blue are BRF428-

438 and TFIIA 67-72; this region of TFIIA makes interactions with the TBP/DNA.

TFIIA interacts with TBP near the N-terminal stirrup and interacts with the DNA
upstream of the TATA box (10,19). TFIIB interacts both with the C-terrninal stirrup and
DNA on one face of TBP/DNA complex (5,7). It was suggested from the Ge] Mobility
Shift Assay that BRF does not interact on the same face of the TBP/DNA complex as
TFIIB, however, BRF likely overlaps in its position with TFIIA for binding to TBP,
DNA, or both (2). The folding of some domains of BRF can be roughly predicted, based
on the structures of the TFIIA/TBP/DNA complex and the TFIIB/TBP/DNA complex
that have already been determined. Figure 2.18 shows the modeled
TFIIA/TFIIB/IBP/DNA quaternary complex, this modeling was done by overlaying the

structure of TBP in the human TFIIBc/TBPc/MLP complex (7) onto the structure of TBP

83

in the yeast TFIIA/TBP/DNA complex (20). The first B strand of the TFIIA small
subunit, Bl , contains the important residues that interact with TBP and shares a good
homology with BRF. Therefore, this region of BRF, BRF 428-43 8, is likely to make
similar interactions with TBP. The primary sequence of the C-tenninal B strand of the

T FIIA small subunit, B3, is quite similar to that of BRF, therefore it can be assumed that
the C-terminus of BRF makes similar interactions to that of TFIIA. It is very interesting
that although the N-terminal half of BRF shares a high homology with TFIIB, it does not
compete with TFIIB for TBP/DNA binding (2). Yet, given the high sequence homology
with TFIIB, the three-dimensional folding of the N-terrninal half of BRF is very likely to

be very similar to that of TFIIB (Figure 2.18).

84

 

Figure 2.18 Modeled structure of the TFIIA/TFIIB/TBP/DNA complex. TFIIA small
subunit is in yellow, TFIIA large subunit is blue, TBP is in cyan, TFIIB is in green, and
DNA is in silver. The presumed folding of BRF 1-286 would be somewhat similar to
TFIIB; the presumed protein-protein and protein-DNA contacts of BRF 428-438 would
be similar to those of B1 of TFIIA small subunit; the presumed folding of BRF 564-585

would mimic B3 of TFIIA small subunit.

85

2.6 Literature cited

1.

10.

11.

12.

l3.

l4.

Kassavetis, G. A., Bardeleben, C., Kumar, A., Ramirez, E., and Geiduschek, E. P.
(1997) Mol Cell Biol 17, 5299-5306

Colbert, T., Lee, S., Schirnmack, G., and Hahn, S. (1998) Mol Cell Biol 18, 1682-
1691

Kassavetis, G. A., Kumar, A., Ramirez, E., and Geiduschek, E. P. (1998) Mol Cell
Biol 18, 5587-5599

Hahn, S. (2000) Unpublished results

Nikolov, D. B., Chen, H., Halay, E. D., Usheva, A. A., Hisatake, K., Lee, D. K.,
Roeder, R. G., and Burley, S. K. (1995) Nature 377, 119-128

Kosa, P. F ., Ghosh, G., DeDecker, B. S., and Sigler, P. B. (1997) Proc Natl Acad
Sci U S A 94, 6042-6047

Tsai, F. T., and Sigler, P. B. (2000) Embo J 19, 25-36

Littlefield, O., Korkhin, Y., and Sigler, P. B. (1999) Proc Natl Acad Sci U S A 96,
13668-13673

Bagby, S., Kim, S., Maldonado, E., Tong, K. 1., Reinberg, D., and Ikura, M.
(1995) Cell 82, 857-867

Geiger, J. H., Hahn, S., Lee, S., and Sigler, P. B. (1996) Science 272, 830-836
Kim, J. L., Nikolov, D. B., and Burley, S. K. (1993) Nature 365, 520-527

Reddy, P., and Hahn, S. (1991) Cell 61, 349-357

Gill, G., and Tjian, R. (1991) Cell 65, 333-340

Cormack, C., Strubin, M., Ponticelli, A. S., and Struhl, K. (1991) Cell 65, 341-

348

86

15.

l6.

17.

18.

19.

20.

Persinger, J ., Sengupta, S. M., and Bartholomew, B. (1999) Mol Cell Biol 19,
5218-5234

Sheri, Y., Kassavetis, G. A., Bryant, G. 0., and Berk, A. J. (1998) Mol Cell Biol
18, 1692-1700

Shah, S. M., Kumar, A., Geiduschek, E. P., and Kassavetis, G. A. (1999) J Biol
Chem 274, 28736-28744

Personal communication with Juo, S.

Tan, 8., Hunziker, Y., Sargent, D. F., and Richmond, T. J. (1996) Nature 381,
127-151

Unpublished 2.5 A structure of yeast TFIIA/TBP/DNA complex.

87

CHAPTER III
X-RAY CRYSTALLOGRAPHIC STRUCTURAL STUDIES OF lL-MYO-INOSITOL
l-PHOSPHATE SYNTHASE
This chapter will describe the structural investigation Of lL-myo-inositol 1-
phosphate (MIP) synthase followed by a proposal Of the mechanism. In this chapter,

“MIP synthase” refers to MIP synthase from S. cerevisiae unless specified otherwise.

3.1 Theory
3.1.1 Structure determination from X-ray diffraction data

When a crystal is exposed to X-rays, constructive interference between rays
scattered from successive planes in the crystal will only take place if the path difference
between the rays is equivalent to an integral number of wavelengths. This is known as

Bragg's Law:

2dsin9 = mi. (3.1)

The structure factor for a particular reflection from a crystal is a complex quantity, which

can be represented by its amplitude and phase:

F(hkl) = ijezmhxj +kyj +121) 2 F(hkl)eia(hkl) (3.2)

Where F (th) is the amplitude and 0t(hkl) is the phase. When the diffracted X-ray is
recorded, all information on the phase is lost and only a measurement of the intensity of

the diffracted beam is recorded. The intensity is given by

1(hkl) = F(hkl) . F*(hkl) = [F(hkl)]2 (3.3)

88

The electron density can be computed by Fourier transform theory. The structure factor
equation 3.2 can be rewritten in terms of a continuous summation over the volume of the

unit cell.

F(hkl) = 2L62mflui +kyj +121) 2 21362711508 = p(r)e2m'rusdv

unit cell
(3.4)
where S is used to denote the position in diffraction Space. By multiplying both sides by

exp (-27rir’OS) and integrating over the volume of diffraction space, it can be shown that

p(r) = L F(S)e’2’°"sdvS

iffraction space (3 '5)

where dvS is a small unit of volume in diffraction space. The integration can be replaced
by a summation since F(S) is not continuous and is non-zero only at the reciprocal lattice

points. Therefore,

_ 1 -27ri(hx+k_v+lz)
p(xyz) — 17 2h: 21." Z F(hkl)e

_ 1 ia(hkl) -27ti(hx+ky+lz)
_V;;;F(hkl)e e (1 6)

Since only intensities but not phases are measured in the recorded diffraction pattern, it is
impossible to determine a structure directly from a recorded diffraction pattern; this is
known as the phase problem in Crystallography.

There are four methods by which the phase problem can be solved. 1) The
Patterson summation. This is a Fourier summation based on the experimentally
observable [F (hkl)]2 . It is basically a vector map of the structure and is applied for

molecules containing relatively few atoms. 2) Direct methods. In this method

89

mathematical relationships between the reflections can be used to provide phase
information. 3) Heavy atom isomorphous replacement. In this method a heavy atom is
introduced to a structure to provide phase information. 4) Anomalous scattering. In this
method phase information is obtained from the information contained in the scattering by
an atom whose natural absorption frequency is close to the wavelength of the incident

radiation.

3 .1 .2 Molecular replacement

The molecular replacement method makes use of a known protein as an initial
phasing model for the structure to be solved. The search and target molecules must have
reasonable sequence identity (> 25 %) for there to be a good chance Of success.
Likewise, having a dataset with good completeness can be crucial. Generally there are
two steps in molecular replacement and these are known as the rotation and translation
functions. If successful, a preliminary model of the target structure will be obtained by
correctly orienting and positioning the search molecule in the target cell. This solution
can be optimized by rigid body refinement. Finally, the target structure can be put
through cycles Of map calculation, model fitting, and refinement, which help to reduce
the bias introduced by the starting model. Even in the absence of a suitable search
molecule, the self-rotation function can be used to determine the direction and nature of
non-crystallographic symmetry elements. The problem is six dimensional because it
involves three rotational and three translational parameters. This is simplified by

separating the search into two stages namely the rotation and translation searches.

9O

Rotation function

The rotation function should allow the orientation of the search molecule, which
produces a maximal overlap with the target structure to be determined in the absence Of
any phases for the unknown structure. TO do this, it compares the Patterson self-vectors
of the known and unknown structures at different orientations of the search model. It
should be noted that Patterson functions can be calculated from the amplitudes only and
using Patterson space means that the translation vector is irrelevant since all intra-
molecular vectors are shifted to the origin. The rotation function is usually calculated as
a function of Eulerian angles: or, B, and y. The molecule is placed in an orthogonal co-
ordinate system with the axis Of highest symmetry along Z (about Ct) to reduce the
amount of computation.
Translation function

Having determined the angles or, B, and y from the rotation search, the rotation
matrix can be calculated and applied to the coordinates of the search molecule. The shift
vector, which is required to position the search molecule correctly relative to the
symmetry elements of the target molecule, can be determined by one of a number Of
translation searches. Patterson methods can be used to measure the overlap of the target
Patterson cross-vectors with those calculated for the oriented search molecule as it ranges
through the target cell. The simpler way to solve the translation problem is R-factor
search. It involves the calculation of an R-factor as the search molecule and its symmetry
mates are moved through the unit cell of the target crystal. The correct position should

give the lowest R-factor. Other parameters, such as the correlation coefficient, can be

91

used to measure the agreement between the Fohss and Fcais as the search model is moved

around. The R-factor is defined as follows.

R : XIIFOb‘II "' chall

ElFobsl (3'7)

The molecular replacement method was used tO determine the structure Of the

 

MIP synthase/NAD+ complex from the diffraction data on a P2. crystal that will be

discussed in this chapter.

3.] .3 Structure refinement

Once a structure is determined and the electron density map calculated, the model
needs to be refined in order to find a best agreement of the model with both observed
diffraction data and chemical restraints. The refinement in macromolecular
crystallography is complicated not only because of the high dimensionality of the
parameter space (typically at least three times the number of atoms in the model), but also
because Of the crystallographic phase problem described in section 3.1.1. Electron
density maps computed from a combination of native crystal amplitudes and multiple
isomorphous replacement (MIR) or multiwavelength anomalous diffraction (MAD)
phases sometimes are insufficiently accurate to allow a complete and unambiguous
tracing of the macromolecule. When structures are solved by molecular replacement
described in section 3.1.2, the resulting electron density maps can be severely “model-
biased” in that they seem to confirm the existence of the search model without providing
clear evidence of actual differences between it and the true crystal structure. Therefore,

initial models require extensive refinement.

