1““ 2-1..
1,:
1.8;”;
5.23: .
x! ‘ n
v...

.3 . :y ‘ 5
. y. 1:. olve‘,

I.
.8: fl}... .

. .I no.
1" .2.)
.c’l~. .

 

. , 4 . ~
'

...&...II E! I.

X-RAY CRYSTALLOGRAPHIC STUDIES OF SNAP19ORcRd (SMALL NUCLEAR
RNA ACTIVATING PROTEIN) COMPLEX AND E. COLI GLYCOGEN SYNTHASE

By

Fang Sheng

A DISSERTATION

Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of
DOCTOR OF PHILOSOPHY
Department of Chemistry

2008

ABSTRACT

X-RAY CRYSTALLOGRAPHIC STUDIES OF SNAP19ORcRd (SMALL NUCLEAR
RNA ACTIVATING PROTEIN) COMPLEX AND E. COLI GLYCOGEN SYNTHASE

By

Fang Sheng

Small Nuclear RNA Activating Protein Complex (SNAPc) is a basal transcription
factor that binds to the Proximal Sequence Element (PSE) of human Small Nuclear RNA
(snRNA) gene promoter. The Myb-domain RcRd repeats of the largest SNAPc subunit
SNAP190, SNAP19ORcRd, was believed to be required and sufficient for SNAPc to bind
to the PSE. Despite extensive crystallization trials, attempts to obtain the
SNAP19ORcRd-PSE complex structure in order to understand their specific interaction
were unsuccessful. Our EMSA DNA-binding assay did not repeat the results reported in
literature. Instead, it shows very limited binding of SNAP19ORcRd to the PSE. The size-
exclusion chromatogram of SNAP19ORcRd suggests a high aggregation state. This result
indicates a folding and aggregation problem with peptide SNAP19ORcRd, which could
hinder SNAP19ORcRd binding to the PSE, causing crystallization failure.

Escherichia coli glycogen synthase (ECGS, EC2.4.1.21) is a retaining
glycosyltransferase (GT) that transfers glucose from adenosine-diphosphate glucose
(ADPGlc) to a glucan chain acceptor. Due to the acceptor analogue HEPPSO, we
obtained the first structure of the catalytically active, closed conformation of two-domain
protein GS, which hosts a deeply buried active site in the narrow interdomain cleft.

Comparison of the ligand-bound E. coli GS structures to that of the apo-deS (C78;

C4085) reveals a 15.2° overall domain-domain closure. The structure of the catalytically
inactive mutant E3 77A complexed with oligosacchrides suggests that the glucan chain
only binds to the GS N-terminal domain surface. This complies with the frequent open-
close motions required by the undisturbed GS catalysis. The absence of a glucose moiety
of ADPGlc in two of the E377A complex structures indicates that Glu377 plays an
important role in positioning and stabilizing the glucose that is to be transferred. Our wild
type E. coli GS (thS) structures demonstrate that residues Arg300 and Lys305 are close
to the ADP phosphate and probably act as Lewis acids, working electrostatically to make
the phosphate a better leaving group. Asp137 is suggested to play an important role in
positioning these acceptor nucleophile analogues. These results are in agreement with
previous biochemical and mutagenesis studies. The observation of the glucose and a
DGM-like species in four of the thS crystal structures strongly suggests a SNl
mechanism for the GT-B retaining enzymes over the once-prevalent double replacement
mechanism. The highly reactive SNl mechanism intermediate, DGM, is probably
stabilized by the nearby phosphate moiety, the incoming nucleophile, and the Hi5161
backbone in the GT-B retaining enzymes. GT-B retaining enzyme family boasts
approximately a thousand important sugar-modifying enzymes that are believed to share
a virtually identical active site. Finally, based on our E. coli GS structures and the
chimeric studies of GBSS and S811, we speculate that the region (0L13 - a18) in starch
synthase is involved in positioning the two domains and that interaction of this region of

GBSS with amylopectin possibly orients the domains for catalysis.

To my teacher, Prof. Zhanru Liao, who introduced me into research and

her infectious passion for science inspired me along the road.

iv

ACKNOWLEDGEMENTS

I thank my advisor Dr. James H. Geiger for his support, inspiration, encouragement
and most of all, his never-ending faith in me. Without that, I couldn’t have ordered and
tried a wide range of rare and expensive chemicals and among them find the magic
reagent HEPPSO for GS crystallization. I thank our collaborator, Dr. Jack Preiss, for
giving me the Opportunity to work on GS. Now the puzzle of the glycogen synthesis finds
its last piece: G8. I also thank our collaborator, Dr. Bill Henry, for being too busy to nag
me but never too busy to be available and helpful. Talking with you has always been
instructive and delightful. I am in gratitude to my committee members Dr. James K.
McCusker, Dr. Thomas J. Pinnavaia, and Dr. David P. Weliky.

I want to say thank you to Dr. Stacy Hovde who took time out fiom her tight
schedule to show me the ropes in the beginning. I appreciate your company in those
around-the-clock synchrotron shifts. Our 2 AM tricycle racing in APS cycle was fun and
did cheer us up from hundreds of poor diffracting crystals. I thank Dr. Xiangshu J in for
her invaluable help with the structure refinement and Dr. Sara E. Cnudde for paving the
path to the usage of CCP4. I want to thank Xiaofei Jia for his great help with NMR trials
and helpfiil discussion. I wish you the best in your career. I also want to thank Joe
Leukyam at GTSF facility for assistance with [1-13C]-ADPGlc purification and Dr. Dan
Holmes for help with NMR experiments. Paul Reed, you are the best computer technician
anyone can ever ask for, thank you for putting up with my numerous and silly questions. I
also want to thank Lisa in the graduate office for having the answers to all things

bureaucratic and accommodating me to teach my favorite classes.

My sincere gratitude extends to all Geiger lab members, Lei Feng, Suzan, Andy,
Blanka, Dorothy, Aimee, Justin and Kathy. Knowing you and working with you in lab
has been a pleasure. I also want to thank my friends Jess Gunn for helping me with
preparing my seminar, and Christine Kalcic for introducing me to line dancing which I
really enjoy.

I want to say that my previous advisor in China, Prof. Zhanru Liao, was one of the
biggest inspirations in my life. Her passion for science encouraged me to fly overseas to
the States to pursue my Ph.D. I thank my parents and my dear brother who believe in me
no matter what and whose love sustain me. Last, but not least, I thank my husband, Jian

Yang, whose love and support have accompanied me and made the pursuit of my Ph.D

joyful.

vi

TABLE OF CONTENTS

ACKNOWLEDGEMENTS ......................................................................... v
TABLE OF CONTENTS .......................................................................... vii
LIST OF TABLES ............................................................................................................. xi
LIST OF FIGURES ................................................................................. xii
LIST OF ABBREVIATIONS ........................................................................................... xx
CHAPTER I
INTRODUCTION .............................................................................................................. 1
1.1. Small Nuclear RNA Activating Protein Complex (SNAPc) l9ORcRd ................... 1
1.1.1. Transcription Regulation and Transcription Factors ................................ 1
1.1.2. Small Nuclear RNA Promoters (PSE, TATA and DSE).............. ..............2
1.1.3. Small Nuclear RNA Activating Protein Complex (SNAPc). ...................... 4
1.1.4. SNAP19ORcRd and Myb Domain ...................................................... 7
1.1.5. Objectives ................................................................................ 14
1.2. Glycogen Synthase ............................................................................. 15
1.2.1. Glycogen ................................................................................. 15
1.2.2. Starch ..................................................................................... 17
1.2.3. Glycosyltransferase .................................................................... 19
1.2.4. The Retaining Glycosyltransferase Mechanism Controversy .................... 22
1.2.5. Previous Studies on Bacterial Glycogen Synthase ................................. 30
1.2.6. Significance of E. coli GS Structural Studies ........................................ 32
1.3. References ....................................................................................... 35
CHAPTER II
CRYSTALLIZATION AND PRELIMINARY X-RAY DIFFRACTION ANALYSIS
OF SNAP19ORcRd ........................................................................................................... 45
11.1. Experimental Procedures ..................................................................... 45
11.1.1. Transformation and Overexpression of SNAP19O RcRd ........................ 45
11.1.2. Purification of SNAP19O RcRd ..................................................... 47
11.1.3. Purification of DNA and Annealing .................................................. 49

11.1.4. Crystallization of SNAP19ORcRd, SNAP19ORcRd/ DNA and
SNAP19ORcRd/TBP/ DNA Complexes.............................................50

11.1.5. Electrophoretic Mobility Shift Assay (EMSA) .................................... 52
11.2. Results and Discussion ........................................................................ 53
11.2.1.Preparation of SNAP19ORcRd ........................................................ 53

vii

11.2.2. Crystallization Trials ................................................................... 55

11.2.3. EMSA DNA-binding assay ........................................................... 55
11.3. Conclusion ...................................................................................... 60
11.4. Appendix ....................................................................................... 61
11.5. References ....................................................................................... 62
CHAPTER III .
THREE DIMENSIONAL STRUCTURES OF Escherichia Cali GLYCOGEN
SYNTHASE AND ITS COMPLEXES ........................................................ 63
111.1. Theory ......................................................................................... 63
111.1.1. Structure Determination from X-ray Diffraction Data .......................... 63
111.1.2. Molecular Replacement ............................................................. 64
111.1 .3. Structure Refinement ................................................................ 66
111.2. Experimental Procedures .................................................................... 68
111.2.1. Protein Overexpression .............................................................. 68
111.2.2. Protein Purification ................................................................... 69
III.2.2.A. His-tagged GS protein ....................................................... 69
111.228. Untagged deS protein .................................................... 7O
111.2.3. GS and GS Complex Crystallization .............................................. 71
111.2.4. Diffraction Data Collection and Processing ...................................... 73
III.2.5.'3C-and‘H-‘3C NMR Experiment .................................................... 73
III.2.5.A. Preparation of D-[l-I3C]-ADPGlc from oc-D-[l-UC] glucopyranosyl
l-phosphate .................................................................. 73
111.253. NMR experiment setting .................................................... 75
111.2.5.C. NMR sample preparation ................................................... 76
IH.2.5.D. NMR spectra ................................................................ 76
111.3. Model Building and Structure Refinement ............................................... 81
111.3.]. apo deS (C7S: C428S) ............................................................ 81
111.3.2. Wild - type GS in Complex with ADP and D—glucopyranosylium (DGM)
(thSa) ............................................................................... 86
111.3.3. Wild - type GS in Complex with ADP and Glucose (thSb, thSc and
GSd) ................................................................................... 88
111.3.4. GS Mutant E377A in Complex with ADP and Oligosaccharides .............95
111.3.5. GS Mutant E377A in Complex with ADP ........................................ 97
111.4. Escherichia coli GS Structures ............................................................. 99
111.4.1. apo deS (C7A;C428S) Structure ................................................. 99

viii

111.4.2. GS- ADP-Glucose-HEPPSO Complex Structures .............................. 102

III.4.2.A. ADP binding site .............................................................................. 104
III.4.2.B. Glucose binding site ......................................................................... 108
III.4.2.C. Acceptor analogue HEPPSO binding site ........................................ 111
111.4.3. GS -ADP-DGM-HEPPSO Complex Structure ........................................ .114
111.4.4. E377A-ADP-HEPPSO Complex Structure ................................................. 118
111.45. E377A- ADP-Oligosaccharide Complex Structure ..................................... 119
III.4.5.A. Interdomain cleft Oligosaccharide-binding site ................................ 120
III.4.5.B. The N—terminal surface Oligosaccharide-binding sites ..................... 126
III.4.5.C. The glucose unit configuration in Oligosaccharide-bound E377A...133
III.4.5.D. Binding mode analysis ..................................................... 134
III.4.5.B. Oligosaccharide-binding causes little protein conformation change134
III.4.5.F. Oligosaccharide conformation analysis .................................. 135
111.5. E.coli GS Structure Analysis ................................................................................. 139
111.5. 1. Open-Close: GS Dynamic Motion ............................................................... 139

111.5. 1 .A.Comparison of ADP conformations in GS open and closed form. . . 142
III.5.1.B. Large displacement of KTGGL loop in GS open and closed form.. 143

111.5. 1.C. What causes the enzyme to close? ................................................... 146
III.5.1.D. Domain-wise ligand binding facilitates GS open-close motion ....... 148
111.5. 1 .E. Substrate binding order ................................................................... .149
111.5.2. Important Residues ...................................................................................... 150
111.5.2.A. Aspl37 is the +1 sugar positioning residue .................................... 151
111.5.2.B. Hisl61 participates in —1 sugar positioning and intermediate DGM
stabilization ................................................................ 152
III.5.2.C. The catalytic Lewis acid Arg300 side-chain switches in and out of the
GS active site in response to the presence of ADP ................... .152
111.5.2.B. Catalytic Lewis acid residue Lys305 .................................... 155
111.5.2.B. Glu377 is responsible for the —1 sugar positioning and helps Lys305
maintain positive charge ................................................... 157

111.53. GS Complex Structures Support SN] Mechanism Over the Double

Displacement Mechanism ........................................................................ 160
III.5.3.A. DGM intermediate ........................................................................... 160
III.5.3.B. DGM stabilization in GS .................................................................. 161

III.5.3.C. Effort to verify the existence of carbocation DGM using NMR. . ...162
III.5.3.D. The substrate phosphate group deprontonates the incoming acceptor

nucleophile .................................................................. 169
III.5.3.E. Evidence against the double-displacement mechanism ................... 169
111.5.4. The GS Catalytic Scenario Suggested by E. coli GS Structural Studies ...... 171

ix

111.6. Mechanism Implication for Starch Synthase .......................................................... 171
111.6.]. Plant Starch Synthases Share a Similar Active Site and Catalytic

Mechanism with Bacterial GS ........................................................... 171

111.6.2. Glucan Chain Binding in Starch Synthase ......................................... 174

111.6.3. Insights into GBSS and SS Specificities ............................................. 175

111.7. Conclusions ............................................................................................................ 177
111.8. References ..................................................................................... 179

LIST OF TABLES

Table 1.1. The sequence alignment of SNAP190 Myb domain and human A-, B-, c-Myb
domains. Important DNA-interacting residues are in bold and underlined. Three well-
defined helices in the human c-Myb domain are illustrated as tandem connected cylinder

(Hl , H2, and H3) .................................................................................... 12
Table 1.2. The sequences of the PSEs present in the wild type and mutant probes.
Uppercase letters correspond to the PSE sequence from the mouse U6 promoter, with
red-colored characters corresponding to mutations ............................................. 13

Table 1.3. Kinetic data of DGM mimics in GT—B retaining enzymes ......................... 29

Table 11.1. The SNAP19ORcRd sequence used in crystallization trials and EMSA assay46

Table 11.2. Oligonucleotides used in crystallization trials ...................................... 51
Table 11.3. DNA sequences used in Figure 11.6 ................................................. 59
Table 111.1. GS complex crystallization trials ................................................... 72

Table 111.2. Data collection and refinement statistics of E.coli GS mutant and their
complexes ............................................................................................ 82

Table 111.3. Data collection and refinement statistics of wild type E.coli GS
complexes ............................................................................................ 92

Table 111.4. Comparison of the environment of DGM and glucose. . . . . . . . . . . . ..1 16
Table 111.5. Conformation of Oligosaccharides when bound to GS, MalP, and GP ....... 137

Table 111.6. Kinetic parameters of wild type glycogen synthase and mutants. . . . . . ....151

Table 111.7. Selective conserved residues in bacterial GS and plant starch synthases.
Those residues critical to GS activity and their equivalents in $83 are colored pink.
Multiple sequence alignment was conducted with DNASTAR. GS sequences used were:
Escherichia. coli GS (POA6U8), Agrobacterium tumefaciens GS (AAD03474),
Pyrococcus abyssi GS (NP_1~25769), Sequences of granule-bound starch synthases from
various organisms were those of barley (AAL77109), maize (PO4713), potato
(CAA41359), and rice (P19395), Other starch synthases used were those of maize SS1
(AAB99957), potato SSI (P93568), wheat SSI (Q43654), wheat SSIIa (BAE48798),
maize SSIIa (AAS77569), potato SSII (CAA61241), potato SSIH (Q43846), wheat SSIII
(AAF 87999) and Chlamydomonas reinhardtii SS (AAC17971) ........................... 173

xi

LIST OF FIGURES

Images in this dissertation are presented in color

Figure 1.1 Composition of pol H and pol 111 snRNA promoters in higher eukaryotes.
Typical eukaryote RNA pol II snRNA genes are U1-U5 genes and RNA pol HI snRNA

genes are U6 and 7-SK genes ....................................................................... 3
Figure 1.2. Architecture of SNAPc complex ....................................................... 5
Figure 1.3. The composition of SNAP190 ......................................................... 6

Figure 1.4. Ternary Myb protein-enhancer binding protein-DNA complex (1H88.pdb).
The Myb protein (39-190) is shown as cartoon and three repeats R1, R2, and R3 are in
green, yellow, and blue, respectively ............................................................... 9

Figure 1.5. The close-up view of the interaction between human Myb protein R2, R3 and
DNA in the ternary protein-DNA complex (1H88.pdb). The conserved tryptophan
residues are shown as magenta sticks ............................................................ 10

Figure 1.6. Top: Typical GT-A fold illustrated by protein SpsA. Bottom: Structure of the

2+
GT-B fold enzyme thB. A Mn ion in thB active site is depicted as blue sphere. The
fold name is after their initial observation in the SpsA (Bacillus subtilis glycosyl-
transferase) and thB (DNA B-glucosyltransferase structures) ............................... 21

Figure 1.7. Cartoon presentation of Agrobacterium tumefaciens GS complexed with ADP
(blue). The N-terminal peptide 15-21 containing the KTGGL loop and Asp21 and the C-
terminal residue Arg299 and Lys304 are colored red because they were indicated by
mutagenesis and kinetic studies to be important in GS catalysis ............................... 22

Figure 1.8. Proposed double-displacement mechanisms for GT retaining enzymes.
Two essential components of this mechanism, the catalytic nucleophile and glucosyl-
enzyme intermediate, are circled and framed, respectively .................................... 24

Figure 1.9. SNl-like mechanism D-glucopyranosylium ion (DGM) intermediate is framed,
which is otherwise considered a transition-state in the SNi-like mechanism. In both cases,

positively charged DGM is stabilized by leaving group AMP-phosphate and the incoming
nucleophile 4-OH group of sugar ................................................................. 26

Figure 1.10. DGM (A) and DGM mimics (B-G) that have been tested in kinetic studies of
GT-B retaining enzymes. The related kinetic data are listed in Table 1.3 ................... 28

Figure 1.11. Reactions catalyzed by GT-B retaining enzyme OtsA, MalP, and GP ........ 31

xii

Figure 11.1. Predicted secondary structure of SNAP19ORcRd (390-518) by method
PSIPRED. Helix and coil are depicted as green cylinder and black straight lines. The
confidence of predication (Cont) is indicated by the height of cyan columns. Pred and
AA stands for the predicted secondary structure and target sequence, respectively ....... 46

Figure 11.2. SDS-PAGE gel showing the GST-tagged RcRd peptide (lane 3 and 4) and the
RcRd peptide alone (lane 6 and 7). The GST-tag was eluted by 100 mM reduced
glutathione solution afterwards (lane 9 and 10) .................................................. 48

Figure 11.3. Diagram of SNAP19ORcRd purification through Source-Q column .......... 49

Figure 11.4. SDS-PAGE gel of SNAP19ORcRd from Source-Q column. Lane 2, 3, and 4
corresponds to the fraction 27, 28, and 29, respectively ........................................ 49

Figure 11.5. Size-exclusion chromatogram results from Sephacryl S-300 HiPrep 16/60
column. The blue line is Bio-Rad molecular weight standards (158 kDa bovinegamma-
globulin, 44 kDa chicken ovalbumin, 17 kDa equine myoglobin, and 1.3 kDa vitamin
B12. The red line is from SNAP19ORcRd. Experiment was performed at lmL /min flow
rate and the fractions were 5 mL .................................................................. 54

Figure 11.6. An EMSA was performed with a probe containing B-MPSE (lane 1-6), S-
MPSE (lane 7-12), or D-MPSE (lane 13-18). B-MPSE, S-MPSE, and D-MPSE are blunt
ends-, single overhang-, and double overhang- mouse U6-21 PSE, respectively

and their sequences are listed in Table 11.2. In lane 1, 7, and 13, no SNAP19ORcRd or
recombinant SNAPc (rSNAPc) were added to the probes. Lane 2-5, 8-11, and 14-17
contain increasing amount of SNAP19ORcRd (0.1, 0.3, l, 3 pg). Lane 6, 12, and 18
contain 3 pg mSNAPc .............................................................................. 56

Figure 11.7. An EMSA was performed with varied length of probes. mSNAPc was added
to the 4* labeled lanes. Lane 1-2, 3-4, 11-12, 23-24, and 29-30, 35-36 contain increasing
amount of mSNAPc (1 and 3 ug). All probe sequences are listed in Table 11.3. Lane 7-
10, 13-16, 19-22, 25-28,31-34, and 37-40 contain increasing amount of SNAP19ORcRd
(0.1, 0.3, 1, 3 pg) ..................................................................................... 58

Figure 111.1. Representative His-tagged E. coli GS SDS-PAGE gel. The most left lane is
the molecular weight standard. Each elution lane accounts for the collection of a 5 mL
elution buffer fraction .............................................................................. 69

Figure 111.2. Ecoli deS SDS-PAGE. The second left lane is the molecular weight
standard and all other lanes are elution from Source-Q column with increasing
concentration of KCl ................................................................................ 7 O

xiii

Figure 111.3. HPLC spectrum fiom the DEAE-5PW column. Flow rate: 0.6 mL /min,
Gradient: 2 min: 0% buffer B; 40 min: 25 % buffer B; 60 min: 100 % buffer B. Fraction
ADPGlc elutes in the range of 14 —1 8 % buffer B .............................................. 75

1
Figure 111.4. NMR spectra of the starting material [1- 3C]-Glc-l-P and batch I [1-13C]-
ADPGlc ............................................................................................... 77

. 13
Figure 111.5. NMR spectra of batch 11 [l- C]-ADPGlc ....................................... 78

. . . l3
Flgure 111.6. NMR spectra of the startlng materlal [UL- C6]-Glc-l-P and batch 1 [UL-

13C6]-ADPGlc ....................................................................................... 79
Figure 111.7. A crystal of deS ................................................................... 81
Figure 111.8. Ramachandran plot of E. coli deS. Phi (degrees) is x and Psi (degrees) is y
........................................................................................................... 85
Figure 111.9. Crystals of thSa. The center crystal is the one the diffraction data was
collected from ....................................................................................... 86
Figure 111.10. Ramachandran plot of E.coli GS /ADP/ DGM /HEPPSO thSa complex
structure. Phi (degrees) is x and Psi (degrees) is y ............................................. 87
Figure 111. 11. Crystals of thSb .................................................................. 88

Figure 111.12. Ramachandran plot of E.c0li GS / ADP / glucose / HEPPSO thSb

complex structure. Phi (degrees) is x and Psi (degrees) is y ................................... 89
Figure 111.13. Ramachandran plot of E. coli GS /ADP / glucose /HEPPSO thSc complex
structure. Phi (degrees) is x and Psi (degrees) is y ............................................. 90

Figure 111.14. Ramachandran plot of E. coli GS /ADP /glucose /HEPPSO thSd complex
structure. Phi (degrees) is x and Psi (degrees) is y .............................................. 91

Figure 111.15. Crystals of ADP, oligosacchride-bound E377A ................................ 95
Figure 111.16. Ramachandran plot of E377A /ADP /oligosaccharides complex structure.
Phi (degrees) is x and Psi (degrees) is y ......................................................... 96
Figure 111.17. Crystals of ADP-bound E377A ................................................... 97

Figure 111.18. Ramachandran plot of E377A-ADP-HEPPSO complex structure. Phi
(degrees) is x and Psi (degrees) is y .............................................................. 98

xiv

Figure 111.19. Overall structure of E. coli deS. The N-terminal domain (1-241) and C-
terminal domain (250-457) are colored yellow and blue, respectively. The interdomain
peptide 242-254 and domain-spanning helix 0L18 (458-476) are colored magenta and
orange, respectively. The N-terminal loop 212-215 and helix a7 (215-220) are in green
whereas the C-terminal helix (114 (398-403) is in light blue. The mutation sites of deS
(C7S: C408S) are shown in red sticks .......................................................... 100

Figure 111.20. Interdomain interactions between Glu218 and Gly399; Thr214 and Gly399,
and Va1248 and Phe460 are shown as dotted lines in E.coli deS. Loop 212-215 and
helix or 7 (215-220) are in green whereas the C-terminal helix 0L14 (398-403) is in light -
blue. The linker peptide 242-254 and domain-spanning helix 0:18 (458-476) are colored
magenta and orange, respectively ................................................................ 101

Figure 111.21. Overall structure of thSb complex. The N-terminal domain is in pink and
C-tenninal domain is in blue. The linker peptide 242-254 and domain-spanning helix (118
(458-476) are colored magenta and orange, respectively ..................................... 103

Figure 111.22. Top: ADP (shown in yellow and in atom colors) bound in the active site of
wild type E. coli GS. Hydrogen bonds between ADP and protein are shown as broken
lines. Residues from the N-terminal domain and the C-terminal domain are colored pink

and blue, respectively. Bottom: 1.0 o contoured 2Fo-Fc electron density map of the
bound ADP in active site ......................................................................... 104

Figure 111.23. Extensive interactions between ADP and residues Arg300 and Lys305 in
the wild type E. coli GS active site. Water molecules are presented as red spheres ...... 106

Figure 111.24. Helix a1(17-32) (cyan) and al3(382-389) (purple) point to the ADP
phosphate group. The NH-ends of both helices are presented as yellow ribbons. . . . . ....107

Figure 111.25. Top: glucose (shown in yellow and in atom colors) bound in the active site
of wild type E.coli GS. The nearby HEPPSO and ADP are shown as black and blue
sticks. Hydrogen bonds between glucose and protein are shown as broken lines. Residues
from the N-terminal domain and the C-terminal domain are colored pink and blue,
respectively. Bottom: 1.0 o' contoured 2Fo-Fc electron density map of the bound
glucose .............................................................................................. 109

Figure 111.26. Interactions between glucose, ADP and HEPPSO in the thSb active
site ................................................................................................... 110

Figure 111.27. Diagram of HEPPSO and its atom numbering ................................ 111

XV

Figure 111.28. Top: Interactions between HEPPSO and N-terminal residues, ADP, and
glucose. Bottom: 1.0 o contoured 2F o-Fc electron density map of the bound HEPPSO] 12

Figure 111.29. 2.5 o' Fo-Fc electron density map of thSa (2.83 A resolution) with DGM
(left), D-arabino-hex-l-enitol (middle) and thSb (2.22 A resolution) with glucose
(right). The ADP molecule is shown as lines for clarity ..................................... 114

Figure 111.30. DGM in WtGSa is in a superimposable position of glucose in thSb.
Ligands in thSb are in yellow whereas those in thSa are atom-wise colored. . . . . ...l 15

Figure 111.31. Positively charged C1 and 05 of DGM are stabilized by ADP, HEPPSO,
and Hisl6l. Interactions starting from the DGM C1 are shown as red dotted lines. ....116

Figure 111.32. Overlay of the active site of E.coli thSb and E377A. The residues and
ligand of E377A structure are colored yellow whereas residues in thSb are in green and
ligands are in magenta. Water W1, W2, and W3 are located between Arg300 and the
ADP distal phosphate oxygen in the E377A structure ........................................ 118

Figure 111.33. Overall structure of E.coli GS E377A in complex with ADP and
Oligosaccharides. ADP and Oligosaccharides are shown as sticks. The secondary structure
elements which host residues interacting with the surface-bound Oligosaccharide at Go
and Gd sites are labeled. .......................................................................... 120

Figure 111.34. 2F o-Fc electron density map contoured at lo with maltotriaose at the Ga
site. The ADP molecule is shown for reference ............................................... 120

Figure 111.35. Schematic representation showing the interactions between E. coli GS
E377A residues and the bound maltotriaose at Ga site. Observed hydrogen bonding and
aromatic stacking are depicted with regular and wide dashed lines, respectively. Van der
Waals interaction is marked with the wavy line ............................................... 121

Figure 111.36. The interaction between enzyme GS and the bound maltotriaose at the Ga
site ................................................................................................... 122

Figure 111.37. Superimposition of the oligosaccharide- binding site in E377A-ADP-
oligosaccharide (green), thS-ADP-glucose-HEPPSO (yellow), and MalP-PLP-G5
(PDB: lL6I) (blue) complexes. The HEPPSO that occupies the comparable position of
the Oligosaccharide was omitted for clarity. GS residues are labeled in black while MalP
residues are in blue ................................................................................. 125

Figure 111.38. Top: 2Fo-Fc electron density map contoured at 10' with maltose at the Gb
site. Bottom: Interaction between protein GS and the bound maltose at the Gb site. ......... 127

xvi

Figure 111.39. Top: 2Fo-Fc electron density map contoured at lo with maltopentaose at

the Go site. Bottom: 2Fo-Fc electron density map contoured at 10' with maltohexaose at
the Gd site ........................................................................................... 128

Figure 111.40. Schematic diagram of interactions between protein and the bound
maltopentaose at the Gb site. Observed hydrogen bonding and Van der Waals interaction
are depicted with dashed lines. The ionic interaction is marked with arrow ............... 129

Figure 111.41. Interaction between protein GS and the bound maltopentaose at the Gb site.
........................................................................................................ 130

Figure 111. 42. Schematic diagram of interaction between protein and the bound
maltohexaose at the Ge site. Observed hydrogen bonding and ionic interactions are
depicted with dashed lines and arrows .......................................................... 132

Figure 111.43. Interactions between protein and the bound maltohexaose at the Gc site133
Figure 111.44. Overlay of Oligosaccharides at the Ga (red, 1-3), Gb (cyan, 4-5), Gc (green,
6-10), and Gd (yellow, 11-16) sites ............................................................. 136

Figure 111.45. Comparison of the ideal glycogen helix model, glycogen phosphorylase —
bound maltopentaose (blue, 1P2B.pdb), Oligosaccharides at GS Ga (red), Gb (cyan, 4-5),
Gc (green, 6-10), and Gd (yellow, 11-16) sites ................................................ 138

Figure 111.46. Superposition of the C-terrninal domains from deS (cyan) and thSb
(yellow). Ligands bound in thSb (HEPPSO, ADP and glucose) are shown as black
sticks. Structural elements essential for catalysis are colored red and their counterparts in
the open form deS are colored blue. Residues critical for glucosyl transfer (Hisl61,
Arg300, and Lys305) are shown as red sticks ................................................. 141

Figure 111.47. Structural comparison of ADP molecule in Open form A. tumefaciens GS
(yellow, 1rzu.pdb) and closed from E.coli GS (blue). The residues from A. tumefaciens
GS are labeled in red and their carbons are colored yellow while those from E. coli G8 are
labeled in black and their carbons are colored blue. The interaction between ADP and
protein are shown as broken lines ............................................................... 142

Figure 111.48. Structural comparison of (bottom) loop 13-20 in apo-deS (yellow) and
thSb complex (red) and (top) the equivalent loop14-24 in the UDP/ imidazole /Glc-6-P
bound OtsA (green, 1g25.pdb) and UDPGlc-bound OtsA (blue, 1uqu.pdb). Gly22 in
OtsA and Gly18 in GS interacting with the phosphate oxygen are shown as sticks.
Ligands other than UDP (OtsA) and ADP (GS) are omitted for clarity .................... 144

xvii

Figure 111.49. Schematic diagram showing the cross-domain network between ligand
HEPPSO, glucose and ADP, which chiefly interacts with the N-terminal (red) and C-
terminal (blue) residues, respectively. ......................................................... 147

Figure 111.50. In apo E. coli deS (cyan, atom-wise colored), apo AtGS (1rzv.pdb, ruby),
apo PaGS (2bis.pdb, orange) and E377A-ADP-HEPPSO- complex (magenta). The side-
chain of Arg300 and its equivalents are out of the GS active site and are in contact with
Ala329 and Gly330 (Lys in PaGS). In ADP, Glc-bound thS (2qzs.pdb, green, atom-
wise colored), Oligosaccharide-bound E377A (blue), and ADP-bound AtGS (1rzu.pdb,
yellow), Arg300 and its equivalent side-chain are close to the ADP phosphate and their
interaction are shown as dotted lines (black lines for thS and red lines for AtGS
complex). Hydrogen bonds between the Arg300 side-chain and Gly330 in apo-deS,
and the Arg300 side-chain and maltotriaose in Oligosaccharide-bound E377A complex
are also shown as dotted lines. Residues are labeled according to the E.coli GS
sequence ............................................................................................. 154

Figure 111.51. Structural comparison of ADP, glucose binding site and active site residues
among E. coli GS (in yellow and atom colored, thSb), MalP (blue, 2asv.pdb), rabbit R-
GP (Fink, 1gpa.pdb), OtsA (green, 1uqu.pdb) and AGT (cyan, 1y6f.pdb). MalP Ligands,
PO4 ', ASO (1,5-anhydrosorbitol) and maltopentaose, occupy the equivalent positions of
the ADP distal phosphate group, glucose and HEPPSO in the thSb structure,
respectively. Maltotriaose in E377A complex was also shown (red sticks). Residues are
labeled according to the E.coli GS sequence. Top: the front view. Bottom: the back
view .................................................................................................. 156

Figure 111.52 Top: The Glu377 environment in the wild-type GS structure. Arrows
indicated the hydrogen bond proton donating direction. Bottom: The interaction between
the modeled Asp3 77 and neighboring residues ................................................ 159

13
Figure 111.53. NMR spectra of [1- C]-ADPGlc -GS complex and [1-13C]-ADPGlc -GS-
HEPPSO complex .................................................................................. 166

l . . .
Figure 111.54. NMR spectra of batch I [1- 3C]-ADPGlc and rts complex at different time
points ................................................................................................ 167

. . 13
Figure 111.55. NMR spectra of buffers 1n production of batch 1 and II [1- C]-ADPGlc.

The Tris and ammonium formate spectra are adopted from Spectral Database for organic
compounds (SDBS) ............................................................................... 168

Figure 111.56. Modeled potato GBSS structure based on E.coli GS structure. The region
ranging from helix 0L13 to 0L18 is colored red which is the determinant of most specific

xviii

properties of GBSS and may play a role in tuning the relative orientation of the two
domains ............................................................................................. 176

xix

LIST OF ABBREVIATIONS

ADPGlc — adenosine diphosphate glucose

AGT - oc-glucosyltransferase

Ala -— alanine

AMP — adenosine monophosphate

Arg - arginine

ASO - 1,5-anhydrosorbitol

AtGS - Agrobacterium tumefaciens glycogen synthase
ATP —- adenosine triphosphate

APS - advanced photon source

C — cysteine

Ca - the alpha carbon in the peptide
CAZY - carbohydrate active enzymes
C-terminal - carboxyl terminal

CC - correlation coefficient

CCD - charged-coupled device

Cys - cysteine

deS - E. coli double mutant (C7S;C408S) glycogen synthase
DGM - D-glucopyranosylium

DEAE - diethlyaminoethyl cellulose

DTT — dithiothreitol

E — glutamic acid
EMSA - electrophoretic mobility shift assay

Fhkl — structural factor

GBSS - granule-bound starch synthase
GH - glycoside hydrolase

Glc - glucose

Glc-l-P - glucose 1-phosphate
Glc-6-P - glucose 6-phosphate

Glu — glutamic acid

Gln - glutamine

Gly - glycine

GP - glycogen phosphorylase

GS - glycogen synthase

GST - glutathione S-transferase

GT - glycosyltransferase

GTA - N-acetylgalactosaminyltransferase

XX

GTB - galactosyltransferase

HEPES - (4-(2-hydroxyethyl)-l-piperazineethanesulfonic acid
HEPPSO - 4-(2-Hydroxyethyl) piperazine-l- (2-hydroxypropane) sulfonic acid
His - histidine

ID - identification
IMD - imidazole
IPTG - isopropyl thiogalactopyranoside

LB —- luria broth
Leu - leucine
Lys — lysine

MalP - maltodextrin phosphorylase
mm — millimeter
M.W.- molecular weight

N-terminal — the amino terminal of protein
ND - not determined
NMR — nuclear magnetic resonance

OtsA - trehalose-6-phosphate synthase

PaGS - Pyrococcus abyssi glycogen synthase
PDB - protein data bank

PEG — polyethylene glycol

Phe - phenylalanine

PLP - pyridoxal phosphate

RMSD - root mean square deviation

PMSF - phenylmethanesulphonylfluoride
Pro - proline

PSE - proximal sequence element

S — serine

SDS-PAGE - sodium dodecyl sulfate- poly acrylamide gel electrophoresis
Ser - serine

SNAPc —— small nuclear RNA activating protein complex

snRNA — small nuclear RNA

SS - starch synthase

TAF - TBP-associated factor

TB — Terrific Broth

TBP - TATA-box binding protein
TDB - thrombin digestion buffer

xxi

TEA - triethanolamine

Thr - threonine

Tris - 2-amino-2- (hydroxymethyl)-1,3-propanediol
Trp — tryptophan

Tyr - tyrosine

UDP - uridine diphosphate

UDPGlc - uridine diphosphate glucose
ug — micrograrn

ul — micro liter

Val - valine
wat - water
wk -— week

thS — wild type E. coli glycogen synthase
w/v — weight/volume

xxii

CHAPTER I

INTRODUCTION

1.1 Small Nuclear RNA Activating Protein Complex (SNAPc)

190 RcRd

1.1.1 Transcription Regulation and Transcription Factors

Gene expression is a fimdamental process in which a gene DNA sequence is
converted into a functional protein. It basically involves two steps, transcription and
translation. RNA polymerase (pol) is the primary enzyme of the first step in which a
specific RNA is made from the sequence information in the DNA genome. The resulting
RNAs participate in the second step in a variety of ways to synthesize functional proteins
and ensure the DNA genome information is faithfully and efficiently expressed(l).

Eukaryotes use three distinct RNA pols to catalyze gene transcription. RNA pol I
transcribes the ribosomal RNA (rRNA) genes; RNA pol 11 is responsible for transcribing
the messenger-RNA (mRNA) genes and the majority of the small nuclear RNA (snRNA)
genes. RNA pol 111 catalyzes the transcription of transfer-RNA (tRN A) genes and some
other snRNA genes (2,3).

To date the only genes that are known to be transcribed by both RNA pol II and pol
III are snRNA genes, which encode essential RNA components of small nuclear
ribonucleoprotein particles (snRNPs) involved in mRNA processing. Many of the
snRNAs are very abundant in cells, ranging from 100,000 to 1 million molecules per
nucleus, therefore, transcription from snRNA-type promoters represents a significant

portion of the total amount of transcription initiated by both RNA pol 11 and pol 111 (4).

Although RNA pols are the primary enzymes that catalyze transcription, they are
not able to recognize their target DNA sequences and initiate transcription on their own.
A series of Transcription Factors (TFs) are required to assemble on the promoter region
and form a preinitiation complex to recruit the specific RNA pol. The preinitaion TF
complex is constructed and maintained through protein-DNA (TF-DNA) and protein-
protein interactions (TF-TF). Some TFs bind to the promoter DNA directly. Some TFs
bind to those TFs that are in direct contact with DNA and modulate their activities (2,3).

Most eukaryotic transcription initiation machinery is complicated in terms of
promoter structure and TF5. The mRNA-type promoters are located upstream from
transcription start sites (toward the 5' end of the transcription initiation site) whereas
tRNA-type promoters reside both downstream and upstream from the transcription start
site(5).

Unlike RNA pol II mRNA-type genes whose promoter structures and transcription
factors are complex, eukaryote RNA pol 11 and pol 111 snRN A genes have a similar
promoter upstream of the transcription start site and similar TFs binding to the promoter
and directing transcription(5-7). Such distinct properties make eukaryote snRN A genes,
including those of humans, a unique model to understand principles of polymerase

selection and the mechanism of RNA pol 11 and pol 111 transcription machineries.

1.1.2 Small Nuclear RNA Promoters (PSE, TATA and DSE)

Like the promoters of most genes, snRN A promoters can be divided into two
distinct functional regions: the core promoter region and the regulatory region. Core
promoter regions contain the binding sites for the basal transcription factors that promote

the assembly of preinitaion complexes, and therefore are sufficient to direct low levels of

transcription in vitro. Regulatory regions are responsible for the recruitment of activator
or repressor proteins, which modulate the level of transcription (6,8,9).

The pol 11 and pol 111 core promoters in human snRNA genes contain a Proximal
Sequence Element (PSE) which is centered ~55 base pairs (bp) upstream of the
transcription start site (or denoted —55 considering zero for the transcription start site) and
is recognized by the same Small Nuclear RNA Activating Protein Complex (SNAPc)
(Figure 1.1)(6,10). Both pol 11 and pol 111 regulatory regions contain a Distal Sequence
Element (DSE) ~220 bp upstream of the transcription start site (-220) where a specific
protein, Oct-1, binds and enhances transcription from the core promoter(8). In addition,
the pol 111 snRN A gene core promoter contains a TATA box (TATAAAAG) at position
25 bp downstream of PSE(6). The presence of the TATA box in the human snRNA gene
promoter is the determinant of RNA pol specificity because addition or deletion of the
TATA box switches RNA pol specificity from 11 to 111 or vice versa, while exchanging
PSEs or DSEs between human RNA pol 11 and 111 snRNA gene promoters does not have

the same effect(6).

.1155! ESE—

Pol II snRNA promoter

43:: M

Pol III snRNA promoter

Figure [.1 Composition of pol 11 and pol III snRNA promoters in higher
eukaryotes. Typical eukaryote RNA pol II snRNA genes are Ul-U5 genes and
RNA pol 111 snRNA genes areU6 and 7-SK genes.

The TATA box is bound by TATA-box Binding Protein (TBP). Interestingly, TBP
is not only required for RNA pol III snRN A gene transcription from a promoter which
has the TATA box, but also RNA pol II snRN A gene transcription from a promoter that
does not have the TATA box. On the RNA pol III snRN A promoter the PSE is located at
a fixed distance from the TATA box. Mutating either PSE or TATA box switches
transcription to RNA polymerase 11(11,12). Thus PSE or PSE-binding factor is likely to
interact with TBP and to be involved in the RNA pol selection for human snRN A
promoters(6).

The DSE is located approximately 160 bp upstream from PSE along the DNA
chain, but in vivo the DSE and the PSE are brought into close proximity in space by a
positioned nucleosome that resides between them and wraps about 140 bp DNA(5,9). The
DSE contains a conserved octamer sequence (ATXXXXAT) (X is a residue not
conserved) and the protein specifically binding to DSE through that octamer motif is
called Oct-1. Oct-1 and the PSE binding factor SNAPc interact with each other and lead
to cooperative binding (5,9,13).

1.1.3 Small Nuclear RNA Activating Protein Complex (SNAPc)
In 1991 , Human snRNA PSE binding factor, SNAPc, was discovered in a nuclear

HeLa cell extract. HeLa is a human cancer cell line commonly used to obtain
biomaterials of human origin for its extremely rapid proliferation ability. SNAPc is a
five subunit protein complex consisting of SNAP190 (190 kDa), SNAP50 (50 kDa),
SNAP45 (45 kDa), SNAP43 (43 kDa), and SNAP19 (19 kDa) (14). The same SNAPc
binds to the PSE of both human U1 and U6- type snRNA gene promoters and is involved

in human snRNA transcription by pol 11 and pol III. Deletion of any subunit results in

reduced transcriptional activity, suggesting that each subunit plays a part in pol II and pol
111 snRNA gene transcription(lS-l 8).

The architecture of SNAPc is shown in Figure 1.2. SNAP190 is the largest subunit,
serving as the backbone of the SNAPc complex. SNAP19 and SNAP45 associate toward
the N- and C- terminus of SNAP190, respectively. SNAP43 associates with the same
region of SNAP190 as SNAP19. SNAP50 joins the complex by associating with
SNAP43 (6,9,19).

OOOH

   
   

Figure [.2 Architecture of SNAPc complex (adopted from

Hernandez et al(l9)).

 

UV cross-linking experiments combined with

90

"Hz

immunoprecipitation indicate that within SNAPc, SNAP50
(15) and SNAP190 (17)are both in close proximity to the
DNA. However, SNAP50 alone does not bind to PSE-containing probes(15), suggesting
that SNAP190 is indispensable for DNA-binding of SNAPc.

SNAP190 bears 1469 residues and there is a Myb-DNA DNA-binding domain on
its N-terrninus (263-503). The rSNAPc that lacks the last two-thirds of SNAPc is still
able to bind DNA and initiate transcription (9). In contrast, the rSNAPc without the N-
terminus of SNAP190, where the Myb domain is located, completely loses the DNA-
binding ability (18). Thus, the N-terminal Myb domain appears to be responsible for the

DNA binding of SNAPc.

SNAP190 interacts with other SNAPc subunits at its two ends; its N-terrninus
interacts with SNAP19, SNAP43, and SNAP50 whereas its C-terminus interacts with
SNAP45(19). The area between these two interacting regions does not mediate any
subunit-subunit interactions and is solely used to interact with DNA (9)and other

transcription factors, such as TBP(20) and Oct-1(18) (Figure 1.3).

 

 

 

 

 

 

 

 

 

 

 

 

Myb
SNAP19/43I50 DNA-Pinding Oct-1 SNAP45
interaction omam interaction interaction
184133 263 503 869 912 1281 1393 1469
TBP
interaction

Figure 1.3 The composition of SNAP190.

The interaction of SNAPc with TBP was first indicated by the observation that a
sub-stoichiometric but detectable level of TBP copurified with SNAPc from human HeLa
cells, which suggests their association in vivo (14). Electromobility shift assay (EMSA)
experiments show that TBP enforces the DNA binding of SNAPc and vice versa (20).
Such cooperative binding between TBP and SNAPc not only facilitates the SNAPc
binding to PSE, but also, more importantly, ensures that TBP binds specifically to the
promoter region. The DNA—binding of TBP is not as specific as its name might indicate
and its dissociation from DNA is slow(21). The protein-protein interaction between

SNAPc and TBP prevents TBP from binding irrelevant A/T rich sequences. TBP consists

of a highly conserved C-terrninal DNA binding domain and an unconserved N-terminal
regulation domain(22). SNAP190 directly interacts with the DNA-binding domain of
TBP. Truncated SNAP190 containing only the Myb DNA binding domain is sufficient
for TBP box in human U6 snRN A promoters(20).

The initiation of snRN A pol II transcription requires the assembly of
transcriptional factors TF 11A, -B, -D, -E, -F and —H on the core promoter(2,23,24). In the
most general case the orderly process of transcription initiation begins with TF 11D
recognizing and binding tightly to the TATA box(2). The TFIID-TATA box complex
directs accretion of the remaining transcription factors to the human snRN A pol 11
promoter. TBP and TBP-associated factors, de1 and Brf2, constitute the TF 111B
complex and nucleated preinitaion complex assembly on the human snRNA pol III
genes(25-27). Mutating either TATA box or PSE which is located at a fixed distance
from the TATA box in the RNA pol III snRN A promoter switches transcription to RNA
polymerase 11(11,12). This suggests that the TATA box, together with the PSE,
determines the assembly of a pol III-specific pre-initiation complex. Considering the
cooperative binding between TBP and the PSE-binding factor, SNAPc, the interaction
between the TBP C-terminal DNA binding domain and the SNAP190 Myb DNA binding
domain may play a determinant role in assembling an RNA polymerase type-specific
preinitiation complex.

1.1.4 SNAP19ORcRd and Myb Domain
The SNAP190 Myb DNA binding domain is located at the N-terminus of

SNAP190 and is important for SNAPc DNA binding(9). The recombinant SNAPc
(rSNAPc) lacking the N-terminal SNAP190 cannot bind the PSE(9). The SNAP190 Myb

DNA-binding domain also interacts with the DNA-binding domain of TBP and

stimulates its recruitment to the neighboring TATA box present in human U6 snRN A
promoters(20). The preinitaion complex made of SNAPc and TBP then anchors RNA pol
III to the promoters to initiate transcription(20).

The Myb domain was originally identified and defined in MYB proteins(28). The
SNAP190 Myb domain is an unusual Myb domain in that it is not in a MYB protein and
it has four and half repeats whereas the typical plant and yeast Myb domains have two
repeats and typical animal Myb domains have three repeats (28,29).

MYB proteins are a group of transcription factors encoded by various myb genes
and exist widely in organisms from fungi, yeast, plants, insects and humans(29). Three
types of MYB protein, c-Myb, A-Myb, and B-Myb, are expressed in higher vertebrates
and they have completely different biological roles as suggested by microarray
experiments(30,31). In human, the c-Myb protein is a DNA binding transcriptional
activator that induces transcription of a group of target genes and regulates both
proliferation and apoptosis of hematopoietic cells(30). c-Myb is the most studied MYB
protein because it is widely involved in the regulation of secondary metabolism, cellular
morphogenesis, cell cycle, development, signal transduction, and disease resistance
(32,33).Tlre DNA binding domain of the c-Myb protein is located at the N-terminus and
specifically binds the DNA sequence 5’-AACNG-3’. This DNA binding domain is well
conserved among Myb families (c-Myb, A-Myb, and B-Myb) and between species from
Drosophila to human(28). This conserved DNA binding domain is a characteristic of the
Myb protein and so is called the“ Myb domain”.

The animal c-Myb domain is a stretch of 155 amino acids containing three

imperfect tandem repeats, R1, R2, and R3, each consisting of 51-52 amino acids) (34).
The function of R1 is unknown. R2 and R3 together are responsible for the recognition of
the specific DNA sequence but neither can interact with DNA separately (Figure 1.4)
(35). Most plant and yeast c-Myb domains do not have R1 repeats and only have R2 and
R3 repeats (32,33). Nuclear magnetic resonance (N MR) studies of R2R3 repeats
complexed with DNA revealed that both R2 and R3 contain three helices and the third
helix of each is the recognition helix (36). The later X-ray structural studies of R1R2R3
repeats complexed with DNA and the enhancer binding protein is consistent with that
NMR result and clearly shows that R2 and R3 are closely packed and intercalate in the
major groove of DNA (Figure 1.4)(3 7). The two recognition helices contact each other
directly and bind to the specific base sequence cooperatively (37). In particular, the
conserved Myb residues such as the K128 in the R2 repeat, and the N179, K182, and
N183 in the R3 repeat, interact with the start AAC and the fifth G residue and. determine

the base pair recognition (Figure 1.5).

 

Enhancer binding protein

Figure 1.4 Ternary human c-Myb protein-enhancer binding protein-DNA
complex (1H88.pdb). The Myb protein (39-190) is shown as cartoon and
three repeats R1, R2, and R3 are in green, yellow, and blue, respectively.

 

Figure 1.5 The close-up view of the interaction between human c-Myb protein repeats
R2, R3 (each constituting three helices H1, H2, and H3) and DNA (containing
characteristic Myb-binding sequence 5’-AACNG-3’) in the ternary protein-DNA
complex (1H88.pdb). The conserved tryptophan residues are shown as magenta sticks.

All Myb -domain repeats contain three tryptophan residues that are highly
conserved and regularly spaced 18 or 19 amino acids apart (29,38-40). They participate
in the hydrophobic core that stabilizes the three-helix structure of Myb domain
repeats(29). Based on these characteristic tryptophan residues, sequence analysis has
identified a few Myb domains in non-myb proteins, including the SNAP190 Myb
domain(9).

The SNAP190 Myb region encompasses 241 amino acids (Trp263 to Gln503) and
constitutes four complete repeats (RaRbRcRd) and a half repeat (Rh, consisting of the
second and third helices)(18). The sequence alignment of the SNAP190 Myb region and
human c-Myb repeats shows considerable similarity between SNAP190 RaRbRcRd
repeats and the R2R3 repeats in the c-Myb domain, especially SNAP190 RcRd (Table
1.1). SNAP190 Rc repeat is 42% identical to SNAP190 Rd repeat(41). The SNAP190 Rc
repeat is 38 % and 23% identical to the Myb R2 and R3 repeats, respectively. The Rd

repeat is 38 % and 30 % identical to the Myb R2 and R3 repeats, respectively (18).

ll

Ami 28 .NE :5 325:? 38258 Song
8 385.2: 08 £996 £26 58:: 05 5 30:0: vocmou-__o>> 35 882.898 new Eon 5 8a mos—each wSBSBE

-<ZQ Hgtoafi~ ..mEmEow 332.0 .-m .-< 92:5: 28 59:8 9A2 ooE<Zm mo Eoficwzm 8:268 2C. fin «Bah.

I.m.Q...3E..UO.Q.mIIOmA..¢H..31I...U.&.>IIA.H.m.Qmm.E3..m.q mDmcmmCOU

 

mma >meZHmZBmZMH<ZDHmltwmqqxflHm§31ImzwqumdltOMHHmDmMMIBRMHMK> Nvfi mm Q>ZIU
vma >xmxHemz3mzu><ZQBmllwmQE¥¢HmflKIImzwq>mmflllmuHHmQMMMIfikomMM> MMH mm D>Zlm
mma >mmmEBmZZEZMHmZQBmIIOmAAX¢HmflSIImzoqmmmdllmwHHQOMMIB3mmmx> hma mm Q>2I¢

Hva ImmquzmkmmmUOfimettomqmm¢H>m3limmmUMKO>IIqu>mOQmmMH3mmeQ om mm Q>ZIU
NMH ImmquzmkmmmoofiwqmlIwMAmMQHAB31IOKBUM¥¥>IlamH>¥Oommxfi3mox>q Hm Nm D>Zlm
mma ImmZAIZEZEMMUOMDHmllwxqmm<HAm3uthmowxo>llqmH>mOQmmxB3mOMHA mm mm D>EI<

mm ImmZQ>MORE$OUO>QHmIIZmAMZ¢H>M3]IDQBozom>iIquqmmommmfikmequ mm Hm Q>EIU
om IQmZA>mQZENOUOOQEmItzmmmm¢AmM31IQOOUm0m>IIqmmqomommzbzx>mox mm Hm Q>Z|m
vm ImmZQ>K03EmOUOmQmmtlzoqmdeABKTIDQHom0m>lIAXXAMQOMQ¢H3M>¢ZZ mm Hm Q>EI¢

mom oxmozsz3XmAUOmommiummqmmaHmazrtmw>owmmHunquqommquzzmoqu Hmv om omfimazm
omv -mmmqmmqwmomoomomm-rom>mmmprzrroomowxm>u[aoqquommmazwwqu mam om omfimazm
mam uwmaquezm»HAOZmomuuomzww>HmmMmHmmo>mzmo>qoeqzmommmtezmxqu «am pm omammzm
mam unamzmoomxoqoomamm-meoqmm<onzrumqmom<<¢uuHaoqmmmmmmmzmomzH mam mm omflmazm
Ham 1mmmmmzozammHmmmmmmomaszmemz. mom em omamazm

12

Previous results have suggested that the Myb domain of SNAP190 is vital for PSE
binding by SNAP190 and the entire SNAPc complex(18). EMSA results reported by Mee
Wa Wong et a] in 1998 show that a truncated SNAP190 peptide containing only the
RcRd repeats (SNAP19ORcRd) can bind to the wild-type but not to the mutant PSE
(sequences shown in Table 1.2), indicating that the SNAP190 Myb domain Rc and Rd
repeats are required and sufficient for PSE binding(l8). In 1999, The same lab found that
the binding of SNAPc containing SNAP190ARcRd (lacking the RcRd) was reduced >90
fold as compared with the wild-type SNAPc whereas the SNAPc containing SNAP190
ARhRaRb (lacking the RhRaRb) bound nearly as efficiently to the PSE as wild-type
SNAPc (9). Taken together, SNAP190 RcRd was suggested to be the minimal and
critical DNA binding element that enables efficient and specific binding of SNAPc to the

core promoter PSE.

Table I2 The sequences of the PSEs present in the wild-type and mutant probes.
Uppercase letters correspond to the PSE sequence from the mouse U6 promoter,
with red-colored characters corresponding to mutations.