92

A crystallographic refinement can be formulated as a search for the global

minimum of the target function E (1),

E=EC +w E

hem xray xray (3.8)
Em," is a function Of all atomic positions and an empirical force field, describing both
covalent and noncovalent interactions. Em, describes the difference between Observed
and calculated diffraction data, and W.,-2,}. is a weight chosen to balance the forces arising
from each term. In a macromolecular structure, the many local minima Of the target
function tend to defeat gradient descent optimization techniques such as conjugate
gradient or least-squares methods (2). These methods are not capable Of shifting the
atomic coordinates enough to correct errors in the initial model. Simulated annealing is
an Optimization technique well suited to the multiple-minima characteristic of

crystallographic refinement.

Simulated annealing refinement

Simulated annealing refinement can overcome barriers between local minima, and
thus can explore a greater volume of the parameter space to find “deeper” minima.
Annealing refers to a physical process wherein a solid is heated until all particles
randomly arrange themselves in a viscous liquid phase, and then cooled slowly so that all
particles arrange themselves in the lowest energy state. By formally defining the target E
in Eq. 3.9 to be the equivalent of the potential energy of the system, one can simulate the
annealing process (3). Simulated annealing is an approximation algorithm; there is no
guarantee that it will find the global minimum except in the asymptotic limit of an

infinite search. Compared to gradient descent methods where search directions must

93

follow the gradient, simulated annealing achieves more Optimal solutions by allowing
motion against the gradient. The likelihood of counter gradient motion is determined by
a control parameter “temperature”, the higher the temperature the more likely the
Optimization can overcome the barriers.

Simulated annealing requires a generation mechanism to create a Boltzmann

distribution at a given temperature T,

_E(q,, ...... (It)
ka

B(ql,......,qi)=e (3,9)

 

where E is the target function given by equation 3.8, k2, is the Boltzmann constant, and q I,
...... ,q, are adjustable parameters such as atomic coordinates. Simulated annealing also
requires an annealing schedule, which is a sequence of temperatures T, 2 T2 2 2 T, at
which the Boltzmann distribution is computed. Monte Carlo (4) and molecular dynamics
(5) simulations are the two most commonly used generation mechanisms.

Simulated annealing has improved the efficiency of crystallographic refinement
significantly; however, simulated annealing alone is still insufficient to refine a crystal
structure without human intervention by manual refitting of the model to electron density
maps using interactive graphics programs (6).

For all structures included in this chapter, structure refinements were done by

simulated annealing methods in combination with manual refitting of the models.

94

3.2 Experimental procedures
3.2.] Protein over-expression and purification
Over-expression

S. cerevisiae MIP synthase was over-expressed in E. coli BL21 (DE3) cells. The
LB medium containing 100 mg/L carbenicillin was inoculated with a single colony of
E. coli strain BL21 (DE3) bacteria containing MIP synthase expression plasmid. Cells
were grown at 37 °C with rapid shaking for 18 hours or until saturated. One liter of
LB/carbenicillin was inoculated with 25 ml of saturated culture and growth was
maintained at 37 °C with rapid shaking until the OD. 600 nm reached 0.6. The inducer,
IPTG, was then added to a final concentration of 0.25 mM, and growth continued for

another 3 hours. The cells were harvested by centrifugation at 5000 rpm.

Purification

0 Cells were re-suspended in the lysis buffer (Tris 20 mM, NH4C1 10 mM, DTT 10
mM, pH 7.7, protease inhibitor cocktail, complete EDTA-free) and lysed by
sonication. Cell debris was spun down by centrifugation at 15000 rpm for 40
minutes.

0 The supernatant was loaded onto a Phast-Q ion exchange column and washed
with the lysis buffer until the OD230 "m of the eluent is lower than 0.5. A gradient
Of NH4C1 (0.01-0.75 M) in the same buffer (20 mM Tris, pH 7.7, 10 mM DTT,
protease inhibitor cocktail, complete EDTA-free) was then applied onto the

column.

95

Eluted fractions containing MIP synthase were then pooled together and passed
directly through a Blue A affinity column to remove glucose 6-phosphate
dehydrogenase.

Activated charcoal was then added to the sample in a ratio of 1 g charcoal to 5mg
protein. This treatment was performed at 4 °C for 30 minutes. The cofactor-
bound charcoal was removed by centrifugation and filtration.

Protein samples were loaded onto another ion exchange column, an F PLC Source-
Q column, a gradient Of NH4C1 (0.01-0.3 M) in the same buffer (20 mM Tris, pH
7.7, 10 mM DTT, protease inhibitor cocktail, complete EDTA-free) was applied.
Fractions containing pure MIP synthase elutes at 0.18 M NH4C1. Fractions were
further analyzed by SDS-PAGE for homogeneity, and were concentrated to

8mg/ml for crystallization setups.

3.2.2 Crystallization

Crystallization Of apo MIP synthase

Crystals Of apo MIP synthase were grown by the hanging drop vapor diffusion

method (Figure 2.4). If good nucleations were not observed, micro or macro-seeding
methods were applied in order to get diffraction quality crystals. Crystals with the

dimensions 0.5mm x 0.5 mm x 0.5 mm grew in a condition previously reported (7).

Crystallization of MIP synthase in complex with cofactors and various ligands

Purified MIP synthase was incubated with cofactors and various concentrations of

different inhibitors at 37 °C for 10 minutes before co-crystallization setups.

96

In cases where co-crystallization setups of MIP synthase with various cofactors
and ligands did not produce suitable crystals for data collection, a soaking approach was
applied. Pre-grown apo MIP synthase crystals were soaked into stabilizers containing
various concentrations of cofactors and ligands. Table 3.] lists all crystals of MIP
synthase with cofactors and ligands obtained for data collection by both co-crystallization
and soaking methods. For each category, multiple data sets were collected on crystals
containing various concentrations Of cofactors and ligands at various pH’s, details will be

described in later sections.

97

Table 3.1 Crystals Of MIP synthase with cofactors and ligands

 

 

 

 

 

 

 

 

 

CO- Ligands CO- Soaked Space Resol
factor crystallized group ution‘k
1 None None N/A N/A C2 2.6 A
2 NAD+ None Yes P21 1.9 A
3 NAD+ None Yes C2 2.4 A
4 NADH None Yes C2 1.7 A
5 NAD+ 2-deoxy-D-glucitol 6-(E)- Yes C2 2.] A
vinylhomophosphonate
6 NAD+ 2-deoxy-D-glucitol 6-(E)- Yes C222. 2.3 A
vinylhomophosphonate
7 NAD+ 2-deoxy-D-glucitol 6-(E)- Yes C2 1.9 A
vinylhomophosphonate
8 NADH 2-deoxy-D-glucitol 6-(E)- Yes C2 2.2 A
vinylhomophosphonate
9 NADH 2-deoxy-D-glucitol 6-(E)- Yes C2 1.8 A
vinylhomophosphonate
10 None 2-deoxy-D-glucitol 6-(E)- Yes C2 2.0 A
vinylhomophosphonate
11 NAD+ 2-deoxy-D-glucose 6- Yes C2 2.0 A
phosphate
12 NAD+ 2-deoxy-D-glucose 6- Yes C2 2.1 A
phosphate

 

98

 

Table 3.1 (cont’d)

 

 

 

 

 

phosphate

 

 

 

 

CO- Ligands CO- Soaked Space Resolu
factor crystallized group tiont

l3 NADH 2—deoxy-D-glucose 6- Yes C2 2.0 A
phosphate

14 NADH 2-deoxy-D-glucose 6- Yes C2 1.9 A
phosphate

15 NAD+ D-glucitol 6- Yes C2 2.0 A
phosphate

16 NADH D-glucitol 6- Yes C2 2.] A
phosphate

17 NAD+ 2-deoxy-D-glucitol 6- Yes C2 1.9 A
phosphate

18 NADH 2-deoxy-D-glucitol 6- Yes Yes C2 2.0 A
phosphate

19 NAD+ Valproate Yes C2 2.0 A

20 NAD+ Valproate Yes R32 2.4 A

21 NAD+ Valproate P2 3.9 A

22 NAD+ Valproate Yes Yes C2 2.0 A

23 NADH Valproate C2 2.6 A

24 NADH Valproate Yes Yes C2 2.4 A

25 NAD+ DL-myo-inositol 1- C2 2.0 A

 

99

 

Table 3.1 (cont’d)

 

 

 

 

 

 

 

 

 

CO- Ligands CO- Soaked Space Resolu
factor crystallized group tion.
26 NADH D-glucose 6- Yes C2 2.0 A
phosphate
27 NAD* NH.+ Yes C2 3.5 A
28 NAD” Rb+ Yes C2 1.8 A
29 NADH EDTA Yes C2 2.] A

 

* For each category, multiple datasets were collected, listed is the highest resolution data

set among each category.

100

 

3.3 Data collection and refinement

For simplicity and clarity, only the datasets that produced the structures tO be
discussed in this chapter are described below.
3.3.] The apo MIP synthase

Crystals were transferred to a cryoprotecting stabilizer (5 % PEG 8000, 0.1 M
CH3COONa, pH 4.5, 30 % glycerol) before flash-frozen for data collection; this method
ofcryoprotection was applied to all the data collection to be described in this chapter.
Data were collected at Michigan State University Macromolecular X-ray diffraction
facility. Data reduction and scaling were performed using Denzo and SCALEPACK
respectively (8). The structure of MIP synthase/NADVZ-deoxy-D-glucitol 6-
phosphate(9) was used as the phasing model, the map was traced and refinement was
done using TURBO-FRODO (6,10) and CNS (11,12) respectively. The data collection
and refinement statistics are listed in Table 3.2. The final refinement model consisted of
residues 10-3 50, 376-533 in both molecules in the asymmetric unit. The structure also
included 133 water molecules. Figure 3.1 represents the Ramachandran plot of the apo

MIP synthase structure; only 3 residues out Of 997 are in disallowed regions.

10]

Table 3.2 Data collection and refinement statistics of the apo MIP synthasea

 

Space group

Resolution range (A)

Unit cell

Number of total reflections
Completeness (%)

Rsm (%)b

[/0

R factor (%)°

R... (%)“

Average B-factor (A2)
RMSD of bond lengths (A)

RMSD of bond angles (°)

 

 

C2

40-2.6 (2.69-2.6)
a=153.543 A, b=97.062 A, c=122.064 A, B=125.716°
44160

98.3 (96.8)

12.0 (60.5)

11.6 (1.9)

19.0

27.8

47

0.0072

1.39

 

a The parentheses denote those values for the last resolution shell.

I — I
b Rn... = II I K 1where I is the observed intensity and <I> is the average intensity

2 III

obtained from multiple observations of symmetry-related reflections.

 

[Fowl - chaiI
2 IF 0175'

d R. _ 2 lFobrI-IFcaII
[W — ElFobsl

 

.R=2

 

 

data.

, where reflections belong to a test set of 10 % randomly selected

102

 

 

 

 

 

 

Figure 3.1 Ramachandran plot of the apo MIP synthase structure.

103

3.3.2 The MIP synthase/NAD+ complex

Co-crystallization trials of MIP synthase and NAD+ have not produced diffraction
quality crystals. Both the P21 form and the C2 forms of apo MIP synthase crystals were
soaked into NAD+-containing stabilizer (1 mM NAD+, 5 % PEG 8000, 0.1 M
CH3COONa, pH 4.5) for 12 hours. Crystals of MIP synthase with fully occupied NAD+
were obtained. Data were collected on both the P21 form and the C2 forms of MIP
synthase/NAD+ complex crystals using synchrotron radiation at the Advanced Photon
Source, Argonne National Laboratory, BIOCARS BM-l4 and IMCA-CAT ID-17
beamlines respectively. Diffraction data reduction and scaling for the P21 MIP
synthase/NAD+ were performed using DENZO and SCALEPACK respectively (8). For
the C2 form of MIP synthase/NAD+ complex crystal, data were processed using
HKL2000. The structure of the P2; form of MIP synthase/NAD+ complex crystal was
solved by molecular replacement method using AMORe (l 3)(CCP4 package). A
monomer from the structure of the MIP synthase/NADVZ-deoxy-D-glucitol 6-phosphate
complex (9) was used as the model. The structure of the C2 form of MIP synthase/NAD+
crystal was solved using the same structure (9) as the initial phasing model. Maps were
traced using TURBO-FRODO (6,10) and refinements were performed using CNS(11).
Data collection and refinement statistics are tabulated in Table 3.3.

The contents of each molecule in the asymmetric unit of both crystal forms are
tabulated in Table 3.4. Figure 3.2 represents the Ramachandran plot Of the P21 MIP
synthase/NAD+ structure, only 6 out of 2001 residues are in the disallowed region; Figure
3.3 represents the Ramachandran plot of the C2 MIP synthase/NADJr structure, only 3 out

Of 1005 residues are in disallowed regions.

104

Table 3.3 Data collection and refinement statistics of the MIP synthase/NAD+ complexesa

 

Space group P21 C2

Resolution range (A) 40-1.9 (1.97-1.9) 50-2.4 (2.49-2.4)

Unit cell a=90.794 A, b=185.581 A, a=152.384 A, b=97.344 A,
c=94 A, l3=114.77° e=122.887 A, B=126.526°

Number of total reflections 214444 56444

Completeness (%) 99.7 (100) 96.6 (99.2)

R,ym (%)b 6.4 (42.2) 9.9 (45.3)

1/0 23.5 (2.5) 28.6 (3.0)

R factor (%)c 20.9 20.9

Rf... (%)d 27.8 28.8

Average B-factor (A2) 49.6 58.8

RMSD of bond lengths (A) 0.007 0.0078

RMSD of bond angles (°) 1.26 1.36

 

 

 

 

 

a The parentheses denote those values for the last resolution shell.

IIII- K111

b
R svm =
III

where I is the observed intensity and <I> is the average intensity

Obtained from multiple Observations of symmetry-related reflections.

C R = "F‘Uhfl — IFIYI“
IFohsI

2| F I , where reflections belong to a test set of 10 % randomly selected
obs

 

 

105

Table 3.4 Contents of the MIP synthase/NAD+ complex structures

 

 

 

Space Molecule Residues Number of water
group molecules
P21 A 10-361, 380-533 1529
B 10-361, 380-464, 472-533
C 10—464, 472-533
D 10-361, 380-390, 410-533
C2 A 10-362, 376-533 160
B 10-362, 376-464, 472-533

 

 

 

 

 

 

106

 

 

 

 

 

-180 0 180
Phi (degrees)

Figure 3.2 Ramachandran plot of the P21crystal form of the MIP synthase/NAD+

complex structure.

107

 

Psi (degrees)

 

 

 

 

-180 0
Phi (degrees)

Figure 3.3 Ramachandran plot of the C2 crystal form of the MIP synthase/NAB+

complex structure.

108

 

 

180

3.3.3 The MIP synthase/NADH complex

The crystals of the MIP synthase/NADH complex were Obtained identically to the
MIP synthase/NAD+ complex crystals, except for the change of NADH for NAD+ in the
stabilizer when the crystals were soaked. Data were collected on a MIP synthase/NADH
complex crystal using synchrotron radiation at the Advanced Photon Source, Argonne
National Laboratory, IMCA-CAT ID-l7 beamline. Diffraction data were integrated,
scaled, and reduced using HKL2000 (8). The structure was solved using the structure of
MIP synthase/NADV2-deoxy-D-glucitol 6-phosphate complex (9) as the initial phasing
model. The electron density map was traced using TURBO-FRODO (6,10) and
refinements were performed using CNS(11). Data collection and refinement statistics are
tabulated in Table 3.5. The final refinement model included residues from 10-533 in
molecule A, 10-464, 472-533 in molecule B, and 1068 water molecules.
Figure 3.4 represents the Ramachandran plot of the MIP synthase/NADH complex

structure, all residues present are in allowed regions evaluated by PROCHECK (14).

109

Table 3.5 Data collection and refinement statistics of the MIP synthase/NADH complex‘ll

 

Space group

Resolution range (A)

Unit cell

Number of total reflections
Completeness (%)

R.,... (%)b

1/0

R factor (%)C

R... (%)d

Average B-factor (A2)
RMSD of bond lengths (A)

RMSD of bond angles (°)

 

 

C2

5017 (1.8-1.7)
a=149.86 A, b=99.851 A, c=122.57 A, B=126.64°
160868

84.4 (69.5)

7.4 (25.6)

36.8 (4.4)

16.5

19

22

0.0053

1.35

 

“ The parentheses denote those values for the last resolution shell.

I — I
b Rm. = II I K xwhere I is the Observed intensity and <I> is the average intensity

2 III

obtained from multiple observations of symmetry-related reflections.

 

IFsiisl - IF call
2 lFobsl

d RH” = 2 lFobyl 4le
leobsl

 

.R=Z

 

 

data.

, where reflections belong to a test set of 10 % randomly selected

110

 

 

Psi (degrees)

 

 

 

 

 

1, fl

0 180
Phi (degrees)

Figure 3.4 Ramachandran plot of the MIP synthase/NADH complex structure.

111

3.3.4 The MIP synthase/NADH/EDTA complex

The crystals of the MIP synthase/NADH/EDTA complex were obtained as
described for the MIP synthase/NADH complex crystals except for the addition Of 5 mM
EDTA in the soaking stabilizer. Data were collected using synchrotron radiation at the
Advanced Photon Source, Argonne National Laboratory, IMCA-CAT ID-17 beamline.
Diffraction data were integrated, scaled, and reduced using HKL2000 (8). The structure
was solved using the structure of the MIP synthase/NADH complex as the initial phasing
model. The electron density map was traced using TURBO-FRODO (6,10) and
refinements were performed using CNS (1]). Data collection and refinement statistics
are tabulated in Table 3.5. The final refinement model included residues from 10-533 in
molecule A, 10-371, 376-464, 472-533 in molecule B, and 180 water molecules.
Figure 3.5 represents the Ramachandran plot of the MIP synthase/NADH/EDTA

structure, only 3 out of 1037 residues present are in disallowed regions evaluated by

PROCHECK (14).

112

Table 3.6 Data collection and refinement statistics Of the MIP synthase/NADH/EDTA

complexa

 

Space group

Resolution range (A)

Unit cell

Number of total reflections
Completeness (%)

R... (%)b

[/0

R factor (%)c

RI... (%)d

Average B-factor (A2)
RMSD of bond lengths (A)

RMSD of bond angles (°)

 

 

C2

502.1 (2.18-2.1)
a=151.98 A, b=97.61 A, c=121.72 A, B=126.15°
161339
99.8(99.6)

6.7 (42.2)

31.8 (3.2)

20.9

26.1

43.4

0.0072

1.30

 

a The parentheses denote those values for the last resolution shell.

I — I
b Ri-i-m = II I K ‘where I is the Observed intensity and <I> is the average intensity

XIII

obtained from multiple Observations of symmetry-related reflections.

 

 

 

 

 

C 2 lFobsl - IFCIIII
R =
2 IFobsI
d llyobsi — chatI
Rfrt't' =
anhsl
data.

, where reflections belong to a test set Of 10 % randomly selected

113

 

 

 

 

 

 

-180

5 Ramachandran plot of the MIP synthase/NADH/EDTA complex structure.

Figure 3

114

3.3.5 The MIP synthase/NAD+(NADH)/2-deoxy-D-glucitol 6-(E)-vinylhomophosphonate
complex

Various conditions were screened for co-crystallization attempts. CO-
crystallization yielded two crystal forms, C2 and C2221, both forms grew from a
condition similar to that for the apo MIP synthase crystals. When the concentration Of
the inhibitor was higher than 1.5 mM, however, co-crystallization yielded no crystals. At
the same time, soaking experiments were done with various concentrations of inhibitors
and pH’s. For soaking experiments, stabilizers with pH’s higher than 5.5 melted crystals.
All attempts are tabulated in Table 3.7.

After more than 30 data sets were collected on the MIP synthase
/NAD+(NADH)/2-deoxy-D-glucitol 6-(E)-vinylhomophosphonate complex crystals, the
data that produced a fully occupied inhibitor structure were collected using synchrotron
radiation source at the Advanced Photon Source, Argonne National Laboratory, IMCA-
CAT ID-17 beam line. Diffraction data were integrated, scaled, and reduced using
HKL2000 (8). The structure was solved using the structure of the MIP synthase/NADH
complex as the initial phasing model. The electron density map was traced using
TURBO-FRODO (6,10) and refinements were performed using CNS(11). Data
collection and refinement statistics are tabulated in Table 3.8. The final refinement
model included residues from 10-533 in molecule A, 10-464, 472-533 in molecule B, and
400 water molecules. Figure 3.6 represents the Ramachandran plot of the MIP
synthase/NADH structure/2-deoxy-D-glucitol 6-(E)-vinylhomophosphonate structure,
only 3 out of 1041 residues present are in disallowed region evaluated by PROCHECK

(14).

115

Table 3.7 Data collected on the MIP synthase/NADYNADH)/2-deoxy-D-glucitol 6-(E)-

vinylhomophosphonate complex crystals

 

 

CO- Concentration pH Method Space Resolu- Observation
factor Of the group tiona(A)
inhibitor
(mM)
NADI 0. l 7 4.5 Co-crystallized C2 3. 1 Not occupied
NADI 0.34 4.5 Co-crystallized C2 2.7 Not occupied
NAD+ 0.68 4.5 Co-crystallized C2 3 .2 Not occupied
NAD+ 1 4.5 Co-crystallized C2 2.] Not occupied
NADH l 4.5 Co-crystallized C2 3.0 Not occupied
NAD+ 1.5 4.5 Co-crystallized C2221 2.3 Not occupied
NADI 2 4.5 Soaked for 12 hrs C2 2.6 Not occupied
NADH 2 4.5 Soaked for 24 hrs C2 1.8 Occupiedb
NAD+ 5 4.5 Soaked for 12 hrs C2 2.7 Not occupied
NADH 5 4.5 Soaked for 7 hrs C2 2.6 Not occupied
NADH 6 4.5 Soaked for 12 hrs C2 2.6 Not occupied
NAD+ 7.5 4.5 Soaked for 12 hrs C2 2.1 Not occupied
NADI 10 4.5 Soaked for 6 hrs C2 2.5 Not occupied
NAD+ 13.5 4.5 Soaked for 24 hrs C2 1.9 Partially
occupied
NADH 13.5 4.5 Soaked for 24 hrs C2 2.0 Partially
occupied
NAD+ 13.5 5.5 Soaked for 24 hrs C2 2.0 Fully
occupied

 

 

 

 

 

 

 

 

 

8The highest resolution data set among multiple data sets in each experiment.
bOccupied with two small molecules that were present in the NADH-bound structure but

not the inhibitor.