 

wild-type mouse U6 ggatccgaaacTCACCCTAACTGTAAAGTaattgtgtttctt
mutant mouse U6 ggatccgaaacTCCCACTACCGGTCCAGTaattgtgtttctt

 

The specific DNA binding ability and the sequence similarity (especially the
conserved Trp residues) between SNAP190 RcRd and the c-Myb domain functional
R2R3 repeats suggest that they share similar tri-helical structures and a similar DNA
recognition manner. Indeed, the consensus AACNG sequence recognized by the c-Myb

domains is in agreement with the AACTG sequence of the mouse U6 PSE, one of the

13

highest affinity PSEs for SNAPc. However, the residues critical for DNA sequence
recognition of the c-Myb domain are not present in the SNAP190 RcRd (in red in Table
1.1), suggesting that repeats Rc and Rd may recognize a different specific sequence and
employ a different DNA-identification mechanism. If that is the case, the match between
c-Myb consensus AACNG and the mouse U6 PSE sequence AACTG would be
incidental. However, this hypothesis needs to be verified by an NMR or X—ray structure

of the SNAP190cd /DNA complex.

1.1.5 Objectives

The specific binding of SNAPc to the PSE is the primary step in the assembly of
the preinitaion transcription complex on human snRN A promoters, which is critical for
RNA polymerase recruitment. Previous studies suggested that the SNAP190 Myb
domain, particularly the RcRd repeats, is responsible for PSE binding of SNAPc.

SNAPc and TBP interact with each other and cooperatively bind to their target
DNA sequence PSE and TATA box. SNAP190Myb domain RcRd repeats mediate the
SNAPc interaction with TBP and are sufficient for TBP recruitment to the TATA box.
TBP bends DNA when it binds to DNA. It is of interest to know how SNAP19ORcRd
interacts with TBP and whether a conformational change would result from their
interaction. Perhaps it is relevant to the specific RNA pol recruitment.

Single crystal X-ray diffraction was used in the hope to obtain structural
information of the SNAP190 RcRd peptide alone, various SNAP19ORcRd /PSE
complexes, and SNAP190 RcRd/ TBP/ DNA complexes. The purpose of structural
studies of SNAP190 RcRd and its PSE complexes was to reveal the DNA binding and

recognition pattern of SNAP19ORcRd, which has been suggested to account for the

14

specific PSE binding of SNAPc. The goal of structural studies of the
SNAP190RcRd/TBP/DNA complex was to map out the protein-protein interactions
(SNAP19ORcRd-TBP) and the protein-DNA interactions (SNAP19ORcRd-PSE, TBP-
TATA box) in the preinitiation complex of human RNA-pol 111 type snRN A genes,

which is sufficient to recruit RNA pol 111 onto promoters.

1.2 Glycogen Synthase
1.2.1 Glycogen

Glycogen is the main carbon source and energy storage molecule in bacteria and

eukaryotic organisms. Glycogen molecules on average weigh 106-107 daltons,

accounting for 5,550-50,000 glucose residues (42). Ninety percent of glucosyl units are
linked by 0t -1, 4 glycosidic bonds and 10 % are linked by OH, 6 glycosidic bonds. This
huge glucose polymer constitutes an efficient way to store energy, glucose, by greatly
reducing the large intracellular osmotic pressure that would result from its storage in
monomeric form. In bacteria, glycogen is consumed rapidly when organisms are exposed
to unfavorable conditions, and those rich in glycogen usually live longer than those
without this energy reserve (43). The glycogen stored in the human body is enough to
provide sugar to blood for 24-36 hours during fasting.

Glycogen metabolism is a major topic in general biochemistry. While catabolism
of these polymers is much better known from a structure/function point of view (the 3D-
structure of glycogen phosphorylase is one of the major milestones in protein

crystallography) (44-46), biosynthesis has lagged behind.

15

Glycogen phosphorylases catalyze the breakdown of glycogen to glucose-1-
phosphate, which enters glycolysis to fulfill the energetic requirements of the
organism(47,48). For a long time, glycogen biosynthesis was considered a simple
reversed phosphorylase degradation reaction. In the 19503 Leloir and coworkers (49)
discovered that glycogen biosynthesis is an independent process and nucleoside
diphosphate sugars are the starting material rather than glucose-l-phosphate. Later, three
individual enzymes were revealed to participate in bacterial glycogen biosynthesis. In
E. coli, they are ADPGlc —pyrophosphorylase EC 2.7.7.27, glycogen synthase
EC.2.4.].21, and branching enzyme EC 2.4.1.18(50,51). ADPGlc —pyrophosphorylase
produces the “activated sugar” donor ADPGlc from ATP and glucose-l-phosphate (eq.1).
Glycogen synthase then transfers the glucosyl unit from ADPGlc to elongate the a-l ,4
glucan chain (Eq.2). In the third step, the branching enzyme (BE) creates branching
points through 1,6-linkage for every 8-12 a-l,4 linked glucose residues (Eq.3). Although
in vitro the first reaction is freely reversible, in vivo the hydrolysis of inorganic
pyrophosphate by inorganic pyrophosphatase and the use of ADPGlc in the second step

makes the first step reaction practically irreversible in the direction of ADP-Glc synthesis

(50-52).
ATP + Glc -1 — phosphate c> ADP-Glc + inorganic pyrophosphate Eq.1
ADP-Glc + a-l, 4-glucan :> ADP + a-l, 4-glucosyl-a-1,4-glucan Eq.2

a-l, 4-glucosyl-oc—l ,4-glucan :> a-l , 6-branched- or-l, 4-glucan polymer Eq.3

Studies on bacterial mutants, either with a deficit or overproduction of glycogen,

have demonstrated that the levels of the polyglucan in these organisms correlates with

16

their levels of ADPGlc pyrophosphorylase and glycogen synthase activity(51). Recently,
the structure of ADPGlc —pyrophosphorylase from potato tuber (53) and structures of
glycogen synthase from A grobacterium tumefaciens (54)and Pyrococcus abyssi (55)
were solved. However, the reported GS structures, which either have no ligand or only
ADP bound in the active site, display the catalytically inactive, open forrrr. The lack of 3
GS structure of the active form, especially one with the entire substrate/ product present
in the active site constitutes a major obstacle to fully understanding glycogen
metabolism, particularly the structure/function relationship of GS and the catalytic

mechanism GS employs.

1.2.2 Starch

While animal and bacterial cells store glucose in the form of glycogen, plants use
starch, which is chemically very similar but contains fewer branches. Starch constitutes
most of the dry matter obtained from crop plants, which is the primary source of the
caloric intake in the diet of humans. Starch is also becoming increasingly important as a
renewable industrial biomaterial in today’s environmentally aware society (56,57).

Unlike glycogen, essentially homologous and water-soluble, starch is an insoluble
semicrystalline material made of two distinct polysaccharide fractions: amylopectin and
amylose. Normal wheat starch usually consists of 20-30 % amylose and 70-80 %
amylopectin(58). The major fraction amylopectin is composed of or-] ,4-linked chains that
are clustered together by 5 % a-l, 6 linkages. Amylose is composed of longer chains
with less than 1 % a-l, 6 linkage(59,60).

The physical and chemical properties of starch are chiefly determined by the

relative amounts of amylose and amylopectin, which is directly related to the specific

17

functionality and has great impact on their industrial applications(61). For example, high
amylose starches are more transparent, flexible, of higher tensile strength and tend to
gelatinize more quickly than regular starches, so it is widely used as gelling agents in
food processes and photographic film production(6l). High amylose starches help create
crispness and hamper the penetration of cooking oils in flied foods, which leads to a
decrease in fat intake by consumers. Starches with high amy10pectin levels are broadly
used in the paper and adhesive industries because of their binding and bonding
properties(56). Since the physicochemical properties of starch are essentially based on the
degree of branching and /or polymerization, the modification of starch biosynthetic
enzymes may enable us to adjust the ratio of amylopectin and amylose as well as to
produce novel starches with improved functionality. The desirable products would be
polymers with features that are intermediate between current amylopectin and amylose,
or more highly branched than amylopectin or have a higher molecular weight than
amylose(56,57,6l).

Like glycogen in bacteria, plant starch synthesis also proceeds via these three steps.
1) to produce ADPGlucose, 2) to transfer glucose from ADPGlucose to a glucan chain, 3)
to create branches(50,51). The second step, the elongation of the glucan chain, is
catalyzed by starch synthase. To date, at least four isoforms of starch synthase (SS) have
been found in plants, granule-bound SS 1 (GBSSI), SS1, SSII, and SSIII (50). While the
precise role of soluble isoforms SS1, SSII and SSIII is less clear, GBSSI has been shown
to be required for the production of the amylose component of starch(62,63).

Bacterial glycogen synthase (~ 450 aa) and plant starch synthase (~ 770-1100 aa)

share an overall 29 -34 % sequence identity. Both of them belong to the GT—35 family

18

(see section 1.2.3) based on thread analysis while a similar three-dimensional structural
organization is suggested. As the residues that were identified to be critical to SS activity
by site-directed mutagenesis and kinetic studies were also found important for GS

activity, a very similar catalytic mechanism is believed to govern both GS and SS.

1.2.3 Glycosyltransferase

A huge amount of sugar-attached biomolecules are present in almost all biological
kingdoms and play roles ranging from energy storage, cell-wall construction, cell-cell
interaction, signaling, to host —pathogen interactions. The production of glycosylated
molecules generally relies on the sugar-attaching glycosyltransferases (GTs). As of 2007,
over 7200 sequences have been identified to be GT related (64) and this number is
expected to grow as more and more genomes are being sequenced (65). Based on
sequence homology, GTs are grouped into 87 families according to the carbohydrate-
active enzyme server CAZy (64) and bacterial GS and plant SS belong to the GT35
family.

The large number of GTs in nature is not so surprising considering the great
diversity of sugar donor and acceptor involved in glycosyltranslation. GTs can transfer a
wide variety of sugars, such as glucose, galactose, N-acetylgalactose, N-
acetylglucosamine, etc., to sugar, protein, nucleotide or lipid. The sugar donors usually
are the lipid —phosphate sugars or nucleotide-sugars where the phosphate group is a good
leaving - group and can drive the reaction in favor of synthesis, considering that most
reactions take place in aqueous solution and hydrolytic processes are usually favored.

In spite of the great number of GTs, only two canonical folds (GT-A and GT-B)

have been demonstrated from the known GT three-dimensional structures. The seeming

19

“dullness” in GT protein organization indicates that all GTs may have evolved from a
small number of progenitor sequences (66). A typical GT-A fold consists of a single
domain in which parallel B-strands are flanked on either side by a-helices. In contrast,
the GT-B fold contains two Rossmann-fold-like domains that are separated by a deep
cleft (Figure 1.6). Based on thread analysis, bacterial glycogen synthase is grouped into
GT 35 with plant starch synthase and will probably adopt the same fold. Crystallographic
structures of GS from Agrobacterium tumefaciens (54) and Pyrococcus abyssi (55)

reported in 2004 and 2006 explicitly reveal that GS is a GT-B fold enzyme (Figure 1.7).

20

 

Figure 1.6 Top: Typical GT-A fold illustrated by protein SpsA. Bottom: Structure of the
GT-B fold enzyme thB. A Mn2+ ion in thB active site is depicted as blue sphere. The

fold name is after their initial observation in the SpsA (Bacillus subtilis

glycosyltransferase(67)) and thB (DNA B-glucosyltransferase) structures (19).

21

 

N-term C-term

Figure 1.7 Cartoon presentation of the structure of GS from Agrobacterium tumefaciens
in complex with ADP (blue). The N-terminal peptide 15-21 containing the KTGGL loop
and Asp21 and the C-terminal residue Arg299 and Lys304 are colored red because they

were indicated by mutagenesis and kinetic studies to be important in GS catalysis.

1.2.4 The Retaining Glycosyltransferase Mechanism Controversy

Glycosylation (the GT catalyzed sugar-attachment) proceeds with a high degree of
stereo- and regio-selectivity. Errors during glycosylation could cause serious malfrmction
of glycosylated products, which is often linked to diseases such as neurodegeneration
diseases, diabetes, and cancer (68). Depending on whether the product anomeric
configuration is retained (or—> a.) or inverted (0t—) e) with respect to the sugar donor, GTs
are classified as retaining and inverting. It is noteworthy that the GTs’ retaining/
inversion feature is not correlated to their folds as the retaining GTs have been found in

both the GT-A and GT-B fold, as have the inverting GTs.

22

Glycogen synthase is a retaining GT enzyme. Unlike inverting GT enzymes whose
mechanism was well recognized to proceed via an SN2-like transition state (66,69), it is

unclear which mechanism the retaining enzymes employ. Both double-displacement and
SNi mechanisms (internal return) have been proposed for retaining GT enzymes in the
past, but neither of them possesses solid and persuasive evidence over the other.

The double-displacement mechanism was first suggested by direct comparison to
the retaining glycoside hydrolase (GHs), which includes branching enzyme, isoamylase,
and cyclomaltodextrin glucanotransferase (70). GHs catalyze similar steps to GTS, the
sugar dissociation and nucleophilic addition, but in GH the nucleophile is simply water,
not a sophisticated acceptor nucleophile from sugar, protein, etc. (71). The double-
displacement mechanism involves a catalytic nucleophilic attack from the B-face of the
nucleotide-sugar donor and the formation of a covalently bound glucosyl — enzyme
intermediate. Afterwards, the activated acceptor undertakes a nucleophile attack from the

a- face of sugar and the retention of the anomeric configuration is achieved (Figure 1.8).

23

Enzyme

OH
H
o ‘0/Lo
H
'o 9 HO HO HO > +
M -00 (f H
?

2P§O

AiflP

AMP

 

H

 

 

 

OI

O OH H Enzyme
o /L'o H
H 0 HO
HO ‘— HO
+ 0
OH H Ow
|

""‘P=o o..-
I lP=o

.1. T.

Figure 1.8 Proposed double-displacement mechanisms for GT retaining enzymes.
Two essential components of this mechanism, the catalytic nucleophile and glucosyl-
enzyme intermediate, are circled and framed, respectively.

In the case of the well-studied retaining glycosyl hydrolyse (GH) enzymes, the
evidence of double-displacement mechanism, a covalent bound glucosyl enzyme
intermediate, was trapped and identified by both mass spectrometric and X-ray
crystallographic characterization (66,72). Many attempts have been made to prove the
existence of such an intermediate in the GT enzyme case in the last twenty years, but
have never been successful. Recently it was reported that a galactosyl moiety covalently
bound at Asp190 was observed by electrospray mass spectrometry of retaining GT-A

enzyme LgtC Q189E (73). This piece of evidence initially appeared convincing and

24

attracted wide attention, but soon the author found that the Asp190, is 8.9 A away from
the anomeric center in either wild -type LgtC or its Q189E mutant, which is too far to
function as a catalytic nucleophile if there were not a large geometric rearrangement at
the active site. Moreover, in those GT-retaining enzymes whose structures have been
determined, including glycogen phosphorylase(7 l ,74), the first structure determined with
a GT-B fold, no convincing potential catalytic nucleophile, a prerequisite for the double-
placement mechanism, is found within an effective distance to the reaction center. In
short, conclusive evidence for the classical double-displacement mechanism in retaining

GTs has not been found.

The remaining option for the retaining GT mechanism is an SN] mechanism with
some steric barrier at the B-side of the sugar. But such a mechanism involves an
oxocarbenium cation intermediate (75). A series of solution experiments have indicated
such electron deficient species have approximately a 10''2 3 half life which is not long
enough for the sugar acceptor to undertake a directed attack (76-78). Almost out of
despair, an SNi mechanism was proposed, in particular for GP, in response to the
apparent lack of an appropriately positioned nucleophile (74,75). The SNi mechanism
features not an oxocarbenium intermediate, but a transition state with significant
oxocarbenium ion character (Figure 1.9). In this mechanism, attack of the nucleophile

occurs on the same face as the departure of the leaving group, and at nearly the same

time, thus avoiding a discrete cation intermediate.

25

 

 

 

 

 

H ow,“
I
(P 9
AMP AMP
OH OH
0
H 0H 0
- Ho
H o , 3
+ H0 ‘— brg ““0 O
0 0 '-. / HO
'0 9” H W ‘60 H 9.
""1520 '07., l OM
$1 /‘p=o
AMP Cl)
AMP

Figure 1.9 SNl-like mechanism D-glucopyranosylium ion (DGM) intermediate
is framed, which is otherwise considered a transition-state in the SNi-like
mechanism. In both cases, the positively charged DGM is stabilized by the
leaving group AMP- phosphate and the incoming nucleophile 4-OH group of

R1102?

The characteristic feature of the proposed SNi /SN1 mechanism is the transition
state/intermediate D-glucopyranosylium (DGM), and the stabilization of the
oxocarbenium cation from the sugar donor phosphate and the attacking anionic
nucleophile. Information derived from inhibitor studies has indicated the presence of
these SNi /SN] mechanistic features for GT-retaining enzymes. A series of DGM
analogues that either resembles the spZ-hybridized anomeric center and the consequential
partly planar conformation of the glucopyranosyl ring or mimics the build-up and

distribution of positive charge of the oxocarbenium ion have shown varied inhibiting

26

effects to GT-retaining enzymes (Figure 1.10). Moreover, the adduct of the DGM
analogue and phosphate has also exhibited a strong inhibiting effect on GT-retaining
enzymes, such as D-gluconic acid 1,5-lactone and phosphate group in Corynebacterium
callunae starch phosphorylase (79), nojirimycin tetrazole and phosphate in rabbit muscle
glycogen phosphorylase (80), orthovanadate (phosphate analogue) and the nucleophile/
leaving group in Schizophyllum commune trehalose phosphorylase (81). This indicates
the involvement of phosphate in the DGM intermediate / transition state in these GT-
retaining enzymes. In addition, the participation of an incoming acceptor nucleophile in
the stabilization of the DGM intermediate /transition state has been implicated in the
striking inhibition of validoxylamine A to Schizophyllum commune trehalose
phosphorylase (Ki: 1.7 uM, Figure 1.10 G). Validoxylamine A not only mimics the
positive charge build-up in DGM, but also mimics the critical ternary complex upon

binding to the enzyme-phosphate complex (82).

27

 

OH

H A OH
/~§/~ it
N
HO \=N HO
H0
HO OH
OH
B C
H H
0 0
H0 Ho 0
Ho
OH
D E
H
H
Ho ”0 OH
ouNH2 Ho L
OH H2
F G

Figure 1.10 DGM (A) and DGM mimics (B-G) that have been tested in
kinetic studies of GT-B retaining enzymes. The related kinetic data are
listed in Table 1.3.

Table 1.3 Kinetic data of DGM mimics in GT-B retaining enzymes

DGM and its mimics Label in Ki (uM) GT-B retaining enzymes
Figure 1.10

 

D -glucopyranosylium (DGM) A
Nojirimycin tetrazole B 700 Rabbit muscle GP(80)

l-deoxynorjirimycin C 1200 Schizophyllum commune
Trehalose phosphorylase(8 1 )

D-glucal D 320 Schizophyllum commune
Trehalose phosphorylase(83)

D-gluconic acid E 90 Corynebacterium callunae
1, 5 -lactone Starch phosphorylase(79)

380 E. coli GS(84)
920 Rabbit muscle GP(85)

Isofagomine F 56 Schizophyllum commune
Trehalose phosphorylase(8] )

Validoxylamine G 1.7 Schizophyllum commune
Trehalose phosphorylase(82)

 

Although the above kinetic evidence apparently supports an SNi / SN] mechanism,

there is no definitive and direct experimental evidence to verify it, especially the
oxocarbenium cation transition state /intermediate. As a result, the SNi / SN] mechanism

has not been firmly established for retaining GTs yet.

29

1.2.5 Previous Studies on Bacterial Glycogen Synthase
E. coli glycogen synthase (GS) is a 52.8 kDa protein containing 477 amino acid

residues. The first successful GS separation from cell extract was done by Jack Preiss and
his coworkers with E. coli (84). The cloning and expression of GS became available and
widely used after the nucleotide sequence encoding E. coli GS, the glgA gene, was
identified in the 19803 (86-88).

In vitro, E. coli GS is able to transfer glucose from ADPGlc to maltose, higher
Oligosaccharides, and rabbit liver glycogen with increasing efficiency, but not to glucose,
even at a concentration as high as 1.4 M, indicating that a long glucan chain is preferred
by glycogen synthase, which distinguishes GS from short-glucan chain elongation
enzymes, such as MalP and OtsA.

Extensive chemical, mutagenesis, and kinetic studies have been performed on
E. coli GS to identify the architecture of the reaction center in Preiss and Fukui’s lab.
Fukui first found that the 8 -amino group of Lys] 5 is involved in ADPGlc binding and
has an impact on GS activity(89). Lys] 5 is the first residue of the Lys-X-Gly- Gly motif,
which is conserved in all glycogen synthases including human GS (X is a residue not
conserved). The neighboring glycine residues were later found not only to affect substrate
binding through interaction with Lys] 5, but also to participate in catalysis
themselves(90).

E. coli GS has three cysteine residues, which were suggested to play two distinct
roles: Cys379 is involved in substrate -binding and catalysis, whereas Cys7 and Cys408
are associated with glycogen binding (91,92). During their investigation with Cys379,
Preiss’s group found that the residue Glu3 77 is a critical catalytic residue because

mutating it to Ala causes a 10, 000-fold decrease in enzyme activity(92). The impact of

30

Glu377 to E. coli GS stems mainly from its negative charge as a much smaller effect (57
fold decrease in enzyme activity) was observed when Glu3 77 is mutated to a similar,
negatively charged residue aspartate. Mutagenesis and kinetic studies have also found
Arg300, Lys305, Hisl61 and Asp] 37 very important for GS activity (93). Interestingly,
all these catalysis-related residues are highly conserved in GSs, SSS, and other GT-B
retaining enzymes, such as OtsA, MalP and GP, indicating they all share a similar active
site (Figure 1.11).

OtsA
UDPGlc + Ot-Glucose-6-P \ : Trehalose-6-phosphate

 

M IP
MOS(glucosyl)n+ pi -—a—-> MOS(glucosyl)n_1+glucose-1-p

PLP

 

, GP
G1ycogen(glucosyl)n+ P1 PLP = Glycogen(glucosyl)n_1 +glucose-1-P

OtsA: Trehalose-6-phosphate synthase
MOS: Malto-oligosaccharides

MalP: Maltodextrin phosphorylase
GP: Glycogen phosphorylase

PLP: Pyridoxal phosphate

Pi: Inorganic phosphate

Figure 1.11 Reactions catalyzed by GT-B retaining enzyme OtsA, MalP, and GP.

3]

Although several residues have been implicated in the substrate binding and GS
catalytic performance, a full understanding of substrate specificity and the catalytic
mechanism requires knowledge of the three —dimensional structure of the enzyme-
substrate complex, especially the structure of the active form. In 2004, the first structure
of GS from Agrobacterium tumefaciens was reported (54). It displays a typical GT-B fold
and a wide cleft between two Rossmann fold domains. Since all available GT-B retaining
enzyme structures show a strikingly identical active site, structural superimposition was
performed with the AtGS structure and structures of OtsA, MalP, and GP. It reveals most
of the critical catalytic residues are located in the interdomain cleft, but a huge
displacement exists between those residues on the unaligned C-terrninal domain.
Therefore, a domain-domain closure movement from the observed AtGS organization
was suggested to bring catalysis residues together and form a competent catalytic center
in the interdomain cleft. Two years later, an apo- Pyrococcus abyssi GS structure
demonstrated a similar open, inactive form of GS by which again only limited details
regarding the active site were provided (55). Apparently, the essential features of the GS
active site have not been revealed by currently available structures. The active, closed
form GS structure is in great demand to elucidate the detailed GS active site architecture

and its catalytic mechanism.

1.2.6 Significance of E.coli GS Structural Studies

Glycogen and starch are the major carbon and energy storage compounds in nearly
all-living organisms. The central step in glycogen/starch biosynthesis involves adding

thousands of glucose units from substrate ADPGlc to build long or-l, 4-glycogen/starch

32

backbones. Structural homologous enzymes glycogen synthase and starch synthase are
responsible for this process in bacteria, animals, and plants, respectively.

Previous biochemical studies have identified a few residues related to GS/ SS
activity. Overall GS organization revealed by the recent AtGS and PaGS structures
demonstrates a typical GT-B fold with twin-Rossmann fold domains. The ADP-bound
AtGS structure represents the first GS structure with ligand in the active site, but the
structural information regarding the sugar to be transferred (the glucose moiety of
ADPGlc) and glucan acceptor is missing. Moreover, in comparison to other GT-B
retaining enzymes, there is a substantial displacement between their active site residues
and AtGS’s, indicating the open form revealed in apo-PaGS, apo-AtGS and ADP-bound
AtGS is not a catalytically active form.

1 am interested in crystallizing the bonafide active form of GS. The atomic
resolution of the GS structure, together with ample experimental data from biochemical
studies on E. coli GS, will reveal the real active site architecture of the enzyme and
elucidate the structural basis for substrate and acceptor binding. The anticipated active
form of the GS structure will be compared to the previously reported GS structures to
confirm whether a significant domain-domain closure occurs to transform GS from an
inactive to an active conformation. The hypothesis that all GT-B retaining enzymes share
a strikingly identical active site will therefore be examined.

Crystallization trials of a number of E. coli GS complexes with various
combinations of substrate, substrate analogue and inhibitor will be initiated. The
anticipated multiple ligand-bound GS structure will provide a revealing picture of the GS

catalysis process, one step closer to nature’s sweet secrets of synthesis. With luck, the

33

structures may even offer three-dimensional insights into the process of glycosidic bond
formation of GS, or even the wider, GT-B retaining family. The most positive scenario
would be some solid piece of evidence from our GS structures that would put an end to
the ongoing mechanism debates and advance our understanding of the biosynthesis of
glycosylated molecules with retention of configuration.

Both glycogen synthase and starch synthase (SS) belong to the GT-35 family and
almost certainly adopt the same GT-B fold. GS and SS share significant sequence identity
and a number of residues conserved between GS and SS are found to be critical to both
enzyme activities. This fact means GS resembles SS a great deal, especially in terms of
their active site core structure. An improved understanding of SS is expected from an
active form GS structure and ultimately benefits the lucrative starch bioengineering field.

In animals, there is an evident link between glycogen metabolism and the
physiopathological events. In particular, important diseases affecting man like Glycogen
Storage Diseases and Type II diabetes mellitus are directly or indirectly linked to
carbohydrate metabolism. Human and yeast GS are comprised of approximately 700
residues. The additional 220 residues compared to E. coli GS (477 residues) were
suggested to form a regulating domain to control GS activity in response to
metabolites(94). Most critical catalytic residues found in E. coli GS are also conserved in
human GS, indicating that the basic catalytic center is shared between E. coli GS and
human GS. The results from our structural studies of E. coli GS will therefore shed light

on the active site of human GS, which would facilitate understanding human GS.

34

10.

1].

References

Voet, D., Voet, J.G. and Pratt, CW. (1998) Fundamentals of biochemistry, 25,
John Wiley & Sons,1nc., Chichester, UK

Patikoglou, G. and Burley, SK. (1997) Eukaryotic transcription factor-DNA
complexes Annual Review of Biophysics and Biomolecular Structure, 26, 289-3 25

Wolberger, C. (1999) Multiprotein-DNA complexes in transcriptional regulation
Annual Review of Biophysics and Biomolecular Structure, 28, 29-56

Dahlberg, IE. and Lund, E. (1998) Structure and function of major and minor
small nuclear ribonucleoprotein particles, 38-70

Schrarnm, L. and Hernandez, N. (2002) Recruitment of RNA polymerase III to its
target promoters Genes & Development, 16, 2593-2620

Hernandez, N. (200]) Small nuclear RNA genes: A model system to study
fundamental mechanism of transcription J. Biol. Chem, 276, 26733-26736

Henry, R.W., Ford, E., Mital, R., Mittal, V. and Hernandez, N. (1998) in Cold
spring harbor symposia on quantative biology, Vol. LX111. l 1 1-120

Mittal, V., Cleary, M.A., Henry, W. and Hernandez, N. (1996) The Oct-1 POU-
specific domain can stimulate small nuclear RNA gene transcription by stabilizing
the basal transcription complex SNAPc Molecular and Cellular Biology, 16,
1955-1965

Mittal, V., Ma, B. and Hernandez, N. (1999) SNAPc: A core promoter factor with
a built-in DNA-binding damper that is deactivated by the Oct-l POU domain
Genes & Development, 13, 1807-1821

Hernandez, N. (1992) in Transcriptional regulation (McKnight and Yamamoto),
Vol. 11. 290, Cold spring harbor laboratory press, Plainview

Das, G., Hinkley, CS. and Henry, W. (1995) Basal promoter elements as a
selective determinant of transcriptional activator function. Nature, 374, 657-660

35

12.

13.

14.

15.

16.

17.

18.

19.

20.

Lobo, SM. and Hernandez, N. (1989) A 7 bp mutation converts a human RNA
polymerase 11 snRN A promoter into an RNA polymerase III promoter. Cell, 58,
55-67

Zhao, X., Pendergrast, PS. and Hernandez, N. (2001) A positioned nucleosome
on the human U6 promoter allows recruitment of SNAPc by the Oct-1 POU
domain. Molecular Cell, 7, 539-549

Henry, R.W., Sadowski, C.L., Kobayashi, R. and Hernandez, N. (1995) A TBP-
TAF complex required for transcription of human snRNA genes by RNA
polymerases 11 and 111 Nature, 374, 653-657

Henry, R.W., Ma, B., Sadowski, C.L., Kobayashi, R. and Hernandez, N. (1996)
Cloning and characterization of SNAP50, a subunit of the snRNA-activating
protein complex SNAPc. EMBOJ., 15, 7129-7136

Sadowski, C.L., Henry, R.W., Kobayashi, R. and Hernandez, N. (1996) The
SNAP45 subunit of the small nuclear RNA (snRN A) activating protein complex
is required for RNA polymerase 11 and 111 snRNA gene transcription and interacts
with the TATA box binding protein Proc. Natl. Acad. Sci. U. S. A. , 93, 4289-
4293

Yoon, J.B., Murphy, S., Bai, L., Wang, Z. and Roeder, R.G. (1995) Proximal
sequence element-binding transcription factor (PTF) is a multisubunit complex
required for transcription of both RNA polymerase 11- and RNA polymerase III-

dependent small nuclear RNA genes Molecular and Cellular Biology, 15, 2019-
2027

Wong, M.W., Henry, R.W., Ma, B., Kobayashi, R., Klages, N., Matthias, P.,
Strubin, M. and Hernandez, N. (1998) The large subunit of basal transcription
factor SNAPc is a Myb domain protein that interacts with Oct-l Molecular and
Cellular Biology, 18, 368-377

Ma, B. and Hernandez, N. (2001) A map of protein-protein contacts within the
small nuclear RNA-activating protein complex SNAPc J. Biol. Chem, 276, 5027-
5035

Hinkley, C.S., Hirsch, H.A., Gu, L., Lamere, B. and Henry, R.W. (2003) The
small nuclear RNA-activating protein 190 Myb DNA binding domain stimulates
TATA box-binding protein-TATA box recognition .1. Biol. Chem, 278, 18649-
18657

36

21.

22.

23.

24.

25.

26.

27.

28.

29.

30.

3].

Coleman, RA. and Pugh, B.F. (1995) Evidence for functional binding and stable
sliding of the TATA binding protein on nonspecific DNA Journal of Biological
Chemistry, 270, 13850-13859

Davidson, 1. (2003) The genetics of TBP and TBP-related factors Trends in
Biochemical Sciences, 28, 391-398

Zawel, L. and Reinberg, D. (1995) Common themes in assembly and function of
eukaryotic transcription complexes Annual review of biochemistry, 64, 533-56]

Kuhlman, T.C., Cho, H., Reinberg, D. and Hernandez, N. (1999) The general
transcription factors 11A, 11B, 11F, and [IE are required for RNA polymerase 11

transcription from the human U1 small nuclear RNA promoter. Molecular and
cellular biology, 19, 2130-2141

Cabart, P. and Murphy, S. (2001) BRF U, a TFIIB-like factor, is directly recruited
to the TATA-box of polymerase 111 small nuclear RNA gene promoters through

its interaction with TATA-binding protein. The Journal of biological chemistry,
276, 43056-43064

Mcculloch, V., Hardin, P., Peng, W., Ruppert, J .M. and Lobo-Ruppert, SM.
(2000) Alternatively spliced hBRF variants function at different RNA polymerase
III promoters. EMBOJ, 19, 4134-4143

Willis, I.M. (2002) A universal nomenclature for subunits of the RNA polymerase
III transcription initiation factor TFIIIB Genes & development, 16, 1337-1338

Lipsick, IS. (1996) One billion years of Myb Oncogene, 13, 223-235

Ogata, K., Tahirov, TH. and Ishii, S. (2004) in Myb transcription factors: Their
role in growth, differentiation and disease (Frampton, ed), Vol. 1 1. 223-238,
Kluwer Academic Publisher

Ness, SA. (2003) Myb protein specificity: Evidence of a context-specific
transcription factor code Blood Cells, Molecules, & Diseases, 31, 192-200

Rushton, J .J ., Davis, L.M., Lei, W., Mo, X., Leutz, A. and Ness, SA. (2003)
Distinct changes in gene expression induced by A-Myb, B-Myb and c-Myb
proteins Oncogene, 22, 308-313

37

32.

33.

34.

35.

36.

37.

38.

39.

40.

4].

Martin, C. and Paz-Ares, J. (1997) Myb transcription factors in plants. Trends in
Genetics, 13, 67-73

Jin, H. and Martin, C. (1999) Multifunctionality and diversity within the plant
myb-gene family. Plant Molecular Biology, 41, 577-585

Ogata, K., Hojo, H., Aimoto, S., Nakai, T., Nakamura, H., Sarai, A., Ishii, S. and
Nishimura, Y. (1992) Solution structure of a DNA-binding unit of Myb: A helix-
tum-helix-related motif with conserved tryptophans forming a hydrophobic core.
Proc. Natl. Acad. Sci. U. S. A. , 89, 6428-6432

Sakura, H., Kanei-Ishii, C., Nagase, T., Nakagoshi, H., Gonda, T.J. and Ishii, S.
(1989) Delineation of three functional domains of the transcriptional activator
encoded by the c-myb protooncogene Proc. Natl. Acad. Sci. U. S. A. , 86, 5758-
5762

Ogata, K., Morikawa, S., Nakamura, H., Sekikawa, A., Inoue, T., Kanai, H.,
Sarai, A., Ishii, S. and Nishimura, Y. (1994) Solution structure of a specific DNA

complex of the Myb DNA-binding domain with cooperative recognition helixes.
Cell, 79, 639-648

Tahirov, T.H., Sato, K., Ichikawa-Iwata, E., Sasaki, M., Inoue-Bungo, T., Shiina,
M., Kimura, K., Takata, S., Fujikawa, A., Morii, H., Kumasaka, T., Yamamoto,
M., Ishii, S. and Ogata, K. (2002) Mechanism of c-Myb-c/ebp beta cooperation
from separated sites on a promoter. Cell, 108, 57-70

Majello, B., Kenyon, LC. and Dalla-Favera, R. (1986) Human c-myb
protooncogene: Nucleotide sequence of cDNA and organization of the genomic
locus Proc. Natl. Acad. Sci. U. S. A. , 83, 9636-9640

Slamon, D.J., Boone, T.C., Murdock, D.C., Keith, D.E., Press, M.F., Larson, RA.
and Souza, L.M. (1986) Studies of the human c-myb gene and its product in
human acute leukemias Science, 233, 347-351

Nomura, N., Takahashi, M., Matsui, M., Ishii, S., Date, T., Sasarnoto, S. and
Ishizaki, R. (1988) Isolation of human cDNA clones of myb-related genes, a-myb
and b-myb Nucleic Acids Res. , 16, 11075-11089

Tatusova, T.A. and Madden, TL. (1999) Blast 2 sequences - a new tool for
comparing protein and nucleotide sequences FEMS Microbiol Lett, 174, 247-250

38

42.

43.

44.

45.

46.

47.

48.

49.

50.

51.

52.

Geddes, R., Harvey, J .D. and Wills, PR. (1977) The molecular size and shape of
liver glycogen Biochem. J. , 163, 201—209

Strange, RE. (1968) Bacterial "glycogen" and survival Nature, 220, 606 - 607

Sprang, S.R., Goldsmith, E.J., Fletterick, R.J., Withers, SO. and Madsen, NE.
(1982) Catalytic site of glycogen phosphorylase: Structure of the T state and
specificity for or-D-glucose. Biochemistry, 21, 5364-5371

Withers, S.G., Madsen, N.B., Sprang, SR. and Fletterick, R.J. (1982) Catalytic
site of glycogen phosphorylase: Structural changes during activation and
mechanistic implications. Biochemistry, 21, 5372-5382

Mclaughlin, P..l., Stuart, D.I., Klein, H.W., Oikonomakos, N.G. and Johnson, L.N.
(1984) Substrate-cofactor interactions for glycogen phosphorylase b: A binding
study in the crystal with heptenitol and heptulose 2-phosphate. Biochemistry, 23,
5862-5873

Johnson, L.N. (1992) Glycogen phosphorylase: Control by phosphorylation and
allosteric effectors FASEB J., 6, 2274-2282

Johnson, L.N., Hu, SH. and Barford, D. (1992) Catalytic mechanism of glycogen
phosphorylase Faraday Discussions, 93, 131-142

Leloir, LP. (197]) Two decades of research on the biosynthesis of saccharides
Science, 172, 1299-1303

Preiss, J. and Sivak, M. (1999) in Comprehensive natural products chemistry
(Pinto, edition), Vol. 3. 441-495, Elsevier Science B. V., Amsterdam, Neth

Iglesias, AA. and Preiss, J. (1992) Bacterial glycogen and plant starch
biosynthesis Biochemical Education, 20, 196-203

Ballicora, M.A., Iglesias, AA. and Preiss, J. (2003) ADP-glucose
pyrophosphorylase, a regulatory enzyme for bacterial glycogen synthesis.
Microbiology and Molecular Biology Reviews, 67, 213-225

39

53.

54.

55.

56.

57.

58.

59.

60.

6].

62.

63.

Jin, X., Ballicora, M.A., Preiss, J. and Geiger, J.H. (2005) Crystal structure of
potato tuber ADP-glucose pyrophosphorylase EMBO J., 24, 694-704

Buschiazzo, A., Ugalde, J.B., Guerin, M.E., Shepard, W., Ugalde, RA. and
Alzari, RM. (2004) Crystal structure of glycogen synthase: Homologous enzymes
catalyze glycogen synthesis and degradation EMBO J., 23, 3196-3205

Cristina, H., Guinovart, J .J ., F ita, I. and Ferrer, J .C. (2006) Crystal structure of an
archaeal glycogen synthase: Insights into oligomerization and substrate binding of
eukaryotic glycogen synthases J. Biol. Chem, 281, 2923-2931

Jobling, S. (2004) Improving starch for food and industrial applications Curr.
Opin. Plant Biol, 7, 210-218

Tharanathan, RN. (2005) Starch — value addition by modification Critical
Reviews in Food Science and Nutrition, 45, 371-384

Akaoka, M., Watanabe, S., Sassa, H., Yamamori, M., Nakamura, T., Sasakuma,
T. and Hirano, H. (1997) Structural characterization of high molecular weight

starch granule-bound proteins in wheat ( T riticum aestivum L.) Journal of
Agricultural and Food Chemistry, 45, 2929-2934

Buleon, A., Colonna, P., Planchot, V. and Ball, S. (1998) Starch granules:
Structure and biosynthesis International Journal of Biological Macromolecules,
23, 85-112

Manners, DJ. (1989) Recent developments in our understanding of amylopectin
structure Carbohydrate Polymers, 11, 87-112

Slattery, J .C., Kavakli, 1.H. and Okita, W.T. (2000) Engineering starch for
increased quantity and quality Trends Plant Sci. , 291-298

Mason-Gamer, R.J., Wei], CF. and Kellogg, EA. (1998) Granule-bound starch
synthase: Structure, function, and phylogenetic utility Molecular Biology and
Evolution, 15, 1658-1673

Ball, S.G. and Morel], MK. (2003) From bacterial glycogen to starch:
Understanding the biogenesis of the plant starch granule Annual Review of Plant
Biology, 54, 207-233

40

64.

65.

66.

67.

68.

69.

70.

71.

72.

73.

74.

Http://Afmb.Cnrs-Mrs.Fr/Cazv/Fam/Acc GT.Html.

 

Coutinho, P.M., Deleury, E., Davies, GI. and Henrissat, B. (2003) An evolving
hierarchical family classification for glycosyltransferases J. Mol. Biol, 328, 307-
317

Lairson, LL. and Withers, S.G. (2004) Mechanistic analogies amongst
carbohydrate modifying enzymes Chem. Commun. (Cambridge, U. K), 20, 2243-
2248

Charnock, SJ. and Davies, GJ. (1999) Structure of the nucleotide-diphospho-
sugar transferase, SpsA from Bacillus subtilis, in native and nucleotide-
complexed forms Biochemistry, 38, 6380-6385

Patenaude, S.I., Seto, N.O.L., Borisova, S.N., Szpacenko, A., Marcus, S.L.,
Palcic, M.M. and Evans, S.V. (2002) The structural basis for specificity in human
ABO(H) blood group biosynthesis Nature Structural Biology, 9, 685 - 690

Unligil, U.M. and Rini, J .M. (2000) Glycosyltransferase structure and mechanism
Curr. Opin. Struct. Biol. , 10, 510-517

Coutinho, PM. and Henrissat, B. (1999) in Recent advances in carbohydrate
bioengineering (Gilbert, Davies, Henrissat and Svensson). 3-12, The Royal
Society of Chemistry, Cambridge

Sinnott, ML. (1990) Catalytic mechanisms of enzymatic glycosyl transfer Chem.
Rev. (Washington, DC, U. S), 90, 1171-1202

Vocadlo, D.J., Davies, G.J., Laine, R. and Withers, S.G. (2001) Catalysis by hen
egg-white lysozyme proceeds via a covalent intermediate Nature, 412, 83 5-838

Lairson, L.L., Chiu, C.P.C., Ly, H.D., He, S., Wakarchuk, W.W., Strynadka,
N.C.J. and Withers, S.G. (2004) Intermediate trapping on a mutant retaining a-

galactosyltransferase identifies an unexpected aspartate residue J. Biol. Chem,
279, 28339-28344

Klein, H.W., 1m, MJ. and Palm, D. (1986) Mechanism of the phosphorylase
reaction. Utilization of D-gluco-hept-l -enitol in the absence of primer. European
journal of biochemistry / FEBS, 157, 107-114

41

75.

76.

77.

78.

79.

80.

8].

82.

83.

Sinnott, ML. and Jencks, WP. (1980) Solvolysis of D-glucopyranosyl
derivatives in mixtures of ethanol and 2, 2, 2-trifluoroethanol J. Am. Chem. Soc,
102, 2026 - 2032

Banait, NS. and Jencks, WP. (199]) Reaction of anionic nucleophiles with or-D-

glucopyranosyl fluoride in aqueous solution through a concerted, ANDN(SN2)
mechanism J. Am. Chem. Soc., 113, 7951-7958

Banait, NS. and Jencks, WP. (1991) General-acid and general -base catalysis of
the cleavage of or-D-glucopyranosyl fluoride J. Am. Chem. Soc, 113, 7958-7963

Chiappe, C., Moro, G.L. and Munforte, P. (1997) Lifetime of the glucosyl
oxocarbenium ion and stereoselectivity in the glycosidation of phenols with l, 2-
anhydro-3, 4, 6-tri-O-methyl-a-D-glucopyranose Tetrahedron, 53, 10471-10478

Schwarz, A., Pierfederici, RM. and Nidetzky, B. (2005) Catalytic mechanism of
(rt-retaining glucosyl transfer by corynebacterium callunae starch phosphorylase:

The role of histidine-334 examined through kinetic characterization of site-
directed mutants Biochem. J., 387, 437-445

Mitchell, E.P., Withers, S.G., Errnert, P., Vasella, A.T., Garman, B.F.,
Oikonomakos, N.G. and Johnson, L.N. (1996) Ternary complex crystal structures
of glycogen phosphorylase with the transition state analog nojirimycin tetrazole
and phosphate in the T and R states. Biochemistry, 35, 7341-7355

Nidetzky, B. and Eis, C. (2001) A-retaining glucosyl transfer catalysed by
trehalose phosphorylase from schizophyllum commune: Mechanistic evidence

obtained from steady-state kinetic studies with substrate analogues and inhibitors
Biochem. J., 360, 727-736

Goedl, C., Griessler, R., Schwarz, A. and Nidetzky, B. (2006) Structure—function
relationships for schizophyllum commune trehalose phosphorylase and their
implications for the catalytic mechanism of family GT-4 glycosyltransferases
Biochem. J. , 397, 491—500

Eis, C., Watkins, M., Prohaska, T. and Nidetzky, B. (2001) Fungal trehalose
phosphorylase: Kinetic mechanism, pH-dependence of the reaction and some

structural properties of the enzyme from schizophyllum commune Biochem. J. ,
356, 757—767

42

84.

85.

86.

87.

88.

89.

90.

91.

92.

Fox, J., Kawaguchi, K., Greenberg, E. and Preiss, J. (1976) Biosynthesis of
bacterial glycogen. Purification and properties of the Escherichia coli b
ADPglucosezl, 4-or-D-glucan 4-oc-glucosyltransferase Biochemistry, 15, 849-847

Papageorgiou, A.C., Oikonomakos, N.G., Leonidas, D.D., Bemet, B., Beer, D.
and Vasella, A. (1991) The binding of D-gluconohydroximo—l , 5-lactone to
glycogen phosphorylase. Kinetic, ultracentrifugation and crystallographic studies.
Biochem. J. , 274

Kumar, A., Larsen, CE. and Preiss, J. (1986) Biosynthesis of bacterial glycogen,
primary structure of Escherichia coli ADP-glucose: or-l, 4-glucan, 4-
glucosyltransferase as deduced from the nucleotide sequence of the glgA gene J.
Biol. Chem, 261, 16256-16259

Okita, T.W., Rodriguez, R.L. and Preiss, J. (1981) Biosynthesis of bacterial
glycogen. Cloning of the glycogen biosynthetic enzyme structural genes of
Escherichia coli. J. Biol. Chem, 256, 6944-6952

Okita, T.W., Rodriguez, R.L. and Preiss, J. (1982) Isolation of Escherichia coli
structural genes coding for the glycogen biosynthetic enzymes Methods in
Enzymology, 83, 549-556

Furukawa, K., Tagaya, M., Inouye, M., Preiss, J. and Fukui, T. (1990)
Identification of lysine 15 at the active site in Escherichia coli glycogen synthase,
conservation of a Lys-X-Gly-Gly sequence in the bacterial and mammalian

enzymes J. Biol. Chem, 265, 2086-2090

Furukawa, K., Tagaya, M., Tanizawa, K. and Fukui, T. (1993) Role of the
conserved Lys-X-Gly-Gly sequence at the ADP-glucose-binding site in
Escherichia coli glycogen synthase J. Biol. Chem, 268, 23837-23 842

Holmes, E. and Preiss, J. (1982) Detection of two essential sulfhydryl residues in
Escherichia coli b glycogen synthase. Arch. Biochem. Biophys. , 216, 736-740

Yep, A., Ballicora, M.A., Sivak, MN. and Preiss, J. (2004) Identification and
characterization of a critical region in the glycogen synthase from Escherichia
coli J. Biol. Chem, 279, 8359-8367

43

93.

94.

Yep, A., Ballicora, MA. and Preiss, J. (2004) The active site of the Escherichia
coli glycogen synthase is similar to the active site of retaining GT-b
glycosyltransferases Biochem. Biophys. Res. Commun., 316, 960-966

Cid, E., Gomis, R.R., Geremia, R.A., Guinovart, J. and Ferret, J.C. (2000)
Identification of two essential glutamic acid residues in glycogen synthase J. Biol.
Chem, 275, 33614-33621

44

Chapter II

CRYSTALLIZATION AND PRELIMINARY X-RAY
DIFFRACTION ANALYSIS OF SNAP190RcRd

11.1 Experimental Procedures

11.1.1 Transformation and Overexpression of SNAP190RcRd

Human SNAP190RcRd (390-518) was overexpressed in E. coli cells as a GST-tag
fusion peptide. The peptide sequence and the predicted secondary structure are shown in
Table 11.1 and Figure 11.1, respectively. The plasmid was provided by Dr. William Henry
(department of Biochemistry and Molecular Biology, Michigan State University) (1) and
transformed into E. coli BL21-Condon Plus competent cell (Cat. No. 230245, Stratagene,
La J olla, CA.). The plasmid-bearing cells were then plated on ampicillin and
chloramphenicol containing, agar Luria Broth (LB)-filled petri dishes and grown for 12 -
16 h at 37 °C. Colonies were picked from the plate and transferred into 50 mL Terrific
Broth (TB) flasks containing 0.1 mg /mL ampicillin and 0.025 mg /mL chloramphenicol.
The cells were grown at 37 °C at 250 rpm shaking for 8-12 h before being transferred to 1
L TB flasks containing the same concentration of ampicillin and chloramphenicol for
further growth. When the cell broth optical density (O.D.) absorbance at 600 nm reached
0.6-0.8, the inducing agent IPTG was added to achieve a final concentration of 1 mM and

the cells were grown for an additional 3h at 16 °C before being collected by

centrifugation at 5000 rpm.

45

Glycerol stocks of cell cultures were made for the ease of transformation and were
used for most overexpression work. Fifty percent of autoclaved glycerol (w/v) was added
to the cell culture that had grown for 12 h to achieve a final concentration of 20 %. Cells
were then incubated with glycerol at 37 °C for l h, aliquoted, and stored at -80 °C. When
needed, flakes were scraped from the resulting frozen glycerol stock and sprayed onto

plates for colony growth.

Table ".1 The SNAP190RcRd sequence used in crystallization trials and EMSA assays

 

Rb 390 RWTKSLDPG- 398
Re 399 LKKGYWAPEEDAKLLQA VAKYGEQD WFKIREEVPG RSDAQCRDRYLRRLHFS- 450
Rd 451 LKKGRWNLKEEEQLIEL IEKYGVGH WAKIASELPH RSGSQCLSKWKIMMGKKQ 503

504 GLRRRRRRARHSVRW 518

 

Conf: lIlI-IIIIII"In!“IIIEIIIIBIIIIIIIIEIIIIE
Pred:

Pred: CCCCCCCCCCCCCCCCHHHHHHHHHHHHHHCCCCHHHHHH
AA: RWTKSLDPGLKKGYWAPEEDAKLLQAVAKYGEQDWFKIRE

399 409 419 4&9

 

 

 

Conf: iIIIIIIll-I33333lllllllllllllififllllflllli
Pred: Hams; {F I

Pred: HCCCCCCHHHHHHHHHHCCCCCCCCCCCHHHHHHHHHHHH
AA: EVPGRSDAQCRDRYLRRLHFSLKKGRWNLKEEEQLIELIE

439 449 459 469

Conf. llllllaillllllllllllllIflllllllllllllnlnlf
Pred: - ‘ - .- ~ , ,

Pred: HHCCCHHHHHHHHCCCCCHHHHHHHHHHHHHHHHHHHHHH
AA: KYGVGHWAKIASELPHRSGSQCLSKWKIMHGKKQGLRRRR

479 489 499 509

 

 

 

 

 

Conf: lull-lull
Pred:

Pred: CCCCCCCCC
AA: RRARHSVRW

Figure 1].] Predicted secondary structure of SNAP190RcRd (390-518) by method
PSIPRED(2). Helix and coil are depicted as green cylinder and black straight lines. The
confidence of predication (Conf) is indicated by the height of cyan columns. Pred and
AA stand for the predicted secondary structure and target sequence, respectively.

46

11.1.2 Purification of SNAP190 RcRd

Cell pellets were resuspended in HEMGT 250 (15 mL/L cell culture)(composition
see appendix). A protease inhibitor cocktail tablet (containing chymotrypsin, thermolysin,
papain, pancreatic extract, and trypsin) (Cat. No.1 1-836-170—001, Roche Applied
Science, Mannheim, Germany) was added to the mixture. After sonication, the lysate
was spun down for 30 min and the presence of protein in the supernatant was confirmed
by SDS-PAGE. The glutathione sepharose column (Cat. No. G4510, Sigma, St. Louis,
MO) was equilibrated with 50 mL of HEMGT 250 buffer. The supernatant was then
loaded into the column and gently shaken overnight for a thorough binding. The column
was washed thoroughly with HEMGT 250 followed by 30 mL HEMGT 100, 20 mL
TDB, and 10 mL TDB-DTT. A small sample of glutathione sepharose resin (approximate
total volume 5-8 uL) was used to run SDS-PAGE in order to check for the bound protein.
Twenty units of thrombin (Sigma) were added to the resin to cleave the GST tag and
release the protein SNAP190RcRd. After 2.5 hours of shaking, PMSF solution was added
to achieve a final concentration of 0.5 mM in the resin to inhibit the thrombin activity.
The protein without the GST tag was eluted with TDB-DTT. The SDS-PAGE gel is

shown in Figure 1.2.

47

 

 

207
129- r‘"
85
42 fi 4* _> RcRd-GST
(kDa) 32 ’ ‘ 43.7kDa
_\ , \w—b GST
28.2 kDa
7 K —> RcRd
15.5kDa

Figure 11.2 SDS-PAGE gel showing the GST-tagged RcRd peptide
(lane 3 and 4) and the RcRd peptide alone (lane 6 and 7). The GST-tag
was eluted by 100 mM reduced glutathione solution afterwards (lane 9
and 10).

The peptide obtained from the previous GST purification step was buffer-
exchanged to QA buffer (20 mM Tris-HCl pH 7.5, 5 % glycerol, 0.2 mM EDTA, and 1
mM DTT) and loaded onto a Source Q ion-exchange column (Cat. No.17-0947-01, GE,
Piscataway, NJ). Pure SNAP190RcRd was eluted within the range of 50 % - 63 %
mixture of QA and QB buffer (1M NaCl, 20 mM Tris-HCl pH 7.5, 5 % glycerol, 0.2 mM
EDTA, and 1 mM DTT) (Figure 11.3). The collected protein fractions were concentrated
to ~ 4.5 mg/mL, pooled, and quick- frozen in liquid N2 for storage at —80 °C. The SDS-

PAGE gel of SNAP190RcRd eluted from the Source-Q column is shown in Figure 11.4.

48

Fractions

5 10 15 20 25 30 35 40 45 50

100% bufferB '
1.75 ~
400.0

 

 

 

 

 

L 50% buffer B '
0.75 100.0
0.50 .

100.0

0.25 .
" m 0.00
025 . 50.0

0 30 60 90

. . Conductivity
AU Time(mrn) (mS/cm)

Figure 11.3 Diagram of SNAP190RcRd purification through Source-Q column.

MW2 3 4
207,

 

N M w‘ -I* RcRd

174:
7.

Figure 11.4 SDS-PAGE gel of SNAP190RcRd from Source-Q column. Lane 2, 3, and 4
corresponds to the fraction 27, 28, and 29, respectively.

11.1.3 Purification of DNA and Annealing

Oligonucleotides were ordered from W. M. Keck Facility (Yale university) and

were purified in an anion exchange column (Source Q-5, Pharmacia) on a Perkin Elmer

49

HPLC system with UV detection at 260 nm and 280 nm. One micromole of DNA was
loaded onto the column with buffer A (10 mM NaOH and 0.2 M NaCl) and eluted with a
shallow gradient of buffer B (10 mM NaOH and 1.0 M NaCl). Collected fractions were
approximately 50-80 mL in total. After dilution with 4 column volumes of 10 mM Tris,
the DNA solution was loaded onto a 1 mL DEAE cellulose column and the liquid and
impurity was allowed to drip off the column completely. The purified DNA was then
eluted with 2-3 mL of 1 M NaCl. Several cycles of concentration and dilution were
utilized to lower the DNA salt concentration to 50 mM for the next step, annealing.
Concentration of DNA was performed using a Centricon—3 Concentrator (Millipore
Corporate, Billerica, MA).