116

Table 3.8 Data collection and refinement statistics of the MIP synthase/NADI/Z-deoxy-

D-glucitol 6-(E)-vinylhomophosphonate complex“I

 

Space group C2

Resolution range (A) 50-2.0 (2.1-2.0)
Unit cell a=151.81 A, b=97.782 A, e=122.293 A, B=126.3°
Number of total reflections 110451
Completeness (%) 98.2(98.6)

R.ym (%)b 7.4 (50)

[/0 17.7 (2.46)

R factor (%)c 18.8

Rs... (%)d 24.4

Average B-factor (A2) 37.3

RMSD of bond lengths (A) 0.0067

RMSD of bond angles (°) 1.29

 

 

 

 

a The parentheses denote those values for the last resolution shell.

I — I
b Rsym = II I K 1where I is the observed intensity and <I> is the average intensity

XIII

Obtained from multiple observations of symmetry-related reflections.

CR 2 2 |F..i..l— 112.1
Emmi

lFobsI — charI
IFobsl

 

 

 

 

d Rim = , where reflections belong to a test set Of 10 % randomly selected

data.

117

 

Psi (degrees)

 

 

 

Figure 3.6 Ramachandran plot of the MIP synthase/NADVZ-deoxy-D-glucitol 6-(E)-

vinylhomophosphonate structure.

118

 

3.4 Structure of the apo MIP synthase

The previous observation that the active site residues 351-409 are missing in the
structure of MIP synthase with partially occupied NAD+ but became ordered in the
inhibitor-bound structure (9) begged the question of whether or not the active site folding
could possibly be due to fully occupied NAD+. In order to answer this question, we
decided to determine the structure of MIP synthase in the absence of NAD+.

The overall structure of the apo MIP synthase is similar to the previously
published low occupancy NAD+-bound MIP synthase structure (9). As expected, there
was no electron density for NAD+ in both molecules in the asymmetric unit, confirming
that the activated charcoal removed the NAD+ from the enzyme. Although a part of the
disordered domain consisting of residues 351—375 is still missing, residues 376-409 that
were missing in the low occupancy NAD+ structure are now well ordered (Figure 3.7).
Most of the residues that are present in this structure overlay very well with those of the
structure with low occupancy NAD+ (9), the RMSD between the two structures is
0.8648A. This observation indicates that while the active site of MIP synthase is very
flexible, the NAD+ binding domain is very rigid.

However, the loop encompassing residues 191-198 differs significantly between
the apo MIP synthase and the previously published low occupancy NAD+-bound
structure (Figure 3.8). The location of this loop in the apo structure overlays well with
the structure of the MIP synthase/NADVZ-deoxy-D-glucitol 6-phosphate complex rather
than the low occupancy NAD+-bound structure, inconsistent with the previous conclusion

that the folding of the active site forces this loop to flip out (9).

119

 

Figure 3.7 Structure of the apo MIP synthase. The region that was disordered in the

previous low occupancy NAD+- bound structure (9) is colored red.

120

 

Figure 3.8 Superposition of the structures of apo MIP synthase in gold and the low

occupancy NAD+-bound MIP synthase (9) in blue. Boxed in red are residues 191-198.

121

3.5 Structure of the MIP synthase/NAD+ complex

In order to investigate any possible conformational changes in the MIP synthase
active site and the Rossmann fold domain in the presence of fully occupied NAD“, and
more specifically, in order to answer the question of whether or not the active site folding
that was observed in the structure of the MIP synthase/NADVZ-deoxy-D-glucitol 6-
phosphate complex could be due to the fully occupied NAD+, the three-dimensional
structure of the MIP synthase/NAD+ complex was elucidated. This structure represents
the enzyme at its state when the substrate is not yet bound or the product is already

produced and released from the enzyme active site.

3.5.1 Structure of the MIP syntliase/NAD+ complex tetramer

The P21 form contains an entire tetramer of MIP synthase in the asymmetric unit
as it is in solution (Figure 3.9). Molecule A and B, C and D are related by a
noncrystallograhic two-fold rotation axis. The AB and CD dimers are related by another
noncrystallographic two-fold axis that is perpendicular to the first two-fold axis. The
interface between two monomers buries a surface area of 9539.66 A2, while the
tetramerization interface between the AB and CD dimers buries an additional 6096.4 A2

of surface area.

122

 

Figure 3.9 The entire tetramer of the P21 structure of the MIP synthase/NAD+ complex.

Molecule A is colored in cyan, B is in magenta, C is in gold, and D is in green.

123

3.5.2 Structure of the MIP synthase/NAD+ complex monomer

In the P21 crystal structure of the MIP synthase/NAD+ complex, molecules A, B,
and D are similar in that they are all missing residues 362-380 as shown in Figure 3.10.
Molecule C in the P21 structure is an exception in that the entire active site is folded. In
the C2 structure, both of the molecules in the asymmetric unit are missing residues 363-
375. In addition, molecules B and C in the P21 structure and molecule B in the C2
structure are missing residues 465-471. The ordered part of the structure in each
monomer is overall similar to that of apo MIP synthase, the RMSD between the two
structures is 1.07 A. On the other hand, 0:13, the helix encompassing residues 352-361,
which was missing in the apo structure became ordered upon NAD+ binding. Residues
N355 and D356 of this helix make direct interactions with NAD+. In all molecules
except molecule C in the P21 structure, the strand that was missing in the previously
published structure of the low occupancy NAD+-bound structure is less well ordered than

in the apo structure.

124

    

Rossmann fold

Figure 3.10 A monomer of the MIP synthase/NAD+ complex. The helix in red is the
newly ordered region upon NAD+ binding; NAD+ is in lavender. The rest of the a

helices are cyan; all B strands are green; the loops are brown.

125

3.5.3 NAD+ binding

In both the P21 and the C2 crystal forms of the MIP synthase/NAD+ complex
structures, the NAD+ molecules are fully occupied. The nicotinamide portion of the
molecule has higher B factor values indicating that this part is less well ordered overall
than the rest of NAD+.

MIP synthase binds NAD+ in a way similar to that of other Rossmann fold
containing NAD+ binding proteins. The NAD+ molecule runs across the bottom of the
Rossmann fold domain. The Rossmann fold domain of MIP synthase contains the core
topology region, B3alflgagfinamfilzag 18.3, the structural motif common in many NAD+
binding proteins. Numerous interactions were observed between NAD+ and MIP
synthase as delineated in Table 3.9. Overall 1054 A2 of surface area was buried by the
interaction of MIP synthase with NAD+. The consensus phosphate binding sequence
GXGGXXG motif is also present in MIP synthase as the loop connecting B3 and on, and
the N-terminus of on (residues 72-78, Figure 3.11). The interaction of this motif and
NAD+ is similar to that of other NAD+ binding proteins. The first conserved glycine,
G72, allows for a tight turn of the main chain from [33 to on, and this is important for
positioning the second and the third glycines, G73 and G74 respectively. Because of its
missing side chain, G74 enables close contact of the main chain to the pyrophosphate of
NAD+. The last glycine in this motif, G78, is important for the close packing of B3 and
on. The main chain nitrogen atoms as well as the side chain nitrogen atoms of N76 and
N77 within this motif make hydrogen bond interactions with the pyrophosphate oxygen

atoms as shown in Figure 3.11. It is important to note that since this GXGGXXG motif

126

is not conserved in A. fulgiudus MIP synthase, NAD+-binding could be significantly
different from that of S. cerevisiae MIP synthase.

Unique in MIP synthase are three insertions. The first insertion encompassing
residues 93-140 contributes to the dimerization via van der Waals interactions and several
tight hydrogen bonds (Table 3.10). The second insertion encompasses residues 149-215,
this insertion completely surrounds the adenine portion of NAD+ making strong
hydrophobic interactions as well as hydrogen bond interactions with NAD+ (Figure 3.12).
These interactions seem to contribute strongly to the high binding affinity of NAD+ to
MIP synthase (Km=0.017 mM). The third insertion encompassing residues 247-276 is
involved in crystal packing as shown in Figure 3.13. This region interacts with the loop
of a neighboring tetramer encompassing residues 464-472.

With the active site residues 362-3 79 being disordered, the nicotinamide portion
of NAD+ is completely exposed to the surface. In fact, there is a very tight hydrogen
bond with an interatomic distance of 2.5 A that stabilizes this conformation as shown in

Figure 3.14.

127

Table 3.9 Interactions between the MIP synthase and NADW

 

 

 

 

 

 

 

 

 

 

Portion van der Hydrogen bonds Portion van der Hydrogen bonds
of Waals (Distances in A) of Waals (Distances in A)
NADA contacts NADJr contacts
171 Phospho G75
G72 diCSter N76 A022N76 N(3.22)
W147 N77 N022N77 N(2.95)
D148 N246
1149 N355 AOlzN355ND2(2.7)
Adenine Sl84 AN1:8184 OG (2.71) D356
1185 AN6:II85 O (2.96)
W243 N77
A245 T244
N246 A245 N03*:A245 O(2.9)
P277 N246
Nicotina T247 N02*:T247OG1(3.2)
mide NO3*:T247 N(3.01)
Ribose G295
8296
D356
G72 N77 NN7:N77 OD.(3.16)
G74 G295
G75 D320
N76 L321
Adenine D148 AO3*:D148 oo.(2.82) Nicotina D438
A02*:D148 OD2(2.88) m‘de A442
Ribose 1149
R198 A02*:R198 NE(3.21)
T244
A245
N246

 

# Listed are residues making van der Waals interactions within a 4.0 A cutoff.

128

 

\
I71 H ."
‘ |‘\‘\ I” \‘ :
G75 \‘v" ‘ '.
G72 ‘ , .
N76 '
‘ G74 N77
L73 G78

Figure 3.11 The GXGGXXG phosphate-binding motif within the MIP synthase

Rossmann fold domain.

129

Table 3.10 Dimerization interactions‘ by the first insertion (93-140)

 

 

 

A B Hydrogen bonds A B Hydrogen bonds
(Distances in A) (Distances in A)
F94 L424 C112 L387
E98 H427 S113 D338
K101 G425 SI 13 N384
N104 M423 SI 13 1386
N104 L424 T114 S383 T114 0G1: S383 0
F106 G340 T114 N384 (2.45)
F106 K342 T114 1386
F106 L387 L117 V37
F106 L392 L117 F45
F106 E421 G118 A35
F106 M423 G118 V37
GIO7 A339 G107 N:G340 O (2.76) 1119 N34
GIO7 G340 1119 A35
G107 1341 1119 V36 Ill9N:A350(3.03)
8108 A339 1119 V37
M109 D338 D120 V37 1119 O : V37 N (2.77)
Q1 11 I386 E122 G496
C1 12 G340 N124 H498
C112 N384 V126 V37
C112 1386 V126 F45
Y127 S383

 

 

 

 

 

 

130

van der Waals interactions are within a 4.0 A cutoff.

 

 

Figure 3.12 Interactions between the adenine portion of NAD+ and MIP synthase.

131

    
 

    

 

”’65:-

, J'
‘tN‘r‘lr $5.1 .C
.‘l I I “I ‘1
“gt, '4' 541.71%, 3
'..,,‘ 31's
\

   

  
 

Figure 3.13 The insertion encompassing residues 247-276 is involved in the interaction
with a neighboring tetramer.

(A) Two neighboring tetramers. Lavender, molecule A; green, molecule B; gold,
molecule C; cyan, molecule D.