Equimolar amounts of complementary DNA strands were added to the annealing
buffer (50 mM NaCl and 10 mM Tris pH 7.0) and heated to their melting point (90 - 95
°C) in a water-bath for one minute. The DNA mixture was then left in the water bath and

allowed to cool to room temperature.

11.1.4 Crystallization of SNAP190 RcRd, SNAP190RcRd / DNA and

SNAP190RcRd / TBP / DNA Complexes

Prior to crystallization, SNAP190RcRd was incubated in a 1:12 molar ratio with
double-stranded DNA and in a 12111.2 molar ratio with yeast TBP and double-stranded
DNA. Table 11.2 lists all of the Oligonucleotides utilized in the crystallization trials. Yeast
TBP was provided by Dr. Bill Henry (Department of Biochemistry and Molecular
Biology, Michigan State University). Crystallization was attempted using the hanging

drop vapor diffusion method at both room temperature and 4 °C. The reservoir in the

50

crystallizing plates was about 1 mL and the primary screening solution kits (Crystal

Screenm, Crystal Screen 2"“) were purchased from Hampton Research (Aliso Viejo,

CA). The hanging drop consisted of 2 uL SNAP19ORcRd or SNAP19ORcRd complexes

and 2 uL reservoir solution.

Table ".2 Oligonucleotides used in crystallization trials

U1-18-blunt ends
U1-185ingle overhang

U1-18-double overhang

U2-18-blunt ends
U2-18-single overhang
U2-18-double overhang

U2-23—double overhang

U4B-18-blunt ends
U4B-18-single overhang

U4B-18-double overhang

U6-18-blunt ends
U6-18-single overhang
U6-18-double overhang

U6-23-double overhang

TGACCGTGTGTGTAAAGA

ACTGGCACACACATTTCT
aTGACCGTGTGTGTAAAGA

ACTGGCACACACATTTCTt
caTGACCGTGTGTGTAAAGA

ACTGGCACACACATTTCTgt

TCACCGCGACTTGAATGT

AGTGGCGCTGAACTTACA
aTCACCGCGACTTGAATGT

AGTGGCGCTGAACTTACAC
caTCACCGCGACTTGAATGT

AGTGGCGCTGAACTTACAgt
cacTCACCGCGACTTGAATGng

gAGTGGCGCTGAACTTACAccgt

TCACCTTTGCGAAATAGG

AGTGGAAACGCTTTATCC
aTCACCTTTGCGAAATAGG

AGTGGAAACGCTTTATCCt
caTCACCTTTGCGAAATAGG

AGTGGAAACGCTTTATCCgt

TTACCGTAACTTGAAAGT

AATGGCATTGAACTTTCA
aTTACCGTAACTTGAAAGT

AATGGCATTGAACTTTCAt
caTTACCGTAACTTGAAAGT

AATGGCATTGAACTTTCAgt
cacTTACCGTAACTTGAAAGTat

gAATGGCATTGAACTTTCAtagt

 

51

Table ".2 (cont'd)
H1 -18-blunt ends

H1 -18-single overhang

H1-18-double overhang

Mouse U6-21—blunt ends
B-MPSE

Mouse U6-21-single overhang
S-MPSE

Mouse U6-21-double overhang
D-MPSE

Mouse U6-22-double overhang

Mouse U6-27-double overhang A
Mouse U6-27-double overhang B

Mouse U6—32-double overhang
with human U6 flanker

Mouse U6-32-double overhang
with mouse U6 flanker

TCACCATAAACGTGAAAT

AGTGGTATTTGCACTTTA
aTCACCATAAACGTGAAAT

AGTGGTATTTGCACTTTAt
caTCACCATAAACGTGAAAT

AGTGGTATTTGCACTTTAgt

ACTCACCCTAACTGTAAAG
TGAGTGGGATTGACATTTC
aACTCACCCTAACTGTAAAG
TGAGTGGGATTGACATTTCt
caACTCACCCTAACTGTAAAG
TGAGTGGGATTGACATTTCgt
caCTCACCCTAACTGTAAAGTa
GAGTGGGATTGACATTTCAtgt
caCTCACCCTAACTGTAAAGTatttcg
GAGTGGGATTGACATTTCAtaaagcgt
caatathTCACCCTAACTGTAAAGTa
tatacGAGTGGGATTGACATTTCAtgt
caatathTCACCCTAACTGTAAAGTatttcg
tatacGAGTGGGATTGACATTTCAtaaagcgt

caggaaaCTCACCCTAACTGTAAAGTaattgt
CCtttGAGTGGGATTGACATTTCAttaacagt

 

11.1.5 Electrophoretic Mobility Shift Assay (EMSA)

To radiolabel DNA probes, 250 ng single-stranded DNA was mixed with 1 uL 10x
T4 polynucleotide kinase buffer, 2 uL [y-3ZP] ATP, and 1 uL T4 polynucleotide kinase
(Amersham Pharmacia Biotech, Piscataway, NJ). The total volume was 10 uL. The
mixture was incubated at 37 °C for 1 h. The resulting y-32P labeled DNA was purified on

5 % TGE gel and annealed with its complementary single strand DNA.

The DNA binding reactions were performed in a 20 ptl total volume. DNA binding

reaction using only rSNAPc or SNAP190RcRd were performed in a buffer containing 60

52

mM KCl, 20 mM HEPES pH 7.9, 5 mM MgC12, 0.2 mM EDTA, 10 % glycerol, 0.5 pg

of poly (dI-dC), and 0.5 ug of pUCl l9 plasmid. The reactions were incubated for 30 min
at room temperature after which 25,000 cpm of 32P-labeled DNA probe was added and
the reactions were then incubated an additional 30 min. The samples were loaded onto a 5
% nondenaturing polyacrylamide gel (acrylamide: bisacrylamide ratio 39:1). The EMSA
(Electrophoretic Mobility Shift Assay) was then run in TGE running buffer (50 mM Tris,

380 mM glycine, 2 mM EDTA) at 150 V for 3-5 h at room temperature. Reactions

containing both SNAPC/ SNAP19ORcRd and TBP were run in a buffer containing 100

mM KCl, 20 mM HEPES pH 7.9, 5 mM MgC12, 0.2 mM EDTA, 10 % glycerol, 1 mM

dithiothreitol, 0.07 % Tween 20,0.2 ug of poly (dG-dC), and 0.2 ug of pUCl 19 plasmid.
Reactions with SNAP190 RcRd and TBP also contained 0.5 uL of fetal bovine serum as a
protein carrier and 50 mM NaF but lacked KCl. The samples were loaded on a 5 %
nondenaturing polyacrylamide gel (39:1) and EMSA was run in TGEM running buffer

(50 mM Tris, 380 mM glycine, 2 mM EDTA, 5 mM MgC12) at 150 V for 3-5 h at room

temperature. The gels were dried and autoradiographed.

11.2. Results and Discussion

11.2.1. Preparation of SNAP190RcRd

The human SNAP19ORcRd plasmid was cloned into a T7-based vector for
expression in E. coli. At first, the expression was tried on ordinary E. coli BL21 (DE3)
competent cells,_but was not successful. Forced high-level expression of a gene
containing condons rarely expressed in E. coli can deplete the E. coli ‘5 internal tRNA

pools. Therefore E. coli BL21-Condor: Plus competent cells (Stratagene, La Jolla, CA),

53

which contain supplemented codons, were specifically chosen to overcome the codon
bias, which resulted in an efficient expression (800 colonies/ ug DNA).

In an effort to obtain a large amount of pure protein for crystallization screens, two
different cell growth media were tested and terrific broth (TB) was chosen over luria
broth (LB) as TB gains 4.5 mg of protein per liter of cell culture, triple the amount from
LB. A detailed recipe for TB and LB are listed in the appendix. Compared to LB, growth
media TB contains a higher amount of yeast extract, extra glycerol and potassium
phosphate salts instead of NaCl. The yeast extract provided vitamins and trace elements
for cell growth. The higher amount of yeast extract in TB allows higher protein yield per

volume. Glycerol was added to the TB as an additional carbon and energy source.
Potassium phosphate buffer were used instead of NaCl in LB to provide K+ for cellular

systems while maintaining an optimal pH that prevents cell death due to a drop in pH.
The size-exclusion chromatography of peptide SNAP19ORcRd suggests that the
molecular weight of SNAP19ORcRd in solution is about 158kDa, if not higher (Figure
11.5). The considerable aggregation state of SNAP19ORcRd is likely to lead to a large
discrepancy between the estimated and experimental pl values and may also hinder the

peptide from binding to DNA.

Figure 11.5 Size-exclusion chromatogram results
from Sephacryl S-3OO HiPrep 16/60 column. The
blue line is Bio-Rad molecular weight standards
(158 kDa bovinegamma-globulin, 44 kDa chicken
ovalbumin, 17 kDa equine myoglobin, and 1.3
kDa Vitamin B12. The red line is from
SNAP190RcRd. Experiment was performed at

 

 

 

lmL/min flow rate and the fractions were 5 mL.

54

11.2.2. Crystallization Trials

Our crystallization trials include the SNAP19ORcRd peptide alone as well as
various SNAP19ORcRd —PSE and SNAP190RcRd —PSE-TBP complexes. Unfortunately,
none of these trials has produced diffraction-quality peptide or peptide complex crystals
to date. Although some crystals were found in our crystallization trays from time to time,
later the dye test (using Izit Crystal Dye, Hampton Research, Aliso Viejo, CA) and
diffraction examination prove that they were all salt crystals from the crystallization

buffer.

11.2.3. EMSA DNA-binding Assay

The failure to obtain SNAP19ORcRd —PSE complex raised doubt about the binding
ability of SNAP19ORcRd to the PSE that was claimed in Wong’s paper (3). Therefore, a
series of EMSA SNAP19ORcRd-PSE binding assays were carried out with the mouse U6
PSE, which has the highest affinity to the SNAP complex and was used in Wong’s
EMSA experiments. Unfortunately none of our assays show that SNAP19ORcRd can
bind to the mouse PSE, nor selectively bind wild type over mutant mouse U6 PSE. One

of our EMSA gels is shown in Figure 11.6.

55

B-MPSE S-MPSE D—MPSE
l l | l

RcR d R cRd RcR d

M/r/lfl

 

1 2 345 6 789101112131415161718

Figure 11.6 An EMSA was performed with a probe containing B-MPSE (lane 1-6), S-
MPSE (lane 7-12), or D-MPSE (lane 13-18). B-MPSE, S-MPSE, and D-MPSE are
blunt ended, single overhang-, and double overhang— mouse U6-21 PSE, respectively,
and their sequences are listed in Table 11.2. In lane 1, 7, and 13, no SNAP190RcRd or
recombinant SNAPc (rSNAPc) were added to the probes. Lane 2-5, 8-11, and 14-17
contain increasing amount of SNAP190RcRd (0.1, 0.3, l, 3 ug). Lane 6, 12, and 18
contain 3 ug mSNAPc.

Mini SNAPc complex (mSNAPc) is composed of SNAP190 (1-505), SNAP50,
SNAP43, and SNAP19. As Figure 11.5 shows, the mSNAPc binds to the double-overhang
mouse U6 PSE (D—MPSE) better than the single-overhang mouse U6 PSE (S-MPSE).
The blunt-overhang mouse U6 PSE (B-MPSE) displays little binding. This result
suggests that the neighboring base pair of the PSE is involved in mSNAPc binding.
Therefore, we thought that adding flanking base pairs to the mouse U6 PSE may enhance

its affinity to the peptide SNAP19ORcRd. A series of mouse U6 PSE derivatives that

56

contain varied lengths of flanking base pairs at the 3’-, the 5’- or at both ends were used
to perform EMSA assays to identify the optimized length of DNA for SNAP19ORcRd
binding.

The EMSA gel does not show the expected results. Instead, almost identical
binding behavior was displayed for each tested oligonnucleotide at high concentrations of
SNAP19ORcRd, regardless of the oligonucleotide lengths. Figure 11.7 shows the EMSA
gel and Table 11.3 lists all of the oligonucleotide sequences used in the EMSA
experiment. mSNAPc specifically binds to probe AD (containing the human U6 promoter
whose PSE is replaced by the wild-type mouse U6 PSE and a mutated TATA box) but
not to probe BD (containing the human U6 promoter whose PSE is replaced by a mutant
mouse U6 PSE and a mutated TATA box). However, no apparent binding of mSNAPc to
the Oligonucleotides was observed, except for the smear underneath the loading position.
Such a smearing phenomenon was also observed for the lower concentrations of
SNAP190RcRd and probably is due to the relatively low ratio of labeled oligonnucleotide
to unlabeled ones. As the Oligonucleotides we tested in the EMSA do not include a
faithful flanker from the human U6 sequence (Table 11.3), the possibility exists that the
oligonucleotide that contains the mouse U6 PSE and an accurate flanker from the human

U6 sequence may bind to SNAP190RcRd.

57

 

 

12345578910 12 14 15 1820 2224 25 2830 3234 3538 40

Figure 11.7 An EMSA was performed with varied length of probe. mSNAPc was added
to the it labeled lanes. Lane 1-2, 3-4, 11-12, 23-24, and 29-30,35-36 contain increasing
amount of mSNAPc (1 and 3 pg). All probe sequences are listed in Table 11.3. Lane 7-
10, 13-16, 19-22, 25-28,31-34, and 37-40 contain increasing amount of SNAP19ORcRd

(0.1, 0.3, 1, 3 pg).

58

Table ".3 DNA sequences used in Figure ".6.

probeAD agcggataacaatttcacacaggaaacagctatgacatgattacgaattcgagctcgg

probe 80

W121

MU21

WT33

MU33

MU43

tacccggggatccgaaaCTCACCCTAACTGTAAAGTaattgtgtttcttggtTCTCGAGCCtt

gggaagctggcactggcctcgttaaattg

agcggataacaatttcacacaggaaacagctatgacatgattacgaattcgagctcgg

tacccggggatccgaaaCTCCCACTACCGGTCCAGTaattgtgtttcttggtTCTCGAGCCtt

gaggaagctggcactggcctcgtacaattg

wild type
mouse U6
double overhang

mutant
mouse U6
double overhang

wild-type

mouse U6
double overhang
12-base flanker

mutant

mouse U6
double overhang
12 -base flanker

wild-type

mouse U6
double overhang
22-base flanker

mutant

mouse U6
double overhang
22-base flanker

human U6 PSE
NCBI [giz577b71]

caCTCACCCTAACTGTAAAGT
GAGTGGGATTGACATTTCAgt

caCTCQCACTAQCQGTQQAGT
GAGQGIGATQGQCAETCAgt

caatathTCACCCTAACTGTAAAGTgattcga
tatacGAGTGGGATTGACATTTCActaagctgt

caatathTCQCACTAQCQGTQQAGTgattcga
tatacGAGgGIGATgGQCAQQTCACtaagctgt

cactatcatathTCACCCTAACTGTAAAGTgattcgatttct
gatagtatacGAGTGGGATTGACATTTCActaagctaaagagt

cactatcatathTCQCACTAQCQGTQQAGTgattcgatttct
gatagtatacGAGQGIGATgGQCAQQTCACtaagctaaagagt

251 gactatcatathTTACCGTAACTTGAAAGTatttcgatttct

Wild-type PSE in probe AD and mutant PSE in probe 80 are in underscored uppermse.
The mutant TATA box in both probes are in italic uppercaseJn other tested DNAs
as well as human U6 sequence, PSE is in uppercase and mutated bases in PSE are undedined.

59

11.3. Conclusion

The RcRd repeats of the N-terminal Myb domain of SNAP190 are required for the
SNAPc complex to bind to the PSE, as the binding of SNAPc containing
SNAP190ARcRd (lacking the RcRd) was reduced >90 fold as compared to that of the
wild-type SNAPc(4). Data reported by Wong et al. in 1998 firrther suggests that a
truncated SNAP190 peptide containing only the RcRd repeats (SNAP190RcRd) can bind
to the wild-type but not to the mutant PSE, indicating that the SNAP190 Myb domain Re
and Rd repeats are required and sufficient for PSE binding(3).

To investigate the interaction between SNAP190RcRd and the PSE, we performed
crystallization trials with a combination of peptide 390-518 (encompassing the RcRd
repeats which was used in the literature of Wong et a1, 1998(3)) and 25 different PSEs.
None of them has yielded any hits to date. A DNA-binding assay carried out on peptide
390-518 and the mouse U6 PSE, which has a high affinity to the SNAPc complex,
revealed an opposite result to that which was reported in the literature(3). SNAP190RcRd
does not favorably bind to the PSE, although the binding is slightly enhanced by the PSE
flanker base pair. Size-exclusion chromatography suggests SNAP190RcRd is present in
solution minimally as a decamer. We ascribe the observed low PSE affinity of
SNAP190RcRd to its high aggregation state, which may bury the DNA-interacting
elements into the interior and hinder SNAP19ORcRd DNA-binding. Thus, efforts to
investigate the interaction between SNAP190RcRd and the PSE in the future may need
not only the RcRd repeats and the PSE, but also structural elements that prevent

SNAP19ORcRd from aggregating.

6O

Appendix
Luria Broth (LB): to make 1 L of LB media, 15 g of tryptone, 5 g of yeast extract, and 5
g of NaCl were mixed and dissolved in 1 L of water. The solution
was autoclaved. Before use appropriate antibiotics were added.
Terrific Broth (TB): to make 1 L of TB media, 12 g of tryptone, 24 g of yeast extract,
and 4 m1 glycerol were mixed and dissolved in 900mL of water

and autoclaved. Before use 100 mL of salt solution containing 2.31
g KH2P04, and 12.54 g K21-IPO4 were added to the broth.

HEMGT 250: 25 mM Hepes pH 7.9, 2 mM EDTA, 12.5 mM MgC12, 10 % glycerol, 0.1

% Tween-20, 250 mM KC], 3 mM DTT
HEMGT 100: 25 mM Hepes pH 7.9, 2 mM EDTA, 12.5 mM MgC12, 10 % glycerol, 0.1

% Tween-20, 100 mM KCl, 3 mM DTT

TDB (thrombin digestion buffer) (10x): 200 mM Tris-HCI pH 8.4, 1.5 M NaCl, 25 mM
CaClz
TDB —DTT (10x): 200 mM Tris-HCl pH 8.4, 1.5 M NaCl, 25 mM CaC12,3 mM DTT

QA: 20 mM Tris-HCl pH 7.5, 5 % glycerol, 0.2 mM EDTA, 1 mM DTT
QB: 1 M NaCl, 20 mM Tris-HCl pH 7.5, 5 % glycerol, 0.2 mM EDTA, 1 mM DTT
5 % TGE gel (10 mL): 1.25 mL Acrylamide (39:1), 2 mL 5x TGE, 0.5 mL 50 %

glycerol, 8 pL Tetramethylethylenediamine, 80 pL 10 %
ammonium persulfate and 6.25mL H20.

5 x TGE buffer: 500 mM Tris, 3.80 M glycine, 20 mM EDTA

61

References

Hinkley, C.S., Hirsch, H.A., Gu, L., Lamere, B. and Henry, R.W. (2003) The
small nuclear RNA-activating protein 190 Myb DNA binding domain stimulates
TATA box-binding protein-TATA box recognition J. Biol. Chem, 278, 18649-
18657

Jones, D.T. (1999) Protein secondary structure prediction based on position-
specific scoring matrices J. Mol. Biol, 292, 195-202

Wong, M.W., Henry, R.W., Ma, B., Kobayashi, R., Klages, N., Matthias, P.,
Strubin, M. and Hernandez, N. (1998) The large subunit of basal transcription
factor SNAPc is a Myb domain protein that interacts with Oct-1 Molecular and
Cellular Biology, 18, 368-377

Mittal, V., Ma, B. and Hernandez, N. (1999) SNAPc: A core promoter factor with
a built-in DNA—binding damper that is deactivated by the Oct-1 POU domain
Genes & Development, 13, 1807-1821

62

Chapter III

THREE DIMENSIONAL STRUCTURES OF
Escherichia coli GLYCOGEN SYNTHASE AND ITS

COMPLEXES

111.]. Theory

111.1.1. Structure Determination from X-fray Diffraction Data

A crystal arranges huge numbers of molecules in a highly ordered and repetitive
pattern. When an X-ray beam strikes the electron clouds(1) of these molecules, the
scattered waves emitted from electrons can add up in phase and generate a measurable
signal. This phenomenon is popularly termed as X -ray diffraction where the orderly
arrayed molecules form imaginary reflecting planes and in-phase scattered waves are

considered as the product of diffractions that comply with Bragg’s law

2d sin0 = n?»

where d is the distance between reflecting planes. 8 is the angle between the incident

beam and the reflected beam, 71. is the wavelength of incident radiation. The most
commonly used X-ray wavelength for biological macromolecule structure determination

is about 1 A, compatible for atomic bond lengths (1.5-1.7 A) and atoms themselves.

63

The X-ray diffraction pattern reflects the spatial distribution of the electrons and a
three-dimensional picture of the electron density of a biological macromolecule can be
inversely derived from the collected X-ray diffraction data with the aid of Fourier
transformation. The Fourier transformation is a mathematical operation that can link the
diffraction pattern and the object diffracting the waves. Electron density at every position

xyz is presented by the following equation

p(xyz) = —I-l/-ZZZ|F(hkl)lexp[—27ri(hx+ ky+lz)+ ia(hkl)]
h k 1

The structure factor F (hkl) is a complete description of all the atoms that

conuibute to the reflection labeled hkl. F (hkl) is a wave function with frequency,

amplitude thkll and phase or (hkl). The frequency of F (hkl) is identical to the X-ray

source. The amplitude is proportional to the square root of the intensity of the reflection,
which can be measured from the recorded X -ray diffraction pattern. The phase cannot be
directly derived from the diffraction pattern and constitutes the well-known Phase
Problem in crystallography.

IH.1.2. Molecular Replacement

Molecular replacement is one of the methods used to solve the “Phase Problem” in
biological macromolecular crystallography. It takes advantage of the structure of a known
protein to calculate the electron density map of a biological entity whose X-ray
diffraction pattern has been recorded. Usually, a protein that shares >25 % sequence
identity with the target protein can serve as a decent initial search model and stands a

good chance of success in molecular replacement. In our case, the AtGS sequence is

64

43% identical to that of E. coli GS (2)and therefore was chosen as the initial model to
solve the E. coli GS structure.

The search for the correct molecule position relative to the counterpart in the
known structure involves three rotational and three transitional parameters. In order to
reduce otherwise enormous computation time, the search for these six parameters is
carried out in two separate steps. The first is the rotation search, which uses the Patterson
function to determine the correct orientation of the search model. The Patterson function
is the Fourier transform using P (hkl)2 and contains sets of peaks representing intra-
atomic vectors (3). As Patterson functions can be calculated from the amplitudes only,
the orientation of the search molecule can be found without the aid of phase information
from the unknown structure. In the correct orientation, a maximal overlap with the target
structure should be reached. The rotation search in the case of E. coli GS is all performed
by the automated Patterson search routine of the program MolRep (4).

Once the correct orientation of the model (or, B, 7) has been found, the correct
position can be determined through the translation function. The translation vector for
each atom can be split into two vectors, one (I) in the rotation axis direction and the other
(3) perpendicular to the rotation axis direction. The vector t is the same for all atoms
while the self -vector depends on the distance of each atom to the rotation axis.
Superimposing the self -vector set onto the intermolecular vectors (cross vectors) gives a

number of positions where some agreement between the vector peaks is obtained. The

structure factor calculated (Fcal) based on the solution of rotational search and

translational function will be compared to the observed ones from data (Fobs)- A properly

65

positioned model should give a relatively high correlation coefficient (C. C.) and low R-

factor.

 

2 (lFobsl2 — |F0bsl2)(chal|2 -|Fca112)
hkl

 

 

 

cc.
2 2 1/2
2 01:01.42 — lFobslz) )3 (in-ail — mall)
hkl hkl
Z “Fobsl — chal”
R = W x 100%
Z IFobsl
hkl

111.1.3. Structure Refinement

After applying the initial phase from the known structure to the diffraction data of
the protein of interest, an approximate model can be obtained, but most of time, that
model is rather coarse and needs to be refined. In fact, an R-factor of 40-50 % is
common for the model fresh from molecular replacement. Afier thorough refinements it
can be lowered to 8-25 %, depending on the data quality.

Due to thermal vibrations and static disorder, the atoms in the crystal are not
strictly fixed and therefore their atomic scattering factors deviate somewhat from the
standard value. Temperature factors (B) reflect the extent of the dynamic disorders and
these are refined in structure determination as well as three positional parameters (x, y,
2). However, the number of independent X-ray reflections of the protein alone generally
is not sufficient for solving all those parameters. Therefore, many additional

“observations” are incorporated, such as geometric restraints: bond lengths, bond angles,

66

torsion angles and van der Waals contacts. The refinement takes advantage of these
additional “observations” in a restrained refinement where the stereochemical parameters
are only allowed to vary around a standard value (from small molecule geometry bank).
On the other hand, the number of parameters to be refined can be substantially reduced if
the entire protein or domain is taken as rigid. This refinement strategy is termed “rigid
refinement”. In the E. coli GS case, a rigid-body refinement was first applied to find a best
overall- fit position. Then a great number of restrained refinements along with manual
refitting of the model to the electron density maps were carried out. The interactive
graphics program Turbo F rodo(5) was used for manual fitting and Refmac(6) in the
CCP4 program suite was used for both types of refinements. The R-factor of the final
models of E. coli GS and its complexes are in the range of 17-21 %, substantially lower
than the R-factor straight from the molecular replacement (31-51 %). Another widely
used parameter to assess the closeness between the atomic model and the observed data is
the free R value (R-free) (7). R-free is computed with a small set of randomly chosen
intensities (“test set”, usually 5-10 % of the data), which are set aside from the beginning
and not used during refinements. R-free measures how well the atomic model predicts a
subset of the measured intensities that were not included in the refinement ( I Ftest I ),
whereas R-factor measures how well the model predicts the entire data set that produced
the model ( l Fobs I ). The difference between R value and R-free value of an accurate
model should not be large, usually less than 7-10 %(7,8). All determined E. coli GS
structures display a nice agreement between R factor and free R factor (Table 111.1).
Besides R and R-free values, a Ramachandran plot is another means commonly

used to evaluate the model quality in biological macromolecular crystallography. Due to

67

steric hindrance, the mainchain of a polypeptide usually assumes preferred, energetically

favorable conformations(9). Typically for each residue, these conformations are
characterized by torsion angles, (I) (Phi) Ci_1-Ni-Cai - Ci and \y (Psi) Ni'Cai - Ci - NH],

The distribution of 4) and u! is called the Ramachandran plot which shows how well the (b
and w angles of protein residues cluster. As 4) and w angles are not usually restrained
during X-ray refinement, a Ramachandran plot provides a simple, sensitive, and
independent means for assessing the quality of a protein model (10). The Ramachandran
plot of E. coli GS structures and the related discussion is described in their individual

sections.

111.2. Experimental Procedures

111.2.]. Protein Overexpression

E. coli GS (both His-tagged and untagged) was overexpressed in E. coli BL21
(DE3) cells. Fifty milliliters of LB medium containing 50 pg/mL kanamycin was
inoculated with a single colony of E. coli strain BL21 (DE3) bacteria containing the GS
expression plasmid (pAY3 for His-tagged GS, pAYl for untagged GS)(11). Cells were
grown at 37 °C with shaking at 250 rpm for 3-4 h or until saturated. One liter of LB /

kanamycin was inoculated with 50 mL of saturated culture and grth was maintained at

37 °C with 250 rpm shaking until the O.D.600 reached 0.6-0.8. The cell culture was then

cooled down to 4 °C before induced by 1 mM IPTG. Growth continued for 12 h at room

temperature (20-22 °C) before the cells were harvested by centrifugation at 5000 rpm.

68

111.2.2. Protein purification

III.2.2.A. His-tagged GS protein

Cells were resuspended in the buffer (50 mM potassium phosphate at pH 8.0, 300
mM NaCl, 10 mM imidazole) and lysed by sonication. Cell debris was spun down by
centrifugation at 5000 rpm for 30 minutes. The supernatant was mixed with Ni-NTA
resin (Qiagen, Valencia, CA) in volume ratio 4:1. The mixture was stirred slowly at 4 °C
for 1 h to allow thorough binding before loaded onto an empty column. The impurity was
removed by the wash buffer (50 mM potassium phosphate at pH 8.0, 300 mM NaCl, 50
mM imidazole). The His-tagged GS protein was eluted with the elution buffer (50 mM
potassium phosphate at pH 8.0, 300 mM NaCl, 250 mM imidazole).

Fractions were analyzed by SDS-PAGE for homogeneity (Figure 111.1), and were
thoroughly buffer- exchanged into the storage buffer (20 mM TEA/HCl pH 8.0 and 5
mM DTT) using desalting columns (Bio-Rad Econo-Pac 10 DG) and concentrated to 6-8

mg/ mL for crystallization.

wash 611.1116

207 .....
129 . .

85¢

 

 

39

32

(kDa)

Figure 111.1 Representative His-tagged E. coli GS SDS-PAGE gel. The
most left lane is the molecular weight standard. Each elution lane

accounts for the collection of a 5 mL elution buffer fraction.

69

III.2.2.B. Untagged deS protein

Cells were resuspended in the buffer (50 mM potassium phosphate at pH 6.8, 100
mM NaCl, 5 mM DTT) and lysed by sonication. Cell debris was spun down by
centrifugation at 10,000 rpm for 30 minutes. The supernatant was then brought to the
25% saturation of ammonium sulfate and the mixture was stirred 30 minutes and
centrifuged at 10,000 rpm. Most of the impurities stay in the supernatant and was
discarded. The pellet was dissolved in buffer A (20 mM TEA/ HCl pH 7.5 and 5 mM
DTT) and was desalted by desalting column or dialysis. Protein was filtered through 0.22
pm filter and the solution was loaded onto a Source -Q column. GS protein eluted
between 250 -350 mM KCl, 20 mM TEA/ HCl pH 7.5 and 5 mM DTT (Figure 111.2). The
purified GS protein was poured onto a desalting column three times to remove 99 % salt

and concentrated to 6-8 mg/mL before crystallization.

elute elute

 

207 m...
129 . .. s

85

39“- MW 2...... _+ 52 kDa

32
(kDa)

Figure 111.2 E. coli deS SDS-PAGE. The second from left lane is the
molecular weight standard and all other lanes are elution from Source -Q

column with increasing concentration of KCl.

70

111.2.3. GS and GS Complex Crystallization

GS protein was incubated with various substrates and inhibitors for 30 minutes to
generate GS complexes. A full list of substrates and inhibitors that have been used in
crystallization trials are shown in Table 111.1. The resulting GS complex solution was
then subjected to spin-filtration using a spin filter with 0.45 pm cut-off membrane
(Microcon® 0.45 pm, Amicon bioseparations) to remove precipitates that would cause
excessive nucleation in crystallization.

Crystallization was performed using both sitting drop and hanging drop vapor
diffusion methods at room temperature and 4 °C. The reservoir in the crystallizing plates
was about 1 mL and the primary screening solution kits (Crystal Screen, Crystal Screen
2) were purchased from Hampton Research (Aliso Viejo, CA).

Once the primary crystallization condition was found, further screening of that
condition was conducted by altering buffer, precipitate, additive, and temperature to
obtain diffraction-quality crystals. Other attempts to improve crystal quality include the
micro-dialysis, micro- and macro-seeding (12). When co-crystallization of GS with
ligand did not produce suitable crystals for data collection, the apo-GS crystal was soaked
in the stabilizer solution with a high concentration of ligand (50-500 times Km or Ki).
Usually, the stabilizer solution, which is able to prevent the crystal from dissolving, is the
original reservoir solution with 5 —l 0 % higher concentration of precipitate (PEG 4000).
For the ligand- bound GS crystals, multiple data sets were collected and the one with the
best quality was used for the structure reported in the later sections except for the wild-

type GS-ADP-HEPPSO complexes.

71

 

Us v 3 m0 meow 256:3 5E 23:. m_ 6388 £80: Ho: a

 

0 Z vacuum—ocoos_w-m+mo<
OZ omomxozozwfi+mfl<
OZ omoficoqozaE+mO<
8886 as; 288 .2 3+ 2 :3 025m: .2 _ .o + 0803 .58 85 0853.555...
OZ 3335+ ma
OZ omomxonozae+0monhdom+20mQ<
$58 2 mg 8.33 23 + 0803 .58 mm; 08.55852053.
<82 889 2 3+ 2 :a 855m: 23 + x53 .58 mm;
$58 3 :5 025m: 23 + 0:63 .58 mm; 085533353655.
<82 peas 2 3+ 2 :5 025m: 23 + 0803 .58 mm;
£58 2.68 2 3+ 2 :5 025m: 23 + 0803 .58 mm; oao§§+2oaa<
< m 3 :a 025m: 23 + 80.“: .58 m; 08:53.65...
Execs-5:50
£58 .< m 38:82 23+; :5 .5 23 + 803 .58
Saga .83 2.2 2 83 + 3 :5 .5 23 + 080mm .58 m5 :25 8m .0: 89028823
€888 ow... 9.5.2 is 3 + 3 :5 a: 23 + 0.5.”: .58
sonata x83 .632 2 3 + 3 IQ a; 23 + 0.885 .58 mm> 528 m: .3 3282978820
oz 52...
< m 5628:. um... .aaaaea :58 .598 28.82.02 23:35.; 23+ 089: .58 mm; :21 8 .3 .5...
.3225
02 Ea 8X .5: 8888.32
OZ 08823232
02 33.5332
< v 2.0z 2 3 + 3 ma .5. 23 + 0:63 .58 mm; Ea 8” =2 vase:
owes; 28:82 .2 3+ 3 :a 025m: .23 + 80.“: .58
893.80.; 085.2 2 3+ E :5 025m: :3 + 0803 .58
«we; 9.02 .2 3+ 3 :5 025m: 23 + 80.“: .58 mm;
E58 3 25 025mm 3: .o + 0803 .58 mm;
88 958 88. £38852 :52 3 :a ...E 23 + 088.. .58 mm; 93 .2 EM V295...
8953.5
82 .5225 3&5 cuawm

$35 cosz=Smbu 5358 mm 2: 03m...

72

III.2.4. Diffraction Data Collection and Processing

All crystals were obtained at 4 °C and were harvested into the mother liquid with
15 % glycerol as cryoprotectant and flash-frozen before the data collection. Data were
collected at the Advance Photon Source of Argonne National Laboratories (APS)
(Argonne, IL) at various beamlines (details in Table 111.2 and 111.3). Data reduction and
scaling were performed using Denzo and Scalepack (13) either at home or at the
beamline (HKL 2000). Complete data collection and process information is shown in the

table of each individual section.

III.2.5 ”C— and 'H-“C NMR Experiment
III.2.5.A Preparation of D- [1-13C]-ADPGlc from a-D- [1-'3 C]

glucopyranosyl l-phosphate
a-D- [1-13C] glucopyranosyl l-phosphate (dicyclohexylammonium salt,

monohydrate) (MW 477.51, Cat GLC-Ol 5) and a-D- [UL-”(36] glucopyranosyl 1-

phosphate (dicyclohexylammonium salt, monohydrate, uniform labeled) (MW477.51,
Cat. GLC-074) were purchased from Omicron Biochemicals, Inc. (South Bend, IN). The
TSK-GEL DEAE-SPW column was purchased from Tosoh Biosciences LLC
(Montgomery Ville, PA).

The reaction mixtures contain 100 mM Bicine or Hepes buffer (pH 8.0), 0.2 mg/mL

bovine serum album, 7 mM MgC12, 1.5 mM ATP, 2 mM fructose 1,6-bisphosphate, 1.5

unit /mL inorganic pyrophosphatase, 0.5 mM glucose-1-phoshate(14). Reaction was
initiated by addition of the enzyme E. coli ADPGlc pyrophosphorylase and the reaction

mixture was incubated for 2 hrs at 37 °C before it was terminated by EDTA.

73

Paper chromatography was used to check the presence of product ADPGlc. Five uL of 20
mM ATP, ADP, and ADPGlc and 30 uL reaction mix were streaked on Whatrnan No.1
paper and chromatographed (descending) for 24 hrs in the ethanol — ammonium acetate
pH 7.5 solvent system (95 % ethanol /1 M ammonium acetate, pH 7.5 (5:2)). The
chromatogram was dried in the hood and the ADPGlc spot was located with UV-light.

Descending paper chromatography revealed that there are two main UV absorbing spots,
which correspond to the product ADPGlc and the unreacted ATP. To separate l3C-

labeled ADPGlc from the reaction mix, particularly from the unreacted ATP, the reaction
mix was injected into the anion exchange DEAE-SPW column afier filtered through a
0.45 mm filter. The separation was done on a DIONEX P680 pump system with gradient
between buffer A (25 mM ammonium formate, pH 7.2) and buffer B (25 mM ammonium
forrnate and 0.5 M NaCl, pH 7.2) (batch I) or buffer A (25 mM Tris, pH 7.2) and buffer B
(25 mM Tris pH 7.2 and 0.5 M NaCl, pH 7.2) (batch 11). Comparison of the HPLC
spectrum of the reaction mix and the ATP, ADPGlc, ADP and AMP standards revealed
that, besides the expected ADPGlc and ATP, there was a small amount of AMP and ADP
resulting from the degradation of ATP/ADPGlc (Figure 111.3). The second peak was
identified as ADPGlc and the fraction at that position was collected and lyophilized to

give a white solid.

74

 

 

 

 

 

 

 

 

 

 

3,000 .
03:5. ADPGlc
-. ' “m 1 ATP
2,000 f l
Abs (mAU) 3
1,000 €
j‘ AMP ADP
0.0 10.0 20.0 30.0 40.0 50.0 62.0

Time(min)

Figure 111.3 HPLC spectrum from the DEAE-SPW column. Flow rate: 0.6 mL
/min, Gradient: 2 min: 0% buffer B; 40 min: 25 % buffer B; 60 min: 100 %
buffer B. Fraction ADPGlc elutes in the range of 14 —1 8 % buffer B.

A typical adenosine derivative displays an absorption maximum at 259 nm (A239;
A260 = ~0.15 and A250:A260 = ~0.80). The spectral analysis of the [LDC] and [U L-13C6]-

ADPGlc at pH 7.0 is in agreement with that. The quantity of ADPGlc was estimated

based on the equation “g - per liter = (Absz59 "m ~Molecu1ar Weight) / (1M, average molar
absorbance index (1M =15.4 x103 L/mol” (15). The yield of [1-‘3C]—ADPGlc and [UL-
l3C6]-ADPGlc are 66 % and 87 %, respectively.

111.2.5.B. NMR experiment setting

1H NMR, 13C NMR and phase sensitive gHMQC two-dimensional spectra were

recorded at 500 MHz (Varian Unity Plus 500) or at 600 MHz (V arian Inova 600). The

75

NMR facility temperature was calibrated with pure methanol prior to experiment and all
data were collected at 4 -5 °C. Solvent D20 with reference TSP (sodium salt of 3-
(trimethylsilyl) propionic -2, 2, 3, 3- d4 acid) was purchased from Sigma-Aldrich
(Catalogue # 293040) and was used without further purification. All NMR data were

processed with the Varian machine software and the spectra images were made with

software Mestrec23.
111.2.5.C. NMR sample preparation

Wild-type E. coli GS (7-9 mg /mL) was thawed from — 80 °C and diluted to ~ 2-3
mg/mL with D20. l3C —labeled ADPGlc and HEPPSO was added in a 1:10:100 ratio to

GS, ADPGlc and HEPPSO. The GS-ADPGlc-HEPPSO adduct was then concentrated by
centrifugation until reaching its maximum concentration, approximately 9 mg /mL. The

total complex volume was ~ 0.7 mL, a minimum volume to obtain decent NMR data.

IH.2.5.D. NMR spectra

The starting material [1-‘3C]-Glc-l-P displays a major doublet 0t -]3C1 peak at
96.50 and 96.44 ppm and a minor doublet B -'3C1 at 99.89 and 99.85 ppm (approximately
1.8 % of the height of the 96.50 ppm peak) (Figure 111.4). The first batch of [1-13C]-
ADPGlc exhibits a large doublet 0t -'3Cl peak at 98.50 and 98.55 ppm and a minor

doublet B -I3C1 peak at 99.95 and 100.02 ppm (approximately 3.0 % of the height of the

98.50 ppm peak) (Figure 111.4).

76

5.8.2-6.. -2 . :2... Es 3-0.0-8.. .2 .82.... 8.9... 2.. ..o 8.8... ”.22 ...E 2:5...—

 

 

 

IO :0
:0 IO
. .I..T.I._.
__ __/o
0 O
z z :0
\ 0:
j _ \V o o:
. 55.586 -2-..
:2 m—
9. 8. :2 8m RN 8m onm
. :!.3:3:: 3.33.3.3 :3.-...: .3. ,3
8...: «Em 18.8.6
on 8. on. 8m Rm can own
L.33L|p333......_.3.3».r33_ 33....3..3333_. .3_.3.._ r33....353333_...._.
dd 4 l 41 1 +1 - all 1
Io
_ $36-6... -zd
mola/
W 0
IO
OI
o...
O

77

5.9.3-82 -: : H.22 .0 8.8% fizz 3: 2&5

 

78

.2083 o o 2-5. 25 3:20-. co: -5 . .3538 3:93. 2: ..o 2.8% fizz 3: 2:5

 

 

 

 

 

 

 

 

can. 5% 30.806
mm 9. m... 8. m2 3.. m: 8... mm... on... S... 8m mm». on «.2
._.3....3....LL3..._..3._3..__...3L.r.._3...h33.__._3.__33._33.3__..3_._
«fl .—
fi. _ iom. _ _ 81.8..
_ w
0.0.5339 -5.
2
Ga... «Em 382.0
mm om m.“ 8. mm. 03.. m: com
3.3_...Lh3_l,3r_L._.._33<.3__3.3_.3.._...3_3.__
__ .

 

 

 

32-0.2391...

In the carbocation DGM, positively charged C1 is deshielded and probably its peak
will appear at lower field than the ADPGlc C1 peak (98.5 ppm). However, in the batch I
[l-‘3C]-ADPGlc NMR spectrum, we see two peaks (173.8 and 188.0 ppm) in the low
field region where the DGM C1 peak may appear. To vacant the low field region for a
clear interpretation of the C1 peak of DGM in GS-ADPGlc-HEPPSO complex, the
enzymatic synthesis buffer bicine and HPLC buffer ammonium formate were purposely
replaced by HEPES and Tris, respectively, in the production of batch 2 [1-13C]-ADPGlc
and [UL-'3C6] -ADPGlc. However, the 173.8 and 188.0 ppm peak persisted, indicating
they were from some other source than buffer bicine or ammonium formate (Figure 111.5
and Figure 111.6). On the other hand, unlike ammonium formate which has low vapor
pressure and most of it was removed in the lypholization process, the replacement buffer,
Tris, was condensed during lypholization and display a large peak at 53.9 and 57.2 ppm
in batch 2 [1-‘3C]-ADPGlc and [UL-'3C6] —ADPGlc (Figure 111.5 and Figure 111.6).

NMR spectra of uniformly labeled Glc-l-P ([UL-I3C6] —Glc-1-P) and ADPGlc
([UL-I3C6] —ADPGlc) are shown in Figure 111.6. Compared to the regular ADPGlc '3 C
NMR spectrum (C1 : 96.24 /95.93 ppm; C2 :73.53 /73.39 ppm; C3 :73.39 ppm; C4 :69.92
ppm; C5 :72.08 ppm and C6: 61.07 ppm) (16), [UL-'3C6] —ADPGlc displayed a
complicated pattern due to the coupling between neighboring sugar 13C-labeled carbons.
The central chemical shift of [UL-'3C6] —ADPGlc sugar carbon peaks are as followed: C1
: 9827/9786; C2 :76.20 ppm; C3 :75.34 ppm; C4 :71.66 ppm; C5 :73.76 ppm and C6 :
63.05 ppm (Figure 111.6). Moreover, the appreciable doublet peak at 99.41 and 99.77 ppm

corresponds to the C1 of B-[UL-'3C6] —ADPGlc, indicating that there is a considerable

amount of the [3 anomer in [UL-”Cd —ADPGlc.

80

III.3. Model Building and Structure Refinement

III.3.1. apo deS (C7S: C428S)

E. coli deS crystals were obtained from the complex of deS (2.5 mg/mL) and
ADPGlc (0.54 mM) in a buffer of 40 % (w/v) PEG 4000,
0.1 M Tris (pH 7.5) and 0.2 M Na tartrate. The crystal
used for data collection incubated for 16 wk before

reaching its maximum size 0.2 x0.2 x0.3 mm (Figure

 

111.7).
The structure of apo-deS was solved by the Figure "1'7 A crystal 0f
deS
Molecular Replacement method with the program
MOLREP (Collaborative Computational Project, CCP4)(17) using the A. tumefaciens GS
structure (lRZV.pdb) as a search model, with non-conserved amino acid residues
mutated to alanine, except for glycine. Initially, plausible rotation and translation
functions for the N-terminus (1—240) were identified and the C-terminus (271-456) was
subsequently incorporated to search for its counterpart. Placement of the model was
optimized by rigid-body refinement. The apo deS as well as all other GS structures
was obtained through REFMACS (CCP4) refinement and map calculations with model
building using TURBO-FRODO. All the structure refinement statistics are listed in Table
111.2. The final structure of apo-deS consists of the first 475 residues of the protein
(total 477 in E. coli GS) and only one molecule in each asymmetric unit. Figure III.8

represents the Ramachandran plot of the apo deS structure.

81

 

mvé
Sod

o...NN
00.2.

.8. 5 mi
.88. m...
88:8
5.3 E...
Avonmmvmwmmmm
.mm.~-~m.~. 8.3.8
8...
mzméqdozo

odm 5.0m 56m
0wa \mdww 3.0g

3;

8.. m3 g %cm ccom

mood «Ed 2.. 59.9 ocom
«2...; .32 Ea... 22332. 6.8...
2.8 3.8 gym”... mu
5...; 8.8.. 3.3.0.... mo
3.3.33 E2553”.
as 52 .8: m8 3 b \ r
8a... as 6.3. ms 3... 3.95. a .a
88. m8 8: m8 3... 88963800..
3%...” 6.9 v... 32.9.3?
:88 88m 833 {EN 82.032 :38...
83-8.8 8.~ - 8.8 85.2.... «on - 8.8 A 3 5.5.8me
8.. 8.. 2V £mc0_m>m>>
8.295.003 89-20200 cc: Emmm
«czar—3m 9.3.3005 Sun.
@858 5.8 c.8208 5.8 C 3 a \ 5
Ba? 8.82.8.8. 08 6.8... 588 2.. o . a \ m
=3 «E:
.3 mm m 8.06 83w

 

montacooamown=odo<€hhnm Own—mmz-mo<.<hhmm AmmcvonmbovaEu
3.8.9.30 :65 new E53: m6 ...oom ..o 3:»sz 2.2.5.62 uca cowoa=oo Ban. «.5 233.

82

 

38..” noon 080 xcmm 98 50.91
8.8 v.8 No» .8. E850 Eozow
:8. 8.8 $8. 8.8 :33. a. .9950
.3: 8.3. £20 .0029 0535028...
5.8.8 0.8 00 8 0250030950
.8. 8.8. 0.8 00 8.. 025008095...
.8. 8.8 0.8 90 8 02288095..
.3”. 3.8 mam m0 8.. oEEoommogOc
a: 8.8 oomtzm 6 08020..
5.88. E. 8.8 Omamwrw
E. 8.2. 5.8.8 ago...
.85 8.8 .85 8.8 :83 8.8 5929.
8.8 3.3. 3.8 3.3 m 03.03..
.308 m
o.o od ed 3060. 303885
No No No 9860. 0026:... 28.9.00
m... r: 3. macaw. 0032... 85:69..
v.8 58 0.8 9.2mm. 859$ 822
3.. 023350 apicuzanum

 

002..a:00umom=0d0<.<hhnm

Own—amI-n_D<.<hhnm

38385805..

$.28. a... 833

83

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

 

z l I I where I is the observed intensity and <1> is

the average intensity obtained from multiple observations of symmetry-related
reflections.

R _ leFobsl - chal”
_ ZlFobsl

 

_ EllFobsl—chal”

d Rfree —

 

z I F 0 b Sl where reflections belong to a test set of 5 %

randomly selected data.

e . . .
Values 1n parentheses are for the atom number of that specres in category.

84

 

 

 

 

 

 

 

 

180

Figure 111.8 Ramachandran plot of E. coli deS. Phi (degrees) is x and Psi

(degrees) is y.

111.3.2. Wild - type GS in Complex with ADP and DGM (thSa)

The three- dimension structure of wild-type GS in complex with substrate
revealing the authentic active site of the protein is of great
importance to elucidate substrate- binding and understand
the glucosyl transfer mechanism. Cocrystallization of thS

and substrate ADPGlc produced several diffraction —

 

quality crystals, among which thSa is the only Figure 111.9 Crystals of thSa.

The center crystal is the one the
diffraction data was collected
from.

one that was obtained from short (4 wk) incubation
and contains ADP and DGM in its active site.
Single crystals of thSa of size ranging from 0.05 x 0.1 x 0.1 mm to 0.1 x 0.2 x 0.2 mm
were obtained afier 4 wk incubation in a solution of 40 % (w/v) PEG 4000, 0.2 M NaAc
and 0.1 M HEPPSO (pH 8.1) (Figure 111.9).

The structure of the first E. coli GS, apo deS, was used as the phasing model for
molecular replacement performed with MOLREP (CCP4). There is only one molecule in
each asymmetric unit. The final structure contains all residues except for the His-6 tag
attached to the thS C-terminus. In addition, one buffer molecule HEPPSO, three PEG
links and two acetate molecules are found in the structure. The structure also includes
239 water molecules. Figure 111.10 represents the Ramachandran plot of the thSa
structure, showing that none of the residues are in the disallowed region. The refinement

statistics are listed in Table 111.3.

86

 

 

 

 

 

 

 

 

 

_'l _
.25":
.2: _
f‘“ f
;.
LI. ‘1
r 96 ' 180

 

Figure 111.10 Ramachandran plot of E. coli GS /ADP/ DGM /HEPPSO thSa

complex structure. Phi (degree) is x and Psi (degree) is y.

111.3.3. Wild - type GS in Complex with ADP and Glucose (thSb,

thSc and GSd)

The crystals of thSb and thSd grew in the
buffer 40 % (w/v) PEG 4000, 0.2 M Na tartrate and
0.1 M HEPPSO (pH 7.7), and reached a maximum
size 0.25 x 0.25 x 0.25 mm and 0.2 x 0.2 x 0.2 mm,

respectively, within 12 wk (Figure 111.11). The

 

0.2 x 0.2 x 0.2 mm single crystal used for

Figure 111. ll Crystals of thSb

thSc data collection was also obtained from a
12 wk incubation under the same conditions except that the pH of HEPPSO was 7.6.
thSb, thSc and thSd are from slightly different conditions (pH) and all
contain ADP and glucose in their active sites. The structure of E. coli GS thSa (PDB
ID: 2QYY) was used as the phasing model for molecular replacement performed by
MOLREP (CCP4). There is only one molecule in each asymmetric unit. The final
structure includes all the protein residues, ADP, glucose, buffer molecule HEPPSO. In
addition, thSc and thSd have one and four PEG links in their structures, respectively.
The His-6 tag attached at the thS C-terminus was not traceable. The thSb, thSc and
thSd structures also include 246, 23 9, and 209 water molecules, respectively. Figure
111.12, 13, and 14 represents the Ramachandran plots of thSb, thSc, and thSd
complex structure, respectively. There are no residues in the disallowed region. The

refinement statistics are listed in Table 111.3.

88

 

 

 

 

 

 

 

180

 

Figure 111.12 Ramachandran plot of E. coli GS / ADP / glucose / HEPPSO

thSb complex structure. Phi (degrees) is x and Psi (degrees) is y.

 

 

 

 

 

 

 

180

Figure 111.13 Ramachandran plot of E. coli GS /ADP lglucose [HEPPSO

thSc complex structure. Phi (degree) is x and Psi (degree) is y.

90

 

 

 

 

 

 

 

Figure 111.14 Ramachandran plot of E. coli GS /ADP /glucose /HEPPSO

thSd complex structure. Phi (degree) is x and Psi (degree) is y.

 

 

21 mm: 8.: 8.: C 0.93 ucom

:55 85.5 35.5 355.5 3.. £92 ccom

«mag .32 Eat cos—255 .m.E.._

«.8 3.: 98 3.3: 55.8.. mu

«.3 m5: m8 8.: §....2 mo

35.335» “552.com

:5. 8.2 SN. 6.2 :53. 8.: SN. 3.2 2. b \r.

8.3. who 8.8. a... 3.8. 8 6.8. 3.2 55. 83mg a

6.8. 53 :33. «.8 8.8. 5.8 6.3. 8.8 :5. 88325800..

68. 8.~ 8d. 38 :58. 3.3 3.5. 3 55:533.)...

88. 83... 633. s 88 38... 8~3 8:. 583 2282.6. Ear.

A3383. 3.~ - 2.3 83-83. 83 - 2.3 $8.93. «an - 2.3 85-33. 3 - 8 A 4.. 525681..w

8.5 8.5 8.5 5.: A4. £mc2o>m>>

8.25-2825 88.2825 88.2825 39-3505 m5: 885

«33353 $53395 flan

5.85.858 5.85.8 5.8 5.85.858 5.8 5.8 5.8 A... 5 a \ e

58.828.828.89 55.822.59.253: 88:38:34.8: 88:38:38: 2.. o \ a \ m

:8 E5

.3 .3 :1 .1 5:95 825
305 owe; 5mg; «mats

3.8353 mo ...oom an? 5:? “—0 «£333» 23:353. new 20:02.3 Sun 3.. «3a...

92

 

 

538 E8 88 >>o« 88 £85 «80 522a
2: 8.3 8.8 8.3 A8. .5858 Emzow
8: 88 $8. 8. 8 $8. 88 a8. 8.8 .99?
A8. 8.8 8. 8.8 so. 38 6525 2238559
5. 8.8 E. 8.8 E. 8.8 E. 8.8 owamwro
f. 8.8 .209
a: 3.8 a: 8.8 a: 8.8 8820»
8. 88 5. 8.8 A3. 8.8 8. 8. 8 nos.
83. 28 83. 8.3 83. 3.3 83. «F8 5299
8.3. 8.8 3.8 «58 A«..<. m m5m$><
.808 m
5.5 5.5 5.5 5.5 8.259 83235
«.5 «.5 «.5 «.5 .258 8328 2855mm
8 8.8 8.8 m8 252 59528. 8555595.
«.58 5.58 5.58 5.55 8562 58:98 822
3... 3.3.5.3» :Eucazanam

393 5mg; 305 «mats

55:8. 2.. 28..

93

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

Z“II-IOU”

Z l I | where I is the observed intensity and <I> is

 

b Rmerge =

the average intensity obtained from multiple observations of symmetry—related
reflections.

R _ XHFObSI - 'Fcal”
c — ZlFobsI

 

ENFObS|—IFCCIIH

Z | F0 b SI where reflections belong to a test set of 5 %

 

d Rfree =

randomly selected data.

e . . .
Values 1n parentheses are for the atom number of that specnes in structure.

94

III.3.4. GS Mutant E377A in Complex with ADP and
Oligosaccharides

The structure of E377A in complex with ADP
and Oligosaccharides was obtained from

Oligosaccharide cocrystallization and soaking. E.

 

coli GS E377A (8.6 mg /mL) was first cocrystallized
with 5 mM ADPGlc and 5 mM maltohexaose in ggfzggloigcghrzisdtiiglfnd
buffer 40 % (w/v) PEG 4000 and 0.1 M HEPPSO (pH E377A-

7.5). Crystals of size (0.05 x 0.05 x 0.05 mm) were spotted after 4 wk and to that 3 ul
crystal-containing drop 2 ul soaking solution was added which is comprised of equal
volumes of well solution and 100 mM maltopentaose (Figure 111.15).