(B) The insertion in molecule C makes interactions with the loop 464-472 of molecule A

in the neighboring tetramer.

132

   
     
   
  
 

‘1 .g .
I 1’ .II‘A.

, H". -.
‘ A? r:.". atk

 

   

 

Figure 3.14 (A) Space-filling model of the NAD+-bound MIP synthase, the nicotinamide
is exposed as shown in atom-color. (B) The tight hydrogen bond between the

nicotinamide nitrogen and the phosphodiester oxygen stabilizes this conformation.

133

3.5.4 Electrostatic charge distribution

An analysis of surface charge distribution (15) of the NAD+-binding site was
performed using the program GRASP (16). The electrostatic potential surface of the
NAD+-binding region is shown in Figure 3.15. The apparent negative charge on the
NAD+-binding surface within the dotted circle defines the binding site of the positively
charged nicotinamide, indicating that there are polar interactions between this site and the
nicotinamide portion of NAD“. Furthermore, this observation implies that as NAD+
oxidizes the substrate to become NADH, these polar interactions will diminish and a
motion of the nicotinamide ring might occur. The potential substrate-binding site is

denoted within the dotted square.

134

 

Figure 3.15 The electrostatic potential surface of MIP synthase with a view looking into
the NAD+ binding surface. Blue, EPS> 6kca1/mol; red, EPS<—6 kcal/mol; white, EPS~0.
The dotted circle denotes the nicotinamide-binding site; the dotted square denotes the

potential substrate-binding site. For clarity, residues 181-201, 351-361 were removed.

135

3.5.5 The exception of molecule C in the P21 structure of the MIP synthase/NAD+
complex

In molecule C in the P21 structure of the MIP synthase/NAD+ complex, the entire
active site was ordered, although the electron density for some regions was not complete.
A significant electron density feature was present in the center of the 2-deoxy-D-glucitol
6-phosphate binding site, indicating that a small molecule is bound in the enzyme active
site, consistent with the previous observation that the enzyme active site folds only when
occupied with a small molecule (9). Most of the active site residues including S323,
N3 54, K369, K412, D438, and K489, occupy similar positions to those seen in the
previously reported structure of the MIP synthase/NADVZ-deoxy-D-gluctiol 6-phosphate
complex. However, D356, which was flipped out of the enzyme active site in the
structure of the MIP synthase/NADVZ-deoxy-D-glucitol 6-phosphate complex (9), is now
swung into the phosphate position of the inhibitor 2-deoxy-D-glucitol 6-phosphate.
Q325, which was implicated in substrate binding when the substrate was modeled in a
pseudocyclic conformation consistent with aldol cyclization (9), has also moved and
makes a hydrogen bond with K412, buttressing its orientation in the enzyme active site.
These changes compared to the structure of the MIP synthase/NADVZ-deoxy-D-g1ucitol
6-phosphate complex (9) brought about questions regarding some details of the model

reported previously (9).

I36

3.6 Structure of the MIP synthase/NADH complex

All non-redox steps during the reaction catalyzed by MIP synthase occur when
the enzyme is coupled with NADH after NAD+ has oxidized the substrate, therefore, the
three-dimensional structure of the MIP synthase/NADH structure must represent the
enzyme in its active catalytic state. In addition, after the hydroxyl group of the substrate
C5 is oxidized by NAD+, either the nicotinamide ring or the substrate may move in order
to reposition the carbonyl oxygen of the substrate C5 since it would collide with the
hydrophobic nicotinamide ring if there were no motion. Consequently, some of the
active site residues may move as well to accommodate any repositioning of the substrate
and/or the nicotinamide ring. In order to answer these questions, it was decided to
elucidate the structure of the MIP synthase/NADH complex. As expected and even
beyond that, the structure of MIP synthase/NADH presented a picture rather different
from any structures of MIP synthase seen thus far. Several dramatic structural changes

were observed in the enzyme active site as well as in the cofactor.

3.6.1 Overall structure of the MIP synthase/NADH complex

When MIP synthase is bound with NADH instead of NAD“, the diffraction of the
crystals improved significantly to 1.6 A though the data is only complete to 1.7 A.
Figure 3.16A shows an example of the 1.6 A electron density map for the NADH-bound
MIP synthase structure.

The entire active site became ordered when the enzyme is bound to NADH.
Although quite surprising, this observation was consistent with the speculation that the

enzyme is at its active stage when coupled with NADH instead of NAD”, since all of the

137

non-redox steps catalyzed by the enzyme occur after NAD+ is reduced to NADH. There
were no significant conformational changes in the Rossmann fold domain of MIP
synthase compared with the apo structure. The two structures overlay well with an
RMSD of 0.87 A. Figure 3.16B depicts the superimposed structures of the apo and the

NADH-bound MIP synthase.

I38

 

Figure 3.16 (A) An example of the 2F o-Fc electron density map of the MIP
synthase/NADH structure contoured at 2.4 o. (B) Superposition of the structures of apo
MIP synthase and the NADH-bound MIP synthase. The apo MIP synthase is in gold and

the NADH-bound MIP synthase is in blue.

I39

3.6.2 Conformation of NADH

When compared with the structure of the MIP synthase/NAD+ complex, a
significant conformational change of the cofactor was observed. The nicotinamide ring
deviates from its position in the NAD+-bound structure by about 1.3 A, and the tight
hydrogen bond that pinches the nicotinamide and phosphodiester together was eliminated
(Figure 3.17). In fact, there is a 130 electron density peak between the nicotinamide and
phosphodiester as shown in Figure 3.18 A at position 0. This feature coordinates the
nicotinamide oxygen and phosphodiester oxygen with interatomic distances of 2.34 A
and 2.36 A respectively. A conserved water molecule 2.32 A away and another 10 a
feature 2.29 A away (Figure 3.18A at position 4) make the third and fourth coordination
sites of the tetrahedrally coordinated feature. These distances are short for hydrogen
bonds; instead this feature is more likely a metal ion. A number of different divalent and
monovalent cations were modeled in this position including Zn2+, Mn”, Ca2+, Mg”, Na+,
and K“. Modeling of Zn2+ and Mn2+ produced negative density around the atom in the
F.,-Fc map, indicating that they are too electron rich to account for this feature. On the
other hand, Mg2+ accounts for the electron density the best with only a bit of positive
density in the F.,-Fc map.

When the apo MIP synthase crystals were soaked in EDTA in addition to NADH,
all the active site residues are ordered as in the MIP synthase/NADH structure, with the
exception of residues 372-3 75 in molecule B. However, the conformation of the
nicotinamide is completely different from that of the NADH-bound structure, in fact it is
similar to that of the NAD+-bound structure. Importantly, the electron density feature

between the nicotinamide and phosphodiester has disappeared (Figure 3.18 B), indicating

I40

that this is indeed a divalent cation. However, it is still unclear which specific divalent
cation this is. The density of the nicotinamide is also much worse, indicating the

increased mobility of the nicotinamide ring in the absence of a putative divalent cation.

141

 

Figure 3.17 (A) Overlay of cofactors in the NAD+-bound structure in yellow and the
NADH-bound structure in cyan. (B) The 2Fo-Fc electron density map of the MIP
synthase/NADH complex structure contoured at 2.40 around the NADH. Shown in blue
is the putative divalent cation; red is a water molecule that coordinates with the divalent

cation.

142

    

      

      

firm???
ox-‘h-‘fit ~
..

    
 
 
 
 

.,_ -
av»? .

       
 
     
  

  

- \".

:l:
.w

'4'.-
135’

   

.\

Figure 3.18 (A) Simulated annealing omit electron density map of the MIP

synthase/NADH complex structure contoured at 56.

(B) Simulated annealing omit electron density map of the MIP synthase/NADH/EDTA

complex structure contoured at 1.8 o.

143

The conformation of nicotinamide and the position of the putative divalent cation
are similar to those of MIP synthase from Mycobacterium tuberculosis (Figure 3.19) (17).
In the M. tuberculosis MIP synthase structure, Zn2+ was modeled at this position,
however, its coordination is different from that of the putative divalent cation in the
structure of the S. cerevisiae MIP synthase/NADH complex. Distances from Zn2+ to the
nicotinamide oxygen and phosphodiester oxygen are 2.16 A, 2.08 A respectively, the
third ligand 2.21 A away is a conserved water molecule residing at the same position in
the S. cerevisiae MIP synthase/NADH structure. The fourth ligand is S311, and is
conserved as S439 in S. cerevisiae MIP synthase. In the structure of the MIP
synthase/NADH complex, S439 is 2 A away from this position making a tight hydrogen
bond with the fourth ligand of the putative divalent cation. S439 is in the same position
in all structures of S. cerevisiae MIP synthase determined so far.

The mechanistic role of the divalent cation is not clear at this point. Previous
experiments done on S. cerevisiae MIP synthase have ruled out the direct involvement of
a divalent cation during the catalysis (18). The structural role of this putative divalent
cation is critical in that it bridges the cofactor and thus might bring the nicotinamide C4
close to the substrate oxidation center C5. This divalent cation also coordinates an
important water molecule that makes hydrogen bonds with two absolutely conserved
aspartate residues, D356 and D438. It might influence the pK. of the acidic side chains
of D3 56 and D438 by chelating the conserved water molecule, which makes hydrogen
bonds with D3 56 and D438. Further biochemical experiments are required to clearly

identify this cation and define its precise role in catalysis.

144

D410 _ 2.9/9
A mgr. S439
2.34/1 .3.» 2.99A
I

2.29 A

      

 

Figure 3.19 (A) The NADH-bound MIP synthase from S. cerevisiae. Cyan, the modeled

putative divalent cation; Orange, the fourth ligand of the putative divalent; Red, the

conserved water molecule. (B) The NAD+—bound MIP synthase from M. tuberculosis.

'7 . .
Zn” is in cyan.

145

3.6.3 The MIP synthase active site

While all of the active site residues are ordered, significant structural changes
compared to the structures of the apo MIP synthase and the MIP synthase/NAD+ complex
are also evident on the opposite side of the active site. There are two small molecules
bound in the enzyme active site. The peak heights of both of these molecules in the F o-Fc
electron density map are higher than 70. One of these molecules is clearly a tetrahedron
with a heavy atom as the central atom consistent with either a phosphate or sulfate ion,
since the peak height at the center is 22 o in the F o—Fc electron density map. Given the
fact that there are often micro-molar quantities of phosphates in the reagents that were
used in protein purification and crystallization, a phosphate anion was modeled in this
position. The other feature overlaps with the position of the inhibitor in the previously
published structure of the MIP synthase/NADVZ-deoxy-D-glucitol 6-phosphate complex.
A glycerol molecule was modeled because of the shape and size of the electron density
and also the fact that a 30 % glycerol solution was used as the cryoprotectant when
fieezing these crystals. Figure 3.20 shows the 70 simulated annealing omit map with
these two molecules modeled in.

A significant conformational change also occurred in a loop connecting [313 and
(112, in order to accommodate the phosphate. The Con of S323 moved 2.1 A and the side
chain oxygen moved 4 A away from their positions in all structures of MIP synthase
determined previously including the apo, NAD+-bound, low occupancy NAD+-bound (9),
and 2-deoxy-D-glucitol 6-phosphate-bound (9). Figure 3.21 depicts the motion of this

loop.

146

,1
t

0'. ‘

‘v' 4:! fi‘e‘
q... :35;

. £15378?” 7

fiat” *‘ "’4‘ A 7

'l 2"- ’-‘ ""“~‘;‘-”és"!'