The structure was solved by the Molecular Replacement method with MOLREP
(Collaborative Computational Project, CCP4) (17) using the E. coli GS thSb (PDB ID:
2QZS) structure as a search model. Refinement and map calculations were carried out
with the program REFMACS (CCP4)(17). All model building including manual
adjustments and fitting of carbohydrate density was performed using TURBO-FRODO
(18) running on a Silicon Graphics workstation. Water molecules were added using coot

and inspected visually prior to deposition. Five percent of the observations were flagged
as “free”(7) and used to monitor Rfree. The final model displays good stereochemistry as

evaluated with the program PROCHECK (CCP4)(17). Figure 111.16 represents the
Ramachandran plot of Oligosaccharide-bound E377A structure. None of residues are in

the disallowed region. The refinement statistics are listed in Table 111.2.

95

 

 

 

 

 

 

 

 

180

Figure 111.16 Ramachandran plot of E377A /ADP /oligosaccharides complex

structure. Phi (degrees) is x and Psi (degrees) is y.

96

III.3.5. GS Mutant E377A in Complex with ADP

Glu377 is a 100 % conserved residue throughout the whole glycogen synthase
family across bacteria and animals. Mutating Glu377 to Ala diminishes the enzyme
activity while the ability of the protein to bind substrate ADPGlc still remains (l 1). To
better understand the role of Glu377 in the GS
catalysis, the mutant E3 77A was co—crystallized with
ADPGlc and crystals were obtained within 14 weeks.

E. coli GS E377A +ADP complex crystals were

obtained from a buffer containing 7.8 mg/mL protein,

 

3 mM ADPGlc, 3 mM maltotriaose, 40 % (w/v) PEG Figure 111.17 Crystals of
4000 and 0.] M HEPPSO (pH 7.5). After a16 wk ADP-bound E377A.
incubation, the crystals grew to 0.2 x 0.2 x 0.2 mm and diffracted to 2.3 A (Figure
111.17).

The structure of E. coli GS thSa (PDB ID: 2QYY) was used as the phasing
model for molecular replacement performed by MOLREP (CCP4). There is only one
molecule in each asymmetric unit. The final structure includes residues 1-476, buffer
molecule HEPPSO and an azide molecule. The last residue 477 and the His-6 tag
attached to the E3 77A C-terminus were not traceable. The structure also includes 298
water molecules. Figure 111.18 represents the Ramachandran plot for the E377A structure;

none of the residues are in the disallowed region. The refinement statistics are listed in

Table 111.2.

97

 

 

 

 

 

 

 

 

180

Figure [11.18 Ramachandran plot of E3 77A-ADP-HEPPSO complex

structure. Phi (degrees) is x and Psi (degrees) is y.

98

111.4. Escherichia coli GS Structures

III.4.1. Apo deS (C7S: C4288) Structure

E. coli GS is a 52.8 kDa protein with a total of 477 amino acid residues. Apo deS
(C7S: C428S) exhibited a typical twin-Rossmann GT-B fold; the N- terminal domain (1-
241) and C-terminal domain (250-457), each of which is composed of a “sandwich” of
parallel B-sheets between a-helices, were separated by a deep cleft. The extended
interdomain linker peptide (242-249) connects the N- and C-terminal halves. The helical
tail, 0:18 (458 - 476), crosses over from the C-terminal domain to pack against the N-
terminal domain (Figure 111.19). These results are consistent with the previously reported
open form structure of PaGS (19) and AtGS (20). E. coli deS (C7S: C4288) exhibits
specific activity and substrate ADPGlc affinity comparable to E. coli wild-type GS (1 l).
The apo deS structure reveals that the mutant sites Ser7 and Ser408 are not close to the
active site and are separately buried in the protein structure (Figure 111.19), suggesting
that this double mutation has little impact on the enzyme tertiary structure and the wild-
type GS, when not bound with ligand, should adopt a similar open, inactive form as seen
in apo deS.

As Figure 111.19 shows, the two domains of apo deS are loosely connected. Most
interdomain interactions occur in a defined region, in particular, between the linker
peptide (242-249) and domain-spanning helix 0:18 (458 - 476) (V a1248 N-Phe460 O,
3.14 A), and the N-terminal helix (17 (212-220) and the C-terminal helix (114 (398-403)
(Thr214 OG-Gly399 O, 2.99 A; Glu218 carboxyl oxygen-Gly399 N, 3.07, 3.09 A)

(Figure 111.20).

99

g8- Z 859-0 EH30 g8- Z

wovaom wciom

   

whom

whom

.869 we E 8527. Pa Amwovo ”mnov mOEu mo 8% 5:938 2:. .053 Em: 5 fl Amonamv 35 £0: 35258-0 2:

380:3 :0on E 08 8mm-2mv 55 x20: 98 m_m-§~ 32 3583? Z 2E. 5038099 .owafio was Snowman @828
Be 83$va w to Vacs wngmméfifiow Ea www-mvm 0233 £98385 2:. iguoommou 6:3 can 25:?» @828
on films €88 3:526 Ea 5m. 5 558 255-2 2: wees mam mo ”EOE: :85 3.5 95$..—

100

Gly399

 

Figure 111.20 Interdomain interactions between Glu218 and Gly399; Thr214
and Gly399, and Va1248 and Phe460 are shown as dOtted lines in E. coli
deS. Loop 212-215 and helix a7 (215-220) are in green whereas the C-
terminal helix al4 (398-403) is in light -blue. The linker peptide 242-254
and domain-spanning helix (x18 (458-476) are colored magenta and orange,

respectively.

101

III.4.2. GS - ADP - glucose - HEPPSO Complex Structures

GS-ADP-glucose-HEPPSO complex structures thSb, thSc, and thSd display
a typical twin-Rossmann fold, which is similar to apo deS (C7S: C428S). But the two
domains are very close together in wild-type GS complexes while the two domains are
spread relatively far apart in apo deS. Several new interdomain interactions are
observed between Asn162 and Gln304, and LyslS and Glu357 in the wild-type GS
complex narrow interdomain cleft. We designate the molecular organization seen in wild-
type GS complexes with a interdomain narrow clefi as the closed form while that of apo
deS with a wide cleft as the open form of GS. A detailed comparison of these two
forms is described in section 111.5.].

Wild type GS complex structures thSb (2.20 A), thSc (2.26 A), and thSd
(2.37 A) are from three independent data sets, but exhibit almost identical structural
organization as the structural displacement index RMSD is less than 0.1 A between them.
Except for small conformational difference of the HEPPSO ethyl end, ligands bound in
the GS active site (ADP and glucose) in these three structures are in superimposable
positions. Therefore we choose the thSb, the highest resolution structure, as a
representative, to illustrate the interaction between ligand and protein.

The narrow clefi at the N- and C-domain interface serves as the ligand-binding
pocket (Figure 111.21). ADP binds mostly along the C-terminal wall and the glucose
resides at the bottom of the cleft. HEPPSO is exclusively on the side of the N-terminal

domain. The detailed description of ligand binding in G8 follows.

102

 

Figure 111.21 Overall structure of the thSb complex. The N-terminal

domain is in pink and C-terminal domain is in blue. The linker peptide 242-
254 and domain-spanning helix (x18 (458-476) are colored magenta and

orange, respectively.

103

III.4.2.A. ADP binding site

Asp2

 

Leu38l

 

Lys305

Glu377

Figure 111.22 Top: ADP (shown in yellow and
in atom colors) bound in the active site of wild
type E. coli GS. Hydrogen bonds between ADP
and protein are shown as broken lines.
Residues from the N-terminal domain and the
C-terminal domain are colored pink and blue,
respectively. Bottom: 1.0 a contoured 2Fo-Fc
electron density map of the bound ADP in the

active site.

104

The ADP heterocyclic adenine ring stacks on the conserved Tyr355 (Phe in animal
GS and plant SS (Figure 111.22). The adenine N1 atom donates a hydrogen bond to the
backbone amide of His356 (3.08 A). The adenine amide N6 is in contact with the Gly3 54
backbone amide (2.91 A).

The ribose of ADP interacts with the Asp21 (2.84 A) and LyslS (2.74 A) side-
chain with its 03 ' hydroxyl group. The ribose moiety adopts a C3 '-endo conformation
relative to the adenine base (Figure 111.22) and the water-mediated hydrogen bond
between the ribose 2- hydroxyl (3.05 A) and the adenine N3 atom (2.88 A) presumably
reinforces it.

The ADP phosphate groups are tucked back with respect to the adenine base so that
05* (connecting adenine and phosphate) and the distal phosphate 03B fall in the
hydrogen bonding range of Glyl 8 (3.24 A and 2.72 A, respectively) (Figure 111.22). The
proximal phosphate is close to the invariant loop 374-381 and both phosphate 01A and
02A hydrogen bond to the main-chain amide of Leu381 (3.32 A and 2.96 A,
respectively). The proximal phosphate 02A also hydrogen bonds to the Thr3 82 backbone
amide (3.23 A). The distal phosphate group is close to the loop 299-306 and extensively
interacts with Arg300 and Lys305 through ionic interaction, direct hydrogen bonds, and

water-mediated hydrogen bonds (Figure 111.23).

105

Arg300

        

glucose

   

Ser374 Glu377

Figure 111.23 Extensive interactions between ADP and residues Arg300, Lys305
and Glu377 in the wild type E. coli GS active site. Water molecules are presented as

red spheres and water-mediated interactions are labeled as red.

106

 

\

Helix al(17-32)

Helix a13(382-389)

Figure 111.24 Helix al(l7-32) (cyan) and (113082-389) (purple) point to
the ADP phosphate group. The NH-ends of both helices are presented as
yellow ribbons.

Besides the local interactions described above, the overall GS protein structural
organization also contributes to ADP binding, particularly the helix dipole moment of
helix al(17—32) and helix a13 (382-3 89) whose NH-terminal ends point towards the
ADP phosphate (Figure 111.24). The helix dipole moment is caused by the aggregate
effect of all the individual dipoles from the carbonyl groups of the peptide bonds pointing
along the helix axis. Because the NH-terminal end of these two helices comprise neutral
amino acid residues (GnGL and ngzQL, respectively), the NH-tenninal positive charge
from the overall helix dipole moment was not neutralized and directly interacts with the

negative-charged ADP phosphates.

107

III.4.2.B. Glucose binding site

Wild-type GS complex thSb, thSc, and thSd were obtained from GS
cocrystallization with substrate ADPGlc. However, the difference density maps of these
thS structures all showed considerable gaps between the distal phosphate oxygen and
the anomeric carbon of the glucose moiety of ADP, indicating the phosphate-glucosyl
bond had been broken and an individual glucose is present.

The individual glucose derived from ADPGlc is exposed at the domain interface
and extensively interacts with the C-terminal 373-380 loop and the N-terminal residues
Hisl61 and Asnl62. The 2- hydroxyl group forms hydrogen bonds with the side chain of
Asnl62 (3.11 A) and Gln304 (3.22 A). The 3-hydroxyl group forms a 2.50 A hydrogen
bond with the carboxyl oxygen of Glu3 77 and makes additional hydrogen bonds with the
backbone amide of Cys3 79 (2.98 A) and Gly380 (2.89 A) (Figure 111.25). The 4-hydroxyl
group makes a hydrogen bond with 01A of the proximal phosphate (2.41 A), and with
the Gly3 80 backbone amide (2.94 A). The 6-hydroxyl group hydrogen bonds to the
interdomain peptide residue Asn246’s side chain carbonyl group (2.90 A) and the His] 61

side chain (2.84 A) (Figure 111.25).

108

HEPPSO

    
 
      
 

 

Gln3 04

:“Gly380 Asn246

Cys379

  
 

Figure 111.25 Top: glucose (shown in yellow and in atom colors) bound in the
active site of wild type E. coli GS. The nearby HEPPSO and ADP are shown as
black and blue sticks. Hydrogen bonds between glucose and protein are shown as
broken lines. Residues from the N-terminal domain and the C-terminal domain are
colored pink and blue, respectively. Bottom: 1.0 0' contoured 2Fo-Fc electron
density map of the bound glucose.

109

In all thS-ADP-glucose structures the anomeric C1 is 3.51 —3.65 A from oxygen
033 of the distal phosphate, which is almost twice a regular carbon-oxygen bond length,
clearly showing that the glucose phosphate bond is broken. On the other hand, the 2.50-
2.99 A distance between the glucose 1- hydroxyl and 03B of the distal phosphate
suggests a hydrogen bond between the glucose and the leaving group phosphate (Figure

H126)

v

ADP

HEPPSO

 

isl61

Figure 111.26 Interactions between glucose, ADP and HEPPSO in the thSb

active site.

110

IH.4.2.C. Acceptor analogue HEPPSO binding site

All thS complex crystals grew in the presence of 4-(2-Hydroxyethyl) piperazine-
1- (2-hydroxypropane sulfonic acid) (HEPPSO) (Figure 111.27). Clear electron density
was consistently observed for HEPPSO in the interdomain cleft of thSb, thSc, and
thSd structures. Comparison of thS ——ADP-HEPPSO complexes (thSb, thSc, and
thSd) with the MalP-PLP-maltopentaose complex structure reveals that HEPPSO
occupies the first two glucose unit positions of the acceptor maltopentaose in the MalP
complex, indicating HEPPSO probably acts as an Oligosaccharide acceptor analogue for
E. coli GS (21). This interpretation is further supported by our GS E377A-ADP-
oligosaccharide structure where the Oligosaccharide also binds at the HEPPSO position in

the active site.

 

o2 H04
ll H2 1 H2 / \ H2 H2

I-Io1 s c ~fi—c—N1 Nz—C—C—Osl-l
03 \__/

Figure 111.27 Diagram of HEPPSO and its atom numbering.

11]

   
 

HEPPSO ADP

     

: p" |
‘23] \‘2.78

|
water 0‘
\

109‘.‘
Trp138

Asp l 37 glucose

 

Figure 111.28 Top: Interactions between HEPPSO and N-terminal
residues, ADP, and glucose. Bottom: 1.0 0' contoured 2Fo-Fc electron
density map of the bound HEPPSO.

112

HEPPSO lies along the N-terminal side of the crevice wall with its ethyl hydroxyl
end near the glucose and ADP (Figure 111.28). The ethyl hydroxyl end of HEPPSO is also
close to the KlsTGGL motif loop and makes a 2.93 A hydrogen bond with the Leu19
backbone. The sulfite end of HEPPSO points to the protein surface and interacts with the
N-terminal residues along the cleft wall. The conserved Trp138 forms hydrogen bonds
with both the branched hydroxyl group and the sulfite oxygen. The conserved residue
Aspl 37 interacts with the branched hydroxyl group through a water-mediated hydrogen
bond (Figure 111.28).

Mutagenesis and modeling studies suggest that the conserved Aspl 37 and Trp138
are involved in recruiting and locking the glucan chain acceptor in place (22). The fact
that HEPPSO cannot be substituted in crystallization experiments with HEPES, which
has an ethane-sulfonic group instead of 2-hydroxypropanesulfonic group as seen in
HEPPSO, indicates that the incorporation of HEPPSO into the protein requires the

interactions of its branching hydroxyl 04 with Aspl 37 and Trp138.

113

[11.4.3 GS-ADP-DGM-HEPPSO complex structure

While an individual glucose fits the density map of thSb (2.22 A), thSc (2.26
A) and thSd (2.37 A) which was cocrystallized with ADPGlc for 12 weeks, the density
map of thSa (2.83 A), which was incubated with ADPGlc for four weeks, is not well-
defined at the glucose site. Tentatively a D-glucopyranosylium cation (DGM) was

assigned though the DGM deprotonation product D-arabino-hex-l-enitol fits just as well

(Figure 111.29).
H OH H 0“ OH
H
/o 0
HO c5 \ Ho fLo
HO I H + H H0\C H / C HO
' /C1 | \C2/ 1\H H0 H
C2 H \ H OH
H \
OH OH H OH
DGM D-arabino-hex-l-enitol glucose

 

Figure 111.29 2.5 o Fo-Fc electron density map of thSa (2.83 A resolution) with
DGM (left), D-arabino-hex-l -enitol (middle) and thSb (2.22 A resolution) with

glucose (right). Co-planar atoms are in red. The ADP molecule is shown as lines
for clarity.

114

The “D-glucopyranosylium cation (DGM)” occupies the equivalent position of
glucose (Figure 111.30) and most interactions between glucose hydroxyl groups and
protein seen in thSb are retained in thSa for “DGM”. DGM differs from glucose in
its positive charge on C1 and 05, the sp2 hybridization of C l , and the subsequent partial
planar conformation of the sugar ring. The distal phosphate O3B, HEPPSO ethyl end, and
Hisl61 backbone carbonyl group are close to the partially positively charged “DGM” Cl
and 05, presumably providing electrostatic stabilization (Figure 111.31). The detailed

interacting distances in thSa and thSb are listed in Table 111.4.

Glc/DGM

 

Hisl61

Figure 111.30 DGM in thSa is in a superimposable position of glucose in
thSb. Ligands in thSb are in yellow whereas those in thSa are atom-wise
colored.

115

 

HEPPSO

Figure 111.31 Positively charged Cl and 05 of DGM are stabilized by ADP,
HEPPSO, and Hisl6l. Interactions starting from the DGM C1 are shown as red

dotted lines.

Table 111.4 Comparison of the environment of DGM and glucose.

 

 

 

Distance A-B (A) in Distance A-B (A) in
Atom A Atom B thSa thSb
Glc/DGM Cl ADP 033 3.26 3.65
HEPPSO 05 3.33 3.31
HiS161 O 2.85 2.90
Glc/DGM 05 ADP 0313 3.14 3.80
HEPPSO 05 3.10 3.12
H1516] 0 3.82 3.46
ADP O33 HEPPSO 05 2.98 2.76
Glc l-OH ADP 033 N/A 2.99
HEPPSO 05 N/A 2.42
Glc/DGM 6-01'1 H18161 ND] 2.69 2.84

 

 

 

116

In spite of the limited resolution of the thSa (2.83 A) crystal, the finding of DGM
(or DGM deprotonation product D-arabino-hex-l-enitol) in the thSa complex active
site is very significant. In particular, it provides strong evidence for an SNi/ SN]
mechanism, as oppose to the once-prevailing double-displacement mechanism. The
possibility that DGM is in the GS active site was reinenforced by the MalP-PLP— PO43’
maltopentaose structure (PDB: 2ASV, 1.95 A) where the density map in its active site
also suggests a glucose-like species lacking 1- hydroxyl group at the superimposable
position. Strangely, the stable 1,5-anhydrosorbitol was built in, although DGM would
also fulfill the density map there. Unfortunately, this MalP structure was only deposited
in the PDB and is not officially published. As a result, the official quotation of DGM is
missing and a firm piece of experimental evidence of DGM must be obtained before
making the claim that DGM is the GS catalysis intermediate. To that end, we attempted
to identify DGM in the GS active site using NMR and more details are described in

section 111.53. C.

117

111.4.4. E377A-ADP-HEPPSO Complex Structure

ADP-bound GS E3 77A displays an essentially identical structure to the thS
complex (R.M.S.D of total 476 residues Con atoms = 0.53 A). Unlike in the thS
structure where ADPGlc was decomposed into ADP and individual Glc/DGM, only ADP

is present in the active site of E377A (Figure 111.32).

Arg300

 

Ala3 77

Figure 111.32 Overlay of the active site of E. coli thSb and E377A. The
residues and ligand of E377A structure are colored yellow whereas residues in
thSb are in green and ligands are in magenta. Water W1, W2, and W3 are
located between Arg300 and the ADP distal phosphate oxygen in the E3 77A
structure.

118

The most significant difference between the active site of E377A and thSb is the
loss of the glucose moiety of ADPGlc whose 3-hydroxyl group hydrogen bonded to the
Glu3 77 side chain in thSb. Moreover, a drastic conformation change was observed at
Arg300. Instead of being close to ADP and interacting with the ADP distal phosphate,
Arg300 flipped its side chain out of the active site in the E377A-ADP complex structure.
111.4.5. E377A-ADP-Oligosaccharide Complex Structure

To understand the molecular basis of glycogen recognition to GS, we have
determined the crystal structure of GS E3 77A in complex with ADP and Oligosaccharide
at 2.3 A resolution. Four Oligosaccharides were found in the interdomain cleft and on the

enzyme’s N-terminal surface (Figure 111.33).

 

C-term N-term

Figure 111.33 Overall structure of E. coli GS E377A in complex with ADP and
Oligosaccharides. ADP and Oligosaccharides are shown as sticks. The secondary
structure elements which host residues interacting with the surface-bound
Oligosaccharide at Gc and Gd sites are labeled.

119

III.4.5.A. Interdomain cleft Oligosaccharide binding

The Icy-weighted difference Fourier map (2Fo-Fc) showed continuous density in
the interdomain active site cleft (“Ga site”) (Figure 111.34), corresponding to three well-
defined glucose rings extending from the center toward the enzyme surface, which
directly mimics the glycogen binding in the GS catalytic site. All protein residues that
directly contact this Oligosaccharide are well defined in the electron density map. The
interaction scheme and the cartoon presentation are shown in Figure 111.35 and 111.36,

respectively.

ADP

 

Figure 111.34 2Fo-Fc electron density map contoured at 10' with maltotriaose
at the Ga site. The ADP molecule is shown for reference.

120

 

/Asp137
,0,
,0. o
x I, I H2N\
I" I” //C\
OH ,,-’HOH Asn162
O -
H\\
,OH‘ \‘O' O
\\ \// O-
'o’P\ /
'I x O IK\\Ade
,’ . ‘, Q
\ ’ .
‘NH \
Gly18 TH \ ‘,
Gly17 ‘_
O\\ ,
/
Glu9

Figure 111.35 Schematic representation showing the interactions between
E. coli GS E377A residues and the bound maltotriaose at Ga site. Observed
hydrogen bonding and aromatic stacking are depicted with regular and
wide dashed lines, respectively. Van der Waals interaction is marked with

the wavy line.

121

ADP

      

His96

via

0

3.44 o'
O

  
  

  
 
    
 
  

-. .3‘33 Thrl.‘
Gly17

3-05 2.85,’
I 3.13

O

.28 '

a ’ . ' '2.62
.13
O . . .

2.96
water
I
2.76 Glu9

    
  
 

   
   

‘ ly18

   
  
  

Asn162
Trp138

Aspl37

Figure 111.36 The interaction between enzyme GS and the bound
maltotriaose at the Ga site.

122

The H sugar is buried close to ADP in the interdomain cleft (Figure 111.35). Its 4-
hydroxyl group is 2.62 A from the ADP phosphate oxygen O3B. Bidentate hydrogen
bonds are made between OH-2, OH-3 and Aspl 37 OD2 (2.44 A and 2.76 A,
respectively). Its 6-hydroxyl group is close to the loop 9-19 that hosts the highly
conserved K. 5TGGL motif, forming hydrogen bonds to the backbone amides of Gly17
(2.85 A) and Gly18 (3.13 A) as well as the side-chain of Glu 9 (3.28 A).

The sugar at subsite +2 stacks against Tyr95 (3.80 A) presenting the only aromatic
stacking between the bound oligosaccharide and GS (Figure 111.36). The 3-hydroxyl
makes interactions to the Hisl39 imidazole NE2 (2.78 A) and Trp138 indole NEl (2.90
A). The 6- hydroxyl is within hydrogen bond distance of the side chain of Thr16 (3.05
A). The sugar at subsite +3 interacts with His96 NDl through its 3-hydroxyl (2.92 A). Its
4-hydroxyl group hydrogen bonds to the Arg300 mainchain and is in van der Waals
contact with Thr302, presenting the only interactions between the oligosaccharide and the
C-terminal domain.

The configuration of the reducing sugar in subsite +3 is in the (1 configuration and
not the [3 configuration, which is the more favored configuration for a terminal glucose
residue. We suggest that this hydroxyl group is not the reducing end of the
oligosaccharide chain but there is at least one more a-l , 4 glycosidic link. The
“invisibility “ of the following sugars in the X-ray structure is probably due to disorder as
there are few opportunities to make strong interactions with the protein on the surface.

The oligosaccharide at the Ga site in the GS interdomain cleft is in most intimate

association with the active site and most interacting residues are highly conserved

throughout GS, SS, MalP, and GP families. Mutating them, such as Aspl37, Glu9,

123

Hisl39 or Tyr95, decreased GS specific activity to a variable extent ((22) and Yep
unpublished result). A close inspection of the overlaid G5- MalP complex (PDB ID:
1L6I)(21) and E3 77A-ADP-oligosacchride structure revealed that those conserved
residues are lining the interdomain clefi and interact with oligosaccharide in a virtually
identical manner (Figure 111.37). This suggests that the glucan chain binding channels are
constructed in a very similar way to guide substrates into or acceptors out of the deeply
buried catalytic site in MalP and GS and this conclusion may apply to other
polysaccharide processing enzymes like SS and GP, which were suggested to share a
similar active site. It is noteworthy that mutating Tyr95 or Hisl39 to Ala does lower the
GS affinity for maltosaccharide, but not much to glycogen (Yep unpublished result).
Such behavior is not difficult to understand as glycogen is a very long, branched glucan
chain and probably has many more sugar binding sites on the GS enzyme than the short

Oligosaccharides have.

124

 

D307

H309 D137
H139

Figure 111.37 Superimposition of the oligosaccharide- binding site in E3 77A-
ADP-oligosaccharide (green), thS-ADP-glucose-HEPPSO (yellow), and
MalP-PLP-GS (PDB: 1L61) (blue) complexes. The HEPPSO that occupies the
comparable position of the oligosaccharide was omitted for clarity.

GS residues are labeled in black while MalP residues are in blue.

As seen in Figure 111.37, the +1 sugar in the E3 77A-ADP-oligosaccharide

complex occupies almost an identical position as the second glucose (+997) of

125

maltopentaose in the GS-bound MalP structure and the rest of the oligosaccharide glucose
units follow a similar orientation (21,23). The edge glucose +998 in the MalP-GS
complex overlaid well on top of the glucose molecule in the thS-ADP- glucose
complex (PDB: 2QZS) and presumably reflects the conformation of transferred glucose
(—1 sugar) in GS. The 174.73° glycosidic angle between the +998 and +997 sugar in
MalP (-1 and +1 sugar in GS) is largely deviated from the average value of 114° for a
regular polysaccharide (24). The neighboring residues presumably contribute much in
the abrupt kink between the acceptor nucleophile (+1 sugar) and sugar donor (-1 sugar) as
both sugars are involved in a number of hydrogen bonds to the protein, in both GS-bound
MalP (23) and oligosaccharide-bound GS.

Glu3 77 makes a hydrogen bond to the —1 sugar (sugar to be transferred), while
Aspl 37 forms a bidentate hydrogen bond to the +1 sugar (sugar acceptor) (Figure 111.37).
Both of them play an important role in precisely positioning and orienting the sugar to be
added (-1 sugar) and the sugar acceptor, which is of extreme importance in the transfer
efficiency. Actually, to date the most detrimental mutations are E377A and D137A,

which cause a 10,000 and 8,400 fold decrease in GS activity, respectively.
111.4.5. B. The N-terminal surface oligosaccharide-binding sites

Three other oligosaccharide chains are observed at Gb, Go and Gd sites on the N-
terminal domain surface (Figure 111.33). The Gb site is close to the interdomain clefi
opening and the sugar +4 at the Gb site is approximately 6.5 A from the +3 sugar at the
Ga site. In contrast, the Go and Gd sites are quite far away from the interdomain clefi and
the closest distance between site Ga and Go, and Ga and Gd are approximately 28.9 A

and 22.31 A, respectively.

126

Two glucose rings (+4 and +5) are found at the Gb site and the electron density
map is shown in Figure 111.38. The +4 sugar OH-2 and OH-3 hydroxyls interact with the
Glnl94 side-chain. The +5 sugar OH-2 hydroxyl forms a close hydrogen bond to the
His173 Ne. The +5 sugar 6-hydroxymethyl group is in a hydrogen bond distance to the
Tyr103 mainchain oxygen. There are several van der Waals contacts between Tyr170 and

sugars at the GB site (Figure 111.38)

 

_ % Gln194

Figure 111.38 Top: 2Fo-Fc electron density map contoured at 16 with maltose at
the Gb site. Bottom: Interaction between protein GS and the bound maltose at the
Gb site.