4w,

1—
(

 

Figure 3.20 70 simulated annealing omit electron map of the MIP synthase/NADH active

site with a phosphate and glycerol modeled in.

147

 

Figure 3.21 Conformational change of the phosphate-binding loop in the NADH-bound

structure in cyan compared with the NAD+-bound structure in silver. Note: The
conformation of this loop in the structure of the apo MIP synthase, the low occupancy
NAD+-bound MIP synthase(9), and the MIP synthase/NADVZ-deoxy-D—glucitol 6-

phosphate complex (9) are all identical to that of the NAD+-bound structure.

148

Numerous interactions were observed between the enzyme active site and
phosphate and glycerol (Figure 3.22). The phosphate oxygens make hydrogen bonds
with the main chain nitrogen atoms of the S323-G324-Q325-T326 sequence. This
sequence mimics the GXGGXXG NAD+ pyrophosphate-binding motif to some extent in
that the glycine residue G324 allows for a tight turn from 813 to a12 to form a stable
phosphate-binding pocket. This SGQT motif is absolutely conserved among eukaryotic
species from yeast S. cerevisiae to human (Figure 1.10). However, in M. tuberculosis,
there are two additional resides, Q201 and V202, inserted between the serine and glycine
making this connection loop extended (Figure 3.23). However, the conformation of the
side chains of 8200, G203-A204-T205 in this loop in M. tuberculosis MIP synthase are
similar to those seen in the structure of the S. cerevisiae MIP synthase/NADH complex.
The side chain nitrogen atom of Q325 makes a hydrogen bond with the phosphate while
the side chain oxygen atom of Q325 buttresses K412 in its position via a hydrogen bond.
The phosphate also makes salt bridge interactions with three absolutely conserved lysine
residues, K373, K412, K489. The putative glycerol molecule makes interactions with
N350, L352, D356, L360, K369, K373, I402, K412, and D438. It is positioned close to
the nicotinamide C-4 where the oxidation and reduction occur. It is important to point
out that the putative divalent ion chelates the water molecule that holds two conserved
aspartate residues, D3 56 and D43 8, together for hydrogen bonding with the glycerol
hydroxyl oxygens.

Clearly, binding of the phosphate and glycerol mimics the substrate binding to the
enzyme active site, and the ordered active site is due to the binding of these two

molecules. This is consistent with the idea that the substrate folds the enzyme active site

149

by encapsulating itself within the enzyme active site. On the other hand, the position of
the phosphate is completely different from that of 2-deoxy-D-glucitol 6-phosphate in the
previously reported structure (9). Figure 3.24 shows the discrepancy between the two
structures. The unambiguous position of the phosphate in the MIP synthase/NADH
complex structure calls into question the mechanism proposed previously based on the
structure of the MIP synthase/NADH-deoxy-D-glucitol 6-phosphate complex (9). In
addition, D356, which was flipped out of the enzyme active site in the structure of the
MIP synthase/NADVZ-deoxy-D-glucitol 6-phosphate complex (9), is now swung into the
position of the phosphate of the inhibitor 2-deoxy-D-glucitol 6-phosphate. The
conformation of D356 in the NADH-bound structure is the same as that seen in the

structures of the MIP synthase/NAD+ complexes where this residue is ordered.

150

Putative
divalent cation

D356

 

Figure 3.22 Interactions in the enzyme active site observed from the MIP
synthase/NADH complex structure. The phosphate and glycerol are in gold; the putative
divalent cation is in cyan. The fourth coordination ligand of the putative divalent cation,
which was modeled as a water molecule in the MIP synthase/NADH complex structure,

is in yellow.

151

 

Figure 3.23 Overlay of the S. cerevisiae MIP synthase/NADH complex structure in cyan
and the M. tuberculosis MIP synthase/NAD+ complex structure (17) in yellow. In the

dotted box are the phosphate-binding loops.

152

 

Figure 3.24 Overlay of the MIP synthase/NADH complex structure in silver and the MIP
synthase/NAD‘VZ-deoxy-D-glucitol 6-phosphate complex structure in lavender. Gold,
the phosphate and glycerol in the NADH-bound structure; blue, 2-deoxy-D-glucitol 6-
phosphate in the previous inhibitor-bound structure (9); yellow, side chains of the

NADH-bound structure; cyan, sides chains of the previous inhibitor-bound structure(9).

153

3.7 Structure of the MIP synthase/NADVZ-deoxy-D-glucitol 6-(E)-
vinylhomophosphonate complex

In order to elucidate completely the mechanism for the reaction catalyzed by MIP
synthase, the structures of MIP synthase in complex with various structural analogues of
the substrate as well as the reaction intermediates are essential. Several questions
emerged from the discrepancies between the structures of the MIP synthase/NADH
complex, the MIP synthase/NAD+ complex, and the previously reported structure of the
MIP synthase/NAD+/2-deoxy-D-glucitol 6-phosphate also brought about a need to
elucidate the structures of MIP synthase in complex with various inhibitors.

Among all of the inhibitors used in crystallization so far, 2-deoxy- D-glucitol 6-
(E)-vinylhomophosphonate was of particular interest, because it is the most potent
inhibitor synthesized so far with K.- = 0.67 M (19). The conformation of 2-deoxy- D-
glucitol 6-(E)-vinylhomophosphonate is constrained by the introduction of a double bond
between C6 and C7; this constraint allows the molecule to be a structural mimic of the
substrate D-glucose 6-phosphate in its transoid conformation.

Obtaining crystals of MIP synthase with fully occupied inhibitor 2-deoxy- D-
glucitol 6-(E)-vinylhomophosphonate was an unexpected challenge. Crystals were
obtainable from co-crystallization setups with the inhibitor concentrations lower than 1.5
mM. However, electron density maps showed no density for bound inhibitor molecules.
Soaking experiments with inhibitor concentrations as high as 10 mM also did not produce
electron density maps with a fully occupied inhibitor, although scattered pieces of
electron density were observed in the enzyme active site. This was surprising given the

high binding affinity of the inhibitor 2-deoxy- D-glucitol 6-(E)-vinylhomophosphonate

154

(K; = 0.67 1.1M). When the inhibitor concentration was increased to 13.5 mM and the pH
of the soaking stabilizer was 4.5, there was unambiguous electron density for phosphate,
and a clear electron density feature was present where the glycerol molecule was
modeled in the NADH-bound structure but with a bit different shape, indicating that the
inhibitor molecule was bound within the enzyme active site but not fully occupied. The
breaking point was another soaking experiment with the same concentration (13.5 mM)
of inhibitor but at pH 5.5, which produced an electron density map indicating a fully
occupied inhibitor in one of two molecules in the asymmetric unit. It is important to
realize that MIP synthase loses activity at pH lower than 7.2. At low pH, many of the
acidic side chains become protonated affecting inhibitor/substrate binding. At higher pH
(5.5), some of the acidic side chains are deprotonated, facilitating binding of the

inhibitor/ substrate.

3.7.1 Overall structure of the MIP synthase/NADVZ-deoxy-D-glucitol 6-(E)-
vinylhomophosphonate complex

The overall structure of the MIP synthase/NADVZ-deoxy-D-glucitol 6-(E)-
vinylhomophosphonate complex is similar to that of the MIP synthase/NADH complex
described in section 3.5. The RMSD between the two structures was only 0.49 A, which
is the lowest RMSD between two structures of MIP synthase seen so far. All of the
active site residues are ordered in both molecules in the asymmetric unit and the side
chains of the active site residues overlap with those of the NADH-bound structure very
well. C436 is the only exception in that its Ca moved 2.4 A and its side chain sulfur

moved 3.5 A away from their positions in the NADH-bound structure (Figure 3.25). In

155

fact, the position of C436 is identical to that of the apo, NAD+-bound, and 2-deoxy-D-
glucitol 6-phosphate-bound structures (9). The reason for this conformational change in

the NADH-bound structure is unclear.

O

'
r" t’

0; r

\

4“

 

Figure 3.25 Overlay of the MIP synthase/NADVZ-deoxy-D-glucitol 6-(E)-
vinylhomophosphonate complex structure in blue with the MIP synthase/NADH complex

structure in gold.

156

3.7.2 Conformation of NAD+

The conformation of NAD+ is similar to that of NADH in the NADH-bound
structure; there was no hydrogen bond pinching the nicotinamide and phosphodiester
together. However, the distance between the amide and phosphodiester is a bit closer
than that of NADH (3.2 A versus 3.8 A). The putative divalent cation is present in this
structure as well, but the fourth ligand in the NADH-bound structure (Figure 3.17 B) is
not present. The fourth ligand is now S439 0, identical to the coordination in the
M. tuberculosis MIP synthase structure (17). Interatomic distances between the putative
divalent cation and its four ligands are 2.26 A, 2.58 A, 2.52 A, 2.83 A respectively, bond
angles are: 1-0-3: 95.0°; 1-0-4: 104.7°; 2-0-3: 92.01°; 2-0-4: 112.8° (numberings are the
same as in Figure 3.18 A). The conclusion to be drawn from this observation is that the
nicotinamide moves away from the phosphodiester when the substrate or a substrate
analogue is bound in the active site, bringing the C4 of nicotinamide close to the substrate
C5 to carry out the oxidation step. Therefore, this conformation of the cofactor may
represent the enzyme in its active state. Also, changes seen between the NAD+-bound
structure and the NADH-bound structure are not purely due to the oxidation state of the

cofactor.

157

 

Figure 3.26 Overlay of cofactors in the MIP synthase/NADVZ-deoxy-D-glucitol 6-(E)-
vinylhomophosphonate complex structure in cyan and the MIP synthase/NADH complex
structure in gold. The putative divalent cation is in aqua in the NADH-bound structure
and in cyan in the 2-deoxy-D-glucitol 6-(E)-vinylhomophosphonate-bound structure.
Shown in orange is the fourth ligand of the putative divalent cation in the NADH-bound
structure. The third ligand water molecule is located at an identical position in both

structures.

158

3.7.3 Structure of the inhibitor 2-deoxy-D-glucitol 6-(E)-vinylhomophosphonate

The inhibitor 2-deoxy-D-glucitol 6-(E)-vinylhomophosphonate is fully occupied
in the enzyme active site as shown in the 1.6 a simulated annealing omit map (Figure
3.27A). The inhibitor molecule is well nestled within the enzyme active site in an
extended conformation, making hydrogen bond interactions as well as van der Waals
interactions with the active site residues. Table 3.11 lists all the interactions observed.
The phosphate group is in a transoid conformation relative to the inhibitor carbon
backbone. It is making hydrogen bonds with the main chain nitrogen atoms of S323-
G324—Q325-T326 as well as conserved lysine residues, K412 and K373. All of the
hydroxyl groups of the inhibitor except 0] make hydrogen bonds with side chains of
conserved active site residues. Figure 3.28 A depicts all of the interactions between the
inhibitor molecule and the active site residues. It is important to note that the putative
divalent cation chelates the water molecule as part of a hydrogen bond network that holds
02 and 03 of the inhibitor molecule in their positions. As opposed to the previously
reported observation (9), the oxidation at CS was not observed from the electron density
map, with the distance fi'om the inhibitor C5 to the nicotinamide C4 being 3.8 A, which is

a bit long for a direct hydride transfer.

159

   

 

0‘: 3%” 1M

 

- ' I '\ v'
. t a . , ./ I . ,
,2 , 39.4%... 1 “27"“! .33 (3324
@fi‘nv,’ A" ‘4’.” a" 33‘)? \
9:31;“: 793‘ t

   
   
    

\‘. ‘VA ['5 f r 0 . A".
"-,. 3: gage/3‘. afar-4i 651%”
Vii“? 4‘ W’”Q‘Qg"flhm-4E{I :-= ‘ “ sift!"
‘33.» l’awfiv‘gbwefizz .3: .9). 9.1.4,; /Q
1 -.. \\“ r .k ‘ - . ' V' 'U’ ' .‘v‘F’v‘
.‘ ”a" 'I ryv’p‘v‘f’q “I. ' ‘ ‘2 4‘o‘
\-. ”Aria flue-.4. ‘- 9:4 = '
yeast .3.’ ”an the. a
‘ .

. 4‘ ' . \
'6' v “'3: ‘

, ‘\\ A
.Q’I' ‘ 1. "a; ’9'
3.. (4‘98“ [”34 ‘ “1'.
0' 11..
. _. if” I
69%! ’4’, ’ ’ l. v’o.

9, 5.. .
v-
' $9:-

  

 

    

‘

{‘0‘ t '
<-'A\ .

Z \i‘“ "

Figure 3.27 (A) Simulated annealing omit map of the MIP synthase/NAD‘VZ-deoxy-D-
glucitol 6-(E)-vinylhomophosphonate structure contoured at 1.6 o. (B) 2Fo-Fc electron
density map of the MIP synthase/NADVZ-deoxy-D-glucitol 6-(E)-

vinylhomophosphonate structure contoured at 1.2 o.

160

 

Putative
divalent cation

D356

 

 

 

 

 

 

 

Figure 3.28 (A) Interactions between MIP synthase active site residues and the inhibitor
2-deoxy-D-glucitol 6-(E)-vinylhomophosphonate; the inhibitor is in lavender.

(B) Overlay of 2-deoxy-D-glucitol 6-(E)-vinylhomophosphonate in gold and phosphate
and glycerol in the NADH-bound structure in cyan. (C) Overlay of 2-deoxy-D-glucitol 6-
(E)-vinylhomophosphonate in gold and 2-deoxy-D-glucitol 6-phosphate in the previously
published inhibitor structure (9) in blue.

I61

Table 3.11 Interactions between MIP synthase and the inhibitor 2-deoxy-D-glucitol 6-

( E )-vinylhomophosphonate within a 3.5 A cutoff.

 

 

 

 

 

2-deoxy-D-glucitol 6-(E)- MIP synthase Distance (A)
vinylhomophosphonate # Residue atom
atom
C2 352 LEU CD1 3.41
C2 360 LEU CD1 3.26
C3 369 LYS CD 3.49
03 356 ASP ODl 2.45
O3 369 LYS NZ 3.28
04 438 ASP OD2 2.64
OS 369 LYS NZ 2.99
OS 489 LYS NZ 2.77
O] P 323 SER N 3.08
OlP 324 GLY N 2.53
OlP 325 GLN N 3.46
OlP 326 THR N 3.07
O2P 324 GLY N 3.06
OZP 325 GLN N 2.57
OZP 412 LYS NZ 2.95
O3P 326 THR N 2.88
O3P 326 THR OGl 2.59
O3P 373 LYS NZ 2.97

 

I62

 

The constellation of new data from the NAD+-bound, NADH-bound, and
2-deoxy-D-glucitol 6-(E)-vinylhomophosphonate-bound structures leads to the
conclusion that the previous modeling of the inhibitor 2-deoxy-D-glucitol 6-phosphate is
incorrect and the mechanism proposed should also be revised (9). First, D3 56 is in the
same conformation in all of the above-mentioned structures, as opposed to the
conformation in the previously reported MIP synthase/NADVZ-deoxy-D-glucitol 6-
phosphate structure (9), where it is flipped out of the active site (Figure 3.24). Secondly,
the phosphate moiety of 2-deoxy-D-glucitol 6-(E)-vinylhomophosphonate is in an
identical position to that of the phosphate in the NADH-bound structure (Figure 3.28B),
quite different from that of the inhibitor from the previously published MIP
synthase/NADVZ-deoxy-D—glucitol 6-phosphate structure (Figure 3.28C) (9). Third, as
shown in Figure 3.29, an analysis of surface charge distribution of the substrate-binding
site indicates the new phosphate-binding pocket (where the phosphate moiety is located
in both the MIP synthase/NADH complex and MIP synthase/NADVZ-deoxy-D-glucitol
6-(E)-vinylhomophosphonate complex structures) to be highly positive in charge, but in
the MIP synthase/NADVZ-deoxy-D-glucitol 6-phosphate structure (9), the phosphate-
binding site is negatively charged. Fourth, the structure of the MIP synthase/NADVZ-
deoxy-D-glucitol 6-(E)-vinylhomophosphonate complex was determined at pH 5.5,
closer to the active pH (7.2) than pH 4.5, at which the previous MIP synthase/NADVZ-
deoxy-D-glucitol 6-phosphate structure was determined. Finally, in the structure of the
MIP synthase/NAD‘VZ-deoxy-D-glucitol 6-(E)-vinylhomophosphonate, the
conformations of the substrate-interacting residues that are different from the previous

MIP synthase/NADVZ-deoxy-D-glucitol 6-phosphate structure (9) agree well with that of

163

 

Figure 3.29 GRASP drawing (16) of the electrostatic potential surface of the MIP
synthase. Circled dotted is the substrate-binding site centered at the phosphate-binding
pocket. Blue, EPS> 6kcal/mol; red, EPS<-6 kcal/mol; white, EPS—4). For clarity, residues
190-200, 351-366 were removed in this figure.

Note: The substrate-binding surface is positively charged for encapsulation of the
substrate. The pK. ’s of D-glucose 6-phosphate are pK.. =2.1, pKa = 6.8. At pH 7.0, the
substrate phosphate is negatively charged; there must be strong polar interactions

between the substrate and the substrate-binding site of MIP synthase.

the recently reported structure of MIP synthase from M. tuberculosis (17), especially the

region surrounding the phosphate-binding pocket

3.7.4 Modeling of the substrate and reaction intermediates

Based on the location and conformation of the inhibitor 2-deoxy-D-glucitol 6-(E)-
vinylhomophosphonate, the substrate D-glucose 6-phosphate was modeled in its
conformation necessary for cyclization (Figure 3.30). The phosphate portion was
overlaid onto that of the inhibitor molecule with 06 at the position of the inhibitor C7;
C6, C5, C4 were overlaid onto the inhibitor C6, C5, C4 respectively. The rest of the
substrate molecule was modeled such that none of the backbone atoms and the hydroxyl
oxygen atoms would collide with the side chains of active site residues.
The result of the modeling provided much information with respect to the interactions in
the enzyme active site. All but 03 of the substrate hydroxyl groups make hydrogen
bonds with the active site residue side chains. Figure 3.31 depicts most of the

interactions seen.

165

 

Figure 3.30 Modeling of the substrate D-glucose 6-phosphate (yellow) based on the

structure of the inhibitor 2-deoxy-D-glucitol 6-(E)-viny1homophosphonate (lavender).

The putative divalent cation is in cyan.

166

Putative
D433 [ divalent cation

D356

 

Figure 3.31 Interactions between MIP synthase and the modeled substrate D—glucose 6-

phosphate.

167

Using the modeled substrate as a guide each of the reaction intermediates was
also modeled. It is important to point out that once the hydride is transferred from the
substrate C5 to the nicotinamide C4, there may be a slight re-position of the substrate in
order to avoid the collision of the C5 carbonyl group with the nicotinamide ring. As a
matter of fact, a modeling of the reaction intermediate 5-keto-D-glucose 6-phosphate
without any change in the rest of the molecule resulted in the C5 carbonyl oxygen only
2.5 A away from the nicotinamide C4. This would never be the case during the catalysis.
A slight rotation of the phosphate portion about the C5-C6 axis resulted in a
conformation where this unfavorable steric collision can be avoided. When the
phosphate was rotated slightly, C4, C5, 05, and C6 were all kept in the same plane, since
the enolization follows oxidation immediately. Figure 3.32 shows the modeling of 5-
keto-D-glucose 6-phosphate (and the enolate intermediate). The slight rotation of the
phosphate moiety did not disrupt the interaction between the substrate and the enzyme,
instead there was an additional hydrogen bond between 03 and the D438 side chain as
denoted by the red dashed line in Figure 3.33.

The final reaction intermediate, myo-2-inosose l-phosphate was also modeled
(Figure 3.34). This cyclic intermediate is in a conformation similar to the substrate but
with an additional hydrogen bond between 03 and D43 8, which contributes to the

stabilization of the cyclic conformation.

168

 

Figure 3.32 Modeling of the reaction intermediate 5-keto-D-glucose 6-phosphate. The

substrate D-glucose 6-phosphate is in yellow, and the 5-keto-D-glucose 6-phosphate (and

the enolate) is in blue.

169

D356

 

Figure 3.33 Interactions between the modeled 5-keto-D-glucose 6-phosphate (enolate)

and the active site residues.

170

Putative divalent
cation

D356

 

K489

K373

Figure 3.34 Interactions between the modeled myo-Z-inosose l-phosphate and the active

site residues.

171

3.8 Proposed mechanism of MIP synthase

The apparent discrepancy between the structure of the MIP synthase/NADVZ-
deoxy-D-glucitol 6-(E)-vinylhomophosphonate complex described above and the
structure of the MIP synthase/NAD+/2-deoxy-D-glucitol 6-phosphate complex published
previously (9) (Figure 3.28 B) calls into question the mechanism proposed previously.
Based on the new structure of MIP synthase/NAD+/2-deoxy-D-glucitol 6-(E)-
vinylhomophosphonate and modeling of the substrate and reaction intermediates, a new
mechanism was proposed.

In the first step, the substrate is oxidized at C5 by NAD+ (Figure 3.35). This
involves a direct hydride transfer from the C5 of D-glucose 6-phosphate to the C4 of the
nicotinamide moiety of NAD+; this is confirmed by the crystal structure where the
nicotinamide is located in a suitable orientation for hydride transfer to occur. In concert,
a proton is lost from the C5 hydroxyl group of D-glucose 6-phosphate; this proton can be
transferred to the K369 terminal nitrogen atom, which is 2.8 A away from the OS of the
substrate. D3 20, adjacent to K3 69, could then accept this proton in a proton-shuffling
system. It is also possible that the terminal nitrogen atom of K489, which is 2.92 A away
from the OS, pulls off the proton from OS.

The second step is the enolization. During the enolization, the pro-R hydrogen of
C6 is eliminated. From the crystal structure, either the dibasic phosphate monoester
(Figure 3.35) or K489 (Figure 3.36) may act as the base at the enolization step. The
phosphate monoester acting as the base at the enolization step is similar to that proposed
for 3-dehydroquinate synthase (20,21). This mechanism takes advantage of phosphate in

a transoid conformation relative to the carbon backbone of the substrate. From the

172

modeling of the 5-keto-D-glucose 6-phosphate based on the crystal structure, K489 is
also in a suitable position to remove the pro-R hydrogen of C6. The negative charge of
the enol can be stabilized by K369.

1n the aldol condensation step, the phosphate could transfer the proton abstracted
from C6 to OI. If the pro-R hydrogen of C6 was removed by K489 at the previous step,
then K412 may deliver a proton to 01 in a proton-shuffling system with N350, and the
developing negative charge on 01 can be stabilized by K3 73. It is also possible that MIP
synthase utilizes the type I aldolase mechanism where K369 could form a Schiff base
with C5, followed by the aldol condensation step, and the developing negative charge on
01 could then be stabilized by K412.

The last step is the reduction by NADH. The hydride that was transferred in the
first step to the nicotinamide C4 returns to the C5 of the intermediate myo-2-inosose 1-
phosphate. Using the same proton-shuffling system, a proton could then be transferred to
the C5 ketone oxygen from D320, via K369.
At this point, there is still not enough structural evidence regarding whether the substrate
binds in its cyclic form followed by ring opening catalyzed by the enzyme, or binds in its
acyclic form, which constitutes less than 0.4 % of D-glucose 6-phosphate in solution.

Verification of the mechanism proposed above requires mutational investigation
and further structural investigation of MIP synthase in complex with various structural

analogues of the reaction intermediates.

173

 

 

 

‘ OH H NO C K373
| 356 2 : H
. H0 “3 H N020
NR 7 ‘ 2
O —
I 0'14 .
t H
+ I \ 0:) iii
K412 l O __
H) “L. o’ “*3"
K489 ’H N
0320 3
0320 H2N 1 LL! K489
2' K36
K36
P438
OH
HO
HO OH
H203PO 0 ”2 *HaN K369
OH *HgN
110139
P438
l
K373 , 0“
H3 H0
’ OH
K431” ”WK ~
_ H-o man:
0
o _ ’0 O
N °./P K369
o +ng

 

Figure 3.35 Proposed mechanism of MIP synthase where the phosphate monoester acts

as the base in the enolization step.

174

    

 

N350 P438
‘, OH HZNOQC T373 I
K373 355 _ I H H
‘ 4 I . HO 3 H NO C
’O 0 NR 2 2
I H —
K412 \ H \/ / ’O\ H OH‘I
I
H203PO * (412 ii
('3 H203 '—
5323,Gs24,0325,T326 H A HzN ‘H3N
, H N
NH3 2 2 032° I_L, K369
0320 z K369
K489 K489
* p438
K373 ,
OH NNHa HO
HO
HO K412
OH NNH3+
O 03”? O 'H3N K369
H3N
OH 1
K489

. P438
I
OH

   
 

t
0320 HzN K489

K369

Figure 3.36 An alternative mechanism of MIP synthase where K489 acts as the base in

the enolization step.

175

3.9 Conclusions

The structures of S. cerevisiae MIP synthase in its apo form, NAD+-bound form,
NADH-bound form, and in complex with an inhibitor, 2-deoxy-D-glucitol 6-(E)-
vinylhomophosphonate were newly determined. It was observed previously from the
structures of S. cerevisiae MIP synthase in the presence of partially occupied NAD+, and
in complex with fully occupied NAD+ and an inhibitor 2-deoxy-D-glucitol 6-phosphate
(9) that active site residues 351-409 were disordered in the presence of partially occupied
NAD+ but became ordered when NAD+ and the inhibitor are bound with full occupancy.
Therefore, it was concluded that the substrate binding folds the active site. Compared to
the previously determined structures, the newly determined structures presented
similarities as well as some significant differences, leading to the proposal of a new
catalytic mechanism of MIP synthase.

In the structure of apo MIP synthase residues 351-3 75 were missing; there were
no structural differences within the Rossmann fold NAD+-binding domain between the
apo structure and the previous low occupancy NAD+-bound structure (9). This indicates
that the NAD+-binding domain of MIP synthase is very rigid, while the enzyme active
site is very flexible. Compared to the previous low occupancy NAD+-bound structure
( 9), residues 376-409 became ordered in the structure of apo MIP synthase; this increased
order appears to be puzzling.

The structure of MIP synthase in the presence of fully occupied NAD+ represents
the state of MIP synthase before the oxidation of the substrate or after the reduction,
which leads to the formation and release of the product. Residues 362-3 79 are disordered

in 5 out of 6 molecules in the asymmetric units in both the P21 and C2 forms of the MIP

176

synthase/NAD+ complex crystal structures. Thus, it appears that the previous observation
that residues 351-409 are disordered in the presence of partially occupied NADJr (9)
could be due to not fully occupied NAD+. The fully folded active site of molecule C in
the structure of the P21 crystal form and the presence of a significant electron density
feature in its active site affirms the previous notion (9) that small molecule binding
causes the folding of the active site.

Consistent with the assumption that the enzyme is at its catalytically active state
after NAD+ is reduced to NADH, the structure of the MIP synthase/NADH complex
presented a phenomenal picture that is completely different from all structures seen thus
far. In this structure all the active site residues are completely ordered, and the
conformation of the nicotinamide changed significantly with a putative divalent cation
eliminating the tight hydrogen bond between the nicotinamide and the phosphodiester
that was present in the NAD+-bound structure. Addition of EDTA in the soaking
stabilizer eliminated the electron density feature between the nicotinamide and the
phosphodiester, indicating that it is very likely to be a divalent cation. Two small
molecules, phosphate and glycerol, are bound in the enzyme active site mimicking the
substrate binding. The unambiguous position of the phosphate-like anion is completely
different from that of the inhibitor molecule in the previous structure of the MIP
synthase/NAD+/2-deoxy-D-glucitol 6-phosphate complex, calling into question some of
the details in the previously published model (9).

The structure of the MIP synthase/NAD+/2-deoxy-D-glucitol 6-(E)-
vinylhomophosphonate complex provided insights into the catalytic mechanism of MIP

synthase. The inhibitor molecule is bound in the enzyme active site in an extended

177

conformation. The position of the phosphate moiety of the inhibitor is identical to that of
the modeled phosphate anion in the NADH-bound structure, as opposed to the previously
published 2-deoxy-D-glucitol 6-phosphate-bound structure (9), where the phosphate
moiety of the inhibitor 2-deoxy-D-glucitol 6-phosphate sits in the opposite side of the
enzyme active site where D356 is located in all newly determined structures. According
to the new structural data, modeling of the substrate in its pseudocyclic conformation
necessary for the aldol cyclization was performed, and a new mechanism proposed. The
fact that 2-deoxy-D-glucitol 6-(E)-vinylhomophosphonate, a structural analogue of the
substrate in its transoid conformation, is the most potent inhibitor synthesized so far
strongly supports the mechanism where the phosphate monoester can act as the base in
the enolization step to pull the pro-R hydrogen off the C6. From the crystallographic
structural data, the possibility that MIP synthase could act as a type I aldolase could not
be ruled out.

In concert, it can be concluded that MIP synthase experiences conformational
changes at various stages during the reaction. At the start of each turnover, the enzyme is
tightly bound to NAD+ since NAD+ is a prosthetic group, and the active site is disordered
as seen in the NAD+-bound structure. Once the substrate is bound in the enzyme active
site, the entire active site folds to completely encapsulate the substrate within its active
site. The nicotinamide changes to the conformation seen in the structures of the MIP
synthase/NADH complex and the MIP synthase/NADVZ-deoxy-D—glucitol 6-(E)-
vinylhomophosphonate complex. This motion of the nicotinamide brings the substrate
C5 close to the nicotinamide C4 to facilitate the oxidation. All non-redox reactions occur

within the same active site, NADH then reduces the cyclized reaction intermediate, myo-

178

2-inosose l-phosphate to produce MIP, and the mobile region of the active site will open
up to release the product as seen in the NAD+-bound structure ready for the next cycle of
turnover.

Further experiments are needed to verify the mechanism proposed above. These
may include: (1) Elucidation of the structures of MIP synthase in complex with different
structural analogues of the substrate as well as reaction intermediates, among which the
structure in complex with one of the reaction intermediates, myo-2-inosose l-phosphate.
would be of particular interest; (2) Elucidation of the structures of MIP synthase in
complex with the previously utilized inhibitors at a higher pH; (3) Elucidation of the
structures of MIP synthase mutants, namely S323A, D356A, K369A, K373A, K4l2A,
D43 8A, and K489A, in complex with the substrate or a substrate analogue; (4)
Evaluation of MIP synthase activity in the presence of various divalent cations, including

Zn”. an+, Mg2+, Ni2+, etc.

179

3.10 Literature cited

1.

12.

Jack, A., and Levitt, M. (1978) Acta CrystallogrA 34, 931

Press, W. H., Flannery, B. P., Teukolosky, S. A., and Vetterling, W. T. (1986)
Numerical Recipes, Cambridge University Press, Cambridge

Kirkpatrick, S., Gelatt, C. D., and Vecchi, M. P. J. (1983) Science 220, 671
Metropolis, N., Rosenbluth, M., Rosenbluth, A., Teller, A., and Teller, E. (1953)
J. Chem. Phys. 21, 1087

Verlet, L. (1967) Phys. Rev. 159, 98

Jones, T. A., Zou, J. Y., Cowan, S. W., and Kjeldgaard, M. (1991) Acta
Crystallogr. A 47, l lO-l l9

Stein, A. J ., and Geiger, J. H. (2000) Acta Crystallogr D Biol Crystallogr 56 ( Pt
3), 348-350

Otwinowski, Z., and Minor, W. (1996) Methods Enzymol 276, 307-326

Stein, A. J ., and Geiger, J. H. (2002) J Biol Chem 277, 9484-9491

Jones, T. A. (1985) Methods Enzymol 115, 157-171

Brunger, A. T., Adams, P. D., Clore, G. M., Delano, W. L., Gros, P., Grosse-
Kunstleve, R. W., Jiang, J.-S., Kuszewski, J., Nilges, M., Pannu, N. 5., Read, R.
J ., Rice, L. M., Simonson, T., and Warren, G. L. (1997) Crystallography & NMR
System, version 1.0, Yale University

Brunger, A. T., Adams, P. D., Clore, G. M., DeLano, W. L., Gros, P., Grosse—
Kunstleve, R. W., Jiang, J. S., Kuszewski, J., Nilges, M., Pannu, N. 3., Read, R.
J., Rice, L. M., Simonson, T., and Warren, G. L. (1998) Acta Crystallogr D Biol

C rystallogr 54 ( Pt 5), 905-921

180

14.

15.

16.

20.

2|.

Navaza, J. (1994) Acta Crystallogra. A 50, 157-163

Laskowski, R. A., MacArthur, M. W., Moss, D. S., and Thornton, J. M. (1993) J.
Appl. Cryst. 26, 283-291

Nicholls, A., Sharp, K. A., and Honig, B. (1991) Proteins 11, 281-296

Nicholls, A. Graphical Representation and Analysis of Structural Properties
Norman, R. A., McAlister, M. S., Murray-Rust, J ., Movahedzadeh, F., Stoker, N.
G., and McDonald, N. Q. (2002) Structure (Camb) 10, 393-402

Maeda, T., and Eisenberg, F., Jr. (1980) J Biol Chem 255, 8458-8464

Tian, F., Migaud, M. E., and Frost, J. W. (1999) J. Am. Chem. Soc. 121, 5795-
5796

Carpenter, E. P., Hawkins, A. R., Frost, J. W., and Brown, K. A. (1998) Nature
394, 299-302

Bender, S. L., Widlanski, T., and Knowles, J. R. (1989) Biochemistry 28, 7560-

7572

181