127

The oligosaccharide at the Ge site consists of five glucose rings and is in proximity
to B-sheet B3 (55-58), B4 (69-73) and loop 123-129. The oligosaccharide at the Gd site is
comprised of six glucose units and is located between helix (18 (229-23 7) and loopl84 -

93 (between a5-B9) (Figure 111.33). Figure 111.39 shows the electron density map for

Oligosaccharides at the Go and Gd sites.

.-

\ ’2’V.\
...‘nc'e Vfl'frx
w‘ ‘ng; ‘-n‘I

“ J ,4.»—

- '.
.-u 953‘.-
.. —-

4 j
...u.
- . .‘A

 

Figure 111.39 Top: 2Fo-Fc electron density map contoured at 16 with
maltopentaose at the Ge site. Bottom: 2Fo-Fc electron density map contoured

at 16 with maltohexaose at the Gd site.

128

Due to the curled-up conformation, the surface-bound Go and Gd exhibit few
interactions with protein GS. The oligosaccharide at the Ge site interacts with GS only at
subsites + 7 and + 8 (Figure 111.40 and 41). The + 7 sugar 6-hydroxyl group is 3.23 A
from Gln55 NE2. +7 sugar 2- and 3- hydroxyls are in the van der Waals distance of Phe
127 CE. The Asp125 carboxyl group is in close contact with the adjacent 2-hydroxyl
(2.52 A) and 3-hydroxyl (3.24 A) of the + 8 sugar and forms “double-handle” hydrogen
bonds. In addition, the + 8 sugar 2-hydroxyl is in contact with the positively charged

Arg38 side chain (3.01 A).

 

NH2

‘0\ /
\?

Gln55

Figure 111.40 Schematic diagram of interactions between protein and the bound
maltopentaose at the Ge site. Observed hydrogen bonding and Van der Waals
interaction are depicted with dashed lines. The ionic interaction is marked with arrow.

129

 

Asp125

Figure 111.41 Interaction between protein GS and the bound maltopentaose at the

Ge site.

130

The oligosaccharide at the Gd site is the most circular oligosaccharide among the
three bound to GS. The first sugar (+11 sugar) almost reaches the last glucose unit (+16
sugar), as the +11 sugar 3- hydroxyl is only 2.71 A away from the +16 sugar l-hydroxyl
group (Figure 111.42).

The interaction between the oligosaccharide at the Gd site and the protein mostly
occurs at one end of the oligosaccharide, in particular at subsite +11, +12, and +13
(Figure 111.43). The Ile186 backbone amide hydrogen bonds to the +11 sugar 6- hydroxyl
(2.63 A). The Lysl99 amine is in close contact with adjacent hydroxyl groups OH-2
(2.93 A) and OH-3 (3.20 A) of the +12 sugar. On the other face of the +12 sugar, Gly227
and Gly230 are in hydrogen bond distance to the 6-hydroxymethyl group (2.62 A and
2.97 A, respectively). At subsite +13, “double-handle” hydrogen bonds occur between
the adjacent 2- and 3-hydroxyls and the Glul90 carboxyl group (2.84 A and 2.83 A,
respectively). The 3-hydroxyl group also hydrogen bonds to the Asnl92 backbone amide
(2.97 A). The only interaction on the other end of oligosaccharide is between the Hisl 87

NE2, the +15 sugar 2-hydroxyl, and the +16 sugar 04.

131

 

O ~~~~~~ \‘ Gly227

Figure 111. 42 Schematic diagram of interaction between protein and
the bound maltohexaose at the Gd site. Observed hydrogen bonding

and ionic interactions are depicted with dashed lines and arrows.

132

 

Glu190 Asn192

Figure 111.43 Interactions between protein and the bound maltohexaose at the Gd site.

III.4.5.C. The glucose unit configuration in oligosaccharide-bound E377A

All the glucose units in the oligosaccharide-bound GSE3 77A complex adopt a
typical 4C1 chair conformation (25). Among 16 sugar units bound in the E377A-ADP-
oligosaccharide complex, twelve of the glucose hydroxymethyl groups hold gauche-
gauche (gg) and three have gauche-trans (gt) conformations. The OH-6 hydroxyl of the
+2 sugar is in close contact with the Thr16 hydroxyl side-chain and the torsion angles of

05-C5-C6-O6 and C4-C5-C6-O6 are 129.13 ° and -106.44 °, respectively, moderately

133

beyond the definition of a gauche conformation (30 ° to 90 ° or -30° to -90°). The
prevalent low- energy gt and gg conformation of hydroxymethyl groups are largely
reinforced by their interaction with the intra ring oxygen 05 (2.98 :t 0.2 A) (Figure 111.35,
40 and 42).
III.4.5.D. Binding mode analysis

Distinct bidentate hydrogen bonds from a protein carboxyl side chain to the
adjacent sugar OH-2 and OH-3 hydroxyls are observed at each binding site (Asp137 at
the G5a site, Asp125 at the GSC site, and Glul90 at the 05d site) and catch our attention.
Such “bidentate” binding was also seen in the polar interaction between the Lysl99 side-
chain N2 and the +12 sugar. In contrast, the carbohydrate - aromatic interaction usually
seen with sugar - processing proteins such as amylase (26,27), endoglucanase CelA(28),
maltodextrin translocation protein maltoporin(29) are rare in our oligosaccharide-bound
E3 77A structure (only between Tyr95 and the +2 sugar). It appears that GS
predominantly relies on double - handle hydrogen bonding to recognize and orient
glycogen.
111.4.5.E. Oligosaccharide-binding causes little protein conformation change

Both E377A-ADP-HEPPSO and E377A-ADP-oligosaccharide complexes belong
to the I 41 space group with cell dimension a =b =l26.3 A, c = 152.3 A. The E377A-
ADP-HEPPSO complex displays a closed conformation, partly due to the presence of
HEPPSO in the interdomain cleft. Co-crystallization and further soaking with
Oligosaccharides brings three Oligosaccharides to E3 77A, specifically in the interdomain
clefi and on the N-terminal surface. The RMSD value of Con atoms between the E377A-

ADP-HEPPSO and E377A-ADP-oligosaccharide complex is only 0.34 A (30). It appears

134

that crystal packing and protein structural organization were not considerably affected by
the action of the oligosaccharide replacing HEPPSO inside the interdomain channel or
binding on the enzyme surface.

The few conformational changes that were observed were at the side-chain of
Hisl39 whose imidazole ring moved 1.03 A to form a hydrogen bond to the +2 sugar and
at 11e186 which rotated its side chain to avoid being too close to the +11 sugar OH-4
group (otherwise 1.81 A). The side - chain of Arg300 adopts an identical conformation as
in the thS- ADP- glucose -HEPPSO complex, but totally differs from that in the
E377A-ADP-HEPPSO complex where the glucose moiety of ADPGlc is absent. More
discussion regarding Arg300 conformational change in response to ligand binding and its
function is in section III.5.3.B.

III.4.5.F. Oligosaccharide conformation analysis

To investigate the possible conformation alteration in oligosaccharide caused by
protein binding, three GS-bound oligosaccharides at binding sites (Ga, Gc, and Gd) are
compared to each other, to another GT-B enzyme GP-bound maltopentaose, and a
proposed glycogen model by Sundaralingam (31) and Goldsmith(32).

The proposed helical conformation of the glycogen chain was referred to as a
minimum energy conformation that has a characteristic intramolecular hydrogen bond
between the 2-hydroxyl from one glucopyranose ring to the 3-hydroxyl of the adjacent
glucopyranose. We construct an ideal glycogen helix to maintain such a hydrogen bond
network as well as the guidelines derived from a typical maltose structure (24,33). They

include: 1) the 1.4 A a- (1,4) glycosidic linkage and the 114° glycosidic angle, 2) the

135

107° torsion angle of OSn—Cln-O4n+1-C4n+1 , 3) the 122.7“ torsion angle of Cln-O4n-
C4n+1-C3n+l.

At the surface Gb, Gc and Gd sites, oligosaccharides maintain O3n - 02"-. hydrogen
bonds between most glucosyl units. In contrast, the oligosaccharide at the Ga site that is
held in the narrow interdomain cleft possesses a relatively rich interaction with protein,
adopts a more extended conformation, and one O3n - 02“-. hydrogen bond out of two
glycosidic linkages is lost. (Figure 111.34). The distances between O3n - 02"-; in all three
oligosaccharides are listed in Table 111.5 and Figure 111.44 shows the overlay of these

oligosaccharides.

 

Figure 111.44 Overlay of oligosaccharides at the Ga (red, 1-3), Gb (cyan, 4-5), Gc
(green, 6-10), and Gd (yellow, 11-16) sites.

136

Table 111.5 Conformation of oligosaccharides when bound to GS, MalP, and GP

 

binding sites
65a

(+1G)—(+2G)
(+ZG)-(+SG)

65b
(+4G)-(+5G)

65¢
(+6G)-(+7G)
(+7G)-(+8G)
(+BG)-(+9G)
(+9G)-(+1OG)

65d
(+11G)-(+12G)
(+1ZG)-(+13G)
(+13G)-(+14G)
(+14G)-(+15G)
(+15G)-(+16G)

Maltopentaose bound to MalP( 1L6I.pdb)
998-997 (equivalent to -1G to +1G)
997-996 (equivalent to +16 to +2G)
996-995 (equivalent to +2G to +36)

Maltopentaose bound to GP(1P2B.pdb)
S7-86
86-85
$5-S4
S4-S3

Phi(<p)

99.48
66.15

95.10

101.96
102.50
107.60
101.58

86.32
101.52
120.20
114.47
47.01

71.96
102.93
81.19

79.84
105.15
74.33
113.05

Psi(\p)

120.22
85.06

115.07

100.76
125.86
124.62
114.42

-51.41
119.05
123.68
128.36
104.43

59
109.82
100.03

103.98
111.77
85.82
111.63

T

114.33
106.00

108.66

113.00
114.16
111.48
109.21

114.64
115.51
116.15
114.49
110.60

-174.73

112.92
110.99

117
114.85
116.87
117.33

02n'03n+1

3.05
3.76

3.14

2.85
2.92 '
2.83
2.96

5.53
3.18
2.78
2.84
4.95

4.59
2.66
3.48

3.66
2.69
3.94
2.55

 

torsion angle: phi: 05(n)-C1(n)-O4(n+1)-C4(n+1)

torsion angle: psi: C1 (n)-O4(n+1 )-C4(n+1 )-C3(n+1)
‘t: glycosidic linkage angle: C1(n)-O4(n+1)-C4(n+1)

lUPCA-IUB 1983

137

 

 

 

 

 

Figure 111.45 Comparison of the ideal glycogen helix model, glycogen
phosphorylase —bound maltopentaose (blue,1P2B.pdb), oligosaccharides at GS Ga
(red), Gb(cyan), Gc (green), and Gd (yellow) sites.

We also overlaid the GP-bound maltopentaose, glycogen helical model and GS-
bound oligosaccharides. The sugar S3 in the GP-bound maltopentaose, sugar +3, +10,
and +15 in GS are the glucosyl units with the least interaction with protein and were
purposely chosen as the superimposition basis. A shown in Figure 111.45, there is
appreciable deviation in each sugar chain conformation compared to the ideal glycogen
helical model, which are likely due to the protein binding. Actually, none of the protein
bound oligosaccharides (n 2 5) in the protein data bank has shown the faithful glycogen

helix conformation to date.

138

111.5. E.coli GS Structure Analysis

Most bacteria transform external nutrients to glycogen for energy storage. Usually
the amount of glycogen in bacteria is huge, because it is not only an energy depot for
daily consumption, but also a reservoir that ensures the bacteria can survive any
unforeseen negative conditions. In some bacterial species the mass of glycogen accounts
for fifty percent of the dry cell weight(34). In humans, glycogen synthesis is under strict
control of “secondary messengers”, but it still can be very fast and effective. For
example, an average adult after a standard meal (approximately 700 Kcals) can soon
produce up to 100 —120 g of glycogen. Apparently, glycogen synthesis is a frequent, fast
and effective biological process. In that process, GS operates like a powerful mini
machine that adds thousands of glucose units from ADPGlc to glucan chains, one at a
time. The structural studies of E. coli GS described here address important questions like
what type of molecular organization is designed to suit such a function? How do the
substrates ADPGlc and the glucan chain bind to the enzyme? What does the active site

look like? What is the mechanism of GS?

111.5.]. Open-Close: GS Dynamic Motion

All of our E. coli GS X-ray structures display a typical twin Rossmann fold with a
cleft between two domains. However, the interdomain cleft of the E. coli GS complex is
remarkably smaller than that in the apo deS, as well as the previously reported GS
structures. Superposition of the C-terminal domain of the apo deS and the
representative thSb structure revealed that the displacement on the apex of their N-
terminal domains is as much as 9.3 A and the overall domain rotation angle is

approximately 152° according to the program DynDom(35) (Figure 111.39). Since the

139

RMSD between the N-terminal domains (the a-carbon atoms of residues 1-241) and the
C-terminal domains (or-carbon atoms of residues 251 -375 and 382-477) between apo
deS and the thS complex are only 0.679 A and 0.605 A, respectively, the domain-
domain movement is basically rigid, and we designate the conformation observed in the
thS complexes as the closed form of GS.

The E. coli GS active site is composed of residues on flexible loops between the B
strand and 01 helix emanating from the domain boundaries (Figure 111.46). Comparison of
the open and closed forms of E. coli GS explains the role of such a motion in enzyme
catalytic turnover: only after the domain—domain closure are these substrate-binding

residues brought together to construct a competent active site.

140

A
4.7.4!
. ‘ . , ’ /
‘ "’7 ‘1’" (oop 13-20 / ‘\(/‘.: "‘ ’
1 4 ‘ ~73!) EPPSo 7‘ ”G
A” "a

   

A - «71153 b»
‘ €1.83: '1 3‘74 I."
. _ . .‘ . ,
521:7; Lys305

Ii" é" A ‘

\v. «r _ > ' ‘ .. \Vlfl.>tv
{hm/‘3‘ '1 %\~A{" ‘
. 8‘ , “ \
Hisl6l _\‘ E

'v ’ loop 376-381
C

N-term C-term

Figure 111.46 Superposition of the C-terminal domains from deS (cyan) and
thSb (yellow). Ligand bound in thSb (HEPPSO, ADP and glucose) are shown
as black sticks. Structural elements essential for catalysis are colored red and their
counterparts in the open form deS are colored blue. Residues critical for

glucosyl transfer (Hisl61, Arg300, and Lys305) are shown as red sticks.

141

111.5.1 A. Comparison of ADP conformations in GS open and closed

form

In an effort to understand how the open and closed form affect the ligand binding,
we compared the binding mode of ADP in the open form with the ADP-bound AtGS

structure (20) and the closed form ADP-containing E. coli thS structure (Figure 111.47).

Lys15

  

Ile297

 
  

2.85 ' Asp21

. .
T11r381 ‘~ g
/ “Leu. 81
2.96 3.32

Figure 111.47 Structural comparison of ADP molecule in open form A.
tumefaciens GS (yellow, 1rzu.pdb) and closed from E. coli GS (blue). The residues
from A. tumefaciens GS are labeled in red and their carbons are colored yellow
while those from E. coli GS are labeled in black and their carbons are colored blue.

The interaction between ADP and protein are shown as broken lines.

142

Compared to ADP in the AtGS open form, the adenosine diphosphate group in the
closed form of E. coli thS exhibited a twisted conformation at the ribose ring and the
proximal phosphate. The ribose ring in the closed form of E. coli thS rotated
approximately 51 .8° and the ribose 3-hydroxyl group then fell within hydrogen bonding
range of the side chain of LyslS and Asp21, which were approximately 8 A away in the
AtGS open form. The proximal phosphate was in contact with Leu381 and formed two
hydrogen bonds to its backbone amide. The distal phosphate of ADP, which extensively
interacts with C-terminal residue Arg300 and Lys305, did not show the same extent of
twist as ribose and the proximate phosphate. It seems that the difference in conformation
of ADP between open and closed forms of GS is initiated at the ribose where the N-
terminal residues approach and interact, suggesting that closure of the enzyme is required

for substrate ADPGlc to adopt a catalytically competent orientation in the protein.

III.5.1.B. Large displacement of the KTGGL loop in GS open and

closed forms

The KTGGL motif is highly conserved among glycogen synthases and starch
synthases and was suggested to be involved in substrate binding (36,37). The N-terminal
KTGGL motif is located in the loop 13-20 at the N- and C-terminal domain interface,
more than 5 A away from the active site in the apo-deS open form (Figure 111.48). The
domain-domain closure brings this N-terminal loop into the vicinity of the C-terminal
domain to interact with both substrate ADPGlc and acceptor. Lys] 5 and Gly18 both
make hydrogen bonds with the ribose hydroxyl and B-phosphate oxygen, respectively

(Figure 111.22). Lys] 5 also participates in a domain-spanning hydrogen bonding network

143

with Glu357 and Tyr355, presumably further facilitating the cross- talk between the N-

and C-terminal domains.

OtsA

GS

 

Figure 111.48 Structural comparison of (bottom) loop 13-20 in apo-deS (yellow)

and thSb complex (red) and (top) the equivalent loopl4-24 in the UDP/ imidazole
/Glc-6-P bound OtsA (green, 1ng.pdb) and UDPGlc-bound OtsA (blue, 1uqu.pdb).

Gly22 in OtsA and Gly18 in GS interacting with the phosphate oxygen are shown as
sticks. Ligands other than UDP (OtsA) and ADP (GS) are omitted for clarity.

144

The last three residues of the KTGGL motif are conserved in GT-B retaining
enzymes, such as GP(3 8), MalP(21,23) and Trehalose —6-phosphate synthase
(OtsA)(3 9,40). The GGL sequence at the beginning of helix 1H1 in rabbit muscle
glycogen phosphorylase has been shown to be in van der Waals contact with heptulose-2-
phosphate located at the enzyme binding site of the natural substrate, glucose-1-
phosphate(38).

OtsA transfers glucose from UDP-glucose to glucose-6-phosphate to form cad-1,1
trehalose-6-phosphate(39,40). The OtsA GGL motif is also on the loop, 14-24,
structurally equivalent to the E. coli GS loop 13-20 (Figure 111.48).

In the substrate UDPGlc-bound OtsA structure, OtsA loop 14-24 is partially
disordered (14-19) while the rest of the loop (20-24) is 7-9 A away from the substrate
UDPGlc (40), reminiscent of the comparable distance between the GS loop 13-20 and
ADPGlc seen in the open apo-deS structure(Figure 111.48). In the presence of the
substrate analogue UDP/imidazole and acceptor Glc-6-P, the entire loop 14-24 of OtsA
becomes structured and adjacent to the UDP molecule where the GGL motif interacts
with the substrate donor and locks the substrate into position for the glucosyl transfer
with its Gly22 (Gly18 equivalent) hydrogen- bonding 05* and B-phosphate oxygen of
UDP (39) (Figure 111.48).

In conclusion, OtsA manages to encapsulate the active site with the GGL motif via
local conformational changes as evidenced by the small N-terminal domain displacement
(~1-1.5 A). In contrast, E. coli GS undergoes a large global domain-domain rotation (9.48

A N-terminal displacement) to move the GGL motif into the interdomain catalytic center.

145

111.5.].C. What causes the enzyme to close?

There are two major forms of E. coli GS, an open form, where the two domains are
spread relatively far apart, and a closed form where the two domains are very close
together. The apo-deS structure adopts the open form while all ligand-bound E. coli
GS structures represent the catalytically active, closed form of the enzyme (Figure
111.46).

Attempts to crystallize GS in its catalytically active, closed form have been made
by several groups over the past five years before we first achieved this goal and obtained
not only one, but several ligand-bound closed form GS structures. The one thing in
common in those closed-form GS structures is the presence of an acceptor analogue
HEPPSO or an oligosaccharide in addition to ADP, or ADP and glucose. In the absence
of an acceptor or acceptor analogue, only the open form has been obtained both in our
and other labs. For instance, incubation of AtGS with substrate ADPGlc alone resulted in
the open form with only ADP visible, which binds to the C-terminal side of the
interdomain cleft (20). Incubation of E. coli deS with ADPGlc results in a vacant
interdomain cleft and wide-open conformation.

HEPPSO was added as a buffer in GS crystallization trials, but turns out to appear
in the active sites of four closed thS structures, suggesting that HEPPSO is critical for
capturing the closed form of E. coli GS. Although ADP, glucose and HEPPSO all bind in
the GS interdomain cleft, HEPPSO exclusively interacts with the N-terminal domain
whereas ADP and glucose predominantly interact with residues from the C-terminal

domain. However, the interaction between HEPPSO, ADP, and glucose, directly and

146

through the KlsTGGL motif, establishes the cross-domain network which brings the two

domains together (Figure 111.49).

Almost equally important in catching the closed form of GS is the impotency of

HEPPSO for glucosyl transfer; mostly due to the distance between the HEPPSO tip

hydroxyl group and the anomeric carbon of glucose (3.3 A). If HEPPSO were able to

approach the glucose of ADPGlc closely enough and perform a nucleophilic attack, the

GS machinery would start to run incessantly with frequent opening and closing, until all

of the substrate ADPGlc is depleted. Then it would be extremely hard to maintain a

closed form on the time scale of crystallization.

ll()

 

Leu19
Asp137
Trp138

 

 

 

{/3 /P§O

OH

HO
OH

34
-o./° o ”

OH

N

 

 

Lys15
Gly18
A$p21
Arg300
Ly5305
Gly354
Tyr355
His356
Leu381

7

 

Thr382

 

 

His161Asn162 Asn246

Gln304 Glu377 Cy5379 Gly380

 

N

NH,

6
2\

Figure 111.49 Schematic diagram showing the cross-domain network

between ligand HEPPSO, glucose and ADP, which chiefly interacts with

the N-terminal (red) and C-terminal (blue) residues, respectively.

147

In the oligosaccharide-bound GS E3 77A structure, the oligosaccharide that
replaces HEPPSO in the interdomain cleft also predominantly interacts with N-terminal
residues (Figure 111.36). The connection between this oligosaccharide and the C-terrninal
domain is mainly mediated through the ADP phosphate group and the KTGGL motif. It
is noteworthy that there is no electron density for the glucose moiety of ADPGlc or an
individual glucose molecule from ADPGlc. Due to such a lack of the sugar to be
transferred, the glucose addition is not able to launch and the enzyme remains in the
closed form.

In summary, our success in capturing the closed form of GS should be attributed
to the simultaneous presence of ADP and oligosaccharide acceptor, or ADP, glucose and
the acceptor analogue HEPPSO. The cross-domain network spanning the N-terminal
residues, oligosaccharide acceptor/ HEPPSO, ADP / ADP+glucose, and the C-terminal

residues provides internal bonds for GS to keep the two domains closed.
[11.5.]. D. Domain -wise ligand binding facilitates GS open-close motion

The structural analysis of the open form apo-deS and closed form thS-ADP-
glucose-HEPPSO complex as well as E377A-ADP-oligosaccharide complex suggests
that domain-domain closure forms a competent active site in the center and domain-
domain opening facilitates the product release. As a result, domain-domain opening and
closing is believed to accompany GS glycogen chain production.

We notice that ligand ADP and glucose are predominantly bound by the C—terminal
domain whereas acceptor oligosaccharide (and acceptor analogue HEPPSO) binds almost
exclusively to the N-terminal domain, either in the interdomain cleft or on the surface.

Such a domain-wise ligand binding facilitates the GS open-close motion because it

148

prevents the constraints that would result from a two-domain spanning bound glycogen

polymer.

III.5.1.B. Substrate binding order

As ADP/Glc is deeply buried in the GS interdomain cleft, one would think that the
binding of glucosyl donor ADPGlc must precede the binding of the acceptor glucan
chain, which occupies the channel. However, the GS open form presents another
possibility regarding substrate-binding order. Both ADPGlc and glucan chain could bind
to the enzyme’s open form without constraint on their binding order. Only when both
bind does closure and formation of the active site happen. This scenario allows
independent binding of substrates and does not require a specific binding order, therefore
avoiding potential substrate inhibition (where the 2nd substrate binds first and must
dissociate before the other binds).

Precedence exists for both the ordered bi-bi mechanism, in which ADPGlucose
binds first, followed by the acceptor and the random kinetic mechanism, in which both
donor and acceptor bind to enzyme without a specific order. Structural and calorimetric
binding studies on GT-B retaining enzyme N-acetylgalactosaminyltransferase (GTA) and
galactosyltransferase (GTB)(41) suggested that the binding of their donor facilitates the
formation of the acceptor-binding site and therefore it occurs prior to acceptor binding.
0n the other hand, the extensive analysis of the GT-B retaining enzyme GP reveals that it
employs a rapid equilibrium, random kinetic mechanism, in which both donor and
acceptor substrates must be bound prior to the chemical event but without constraint on

their binding order (reviewed in (42)).

149

Although there is no experimental evidence to rule out or support one over the
other regarding the GS substrate binding order, the competitive inhibiting effect of
acceptor analogue HEPPSO to ADPGlc (Ki = 0.15 M) suggests that the acceptor binding
somewhat weakens substrate ADPGlc binding to GS, indicating ADPGlc might bind to
GS prior to the acceptor.

Interestingly, the interdomain-bound acceptor analogue HEPPSO shows
substantially low inhibiting effect to glycogen (Ki= 0.8M), indicating the majority of
glucan may bind outside the active site where they are not disturbed by the presence of
HEPPSO. Such a hypothesis is consistent with the mutagenesis studies, which show that
mutating residues that interact with the interdomain bound oligosaccharide such as Tyr95

or Hisl39 to Ala does not lower the GS affinity to glycogen (Yep unpublished result).

111.5.2. Important Residues

Previous mutagenesis and kinetic studies identified several residues that are
significant to GS activity (Table 111.6) (11,22). Now, based on our GS-ligand complex
structures, we are able to illustrate their particular roles in GS catalyzed glucan chain

elongation.

150

Table III.6 Kinetic parameters of wild type glycogen synthase and mutants

 

 

Decrease Km(ADPGIc) Residue roles suggested
V.,,“ (Ulmg) (-fold) 11111 by our 68 structures
WT 500 :l: 65 1 18:3
D137A 0.07 :1: 0.01 8140 2114 holding +1 Glc in place
H161A 0.8 t 0.1 710 200116 positioning -1 Glc and
stabilizing intermediate DGM
R300A 0.22 :t 0.02 2590 150115 acid catalyst
making ADP a better leaving group
K305A 0.46 :t 0.05 1240 3215 acid catalyst
making ADP a better leaving group
E377A 0.05 t 0.01 10,000 70:30
E3770 0.020 :t 0.001 25,000 61112 holding -1 Glc in place for transfer
E377D 10 1: 1 57 1500:300

 

111.5.2.A. Aspl37 is the +1 sugar - positioning residue

Mutating Aspl37 to Ala is an almost lethal mutations of GS as the enzyme activity
drops 8,140- fold (22). In the interdomain catalytic cleft of the oligosaccharide-bound
E3 77A structure, Aspl 37 side chain 0D2 makes bidentate hydrogen bonds to the +1
sugar OH-2 (2.44 A) and 0H-3 hydroxyl (2.76 A) (Figure 111.36). As the +1 sugar is the
immediate acceptor of the substrate ADPGlc glucose, its orientation is critical to the
transfer efficiency and the bidentate hydrogen bonds from Aspl 37 play a critical role in

correctly positioning the +1 sugar.

151

111.5.2.B. His161 participates in —1 sugar positioning and intermediate

DGM stabilization

Hisl6l forms a 2.84 A hydrogen bond between its side chain NDl to the 6—
hydroxyl of Glc in the thSb structure (Figure 111.26) and a 2.69 A hydrogen bond to
DGM 6-0H in the thSa structure (Figure 111.31). This hydrogen bond plays an
important role in positioning the —l Glc as mutating His to Ala at this position caused a
710- fold decrease in enzyme specific activity (22). The Hisl61 backbone carbonyl group
is only 2.85 A from the DGM C1, presumably offering electro static stabilization to the

positively charged intermediate DGM.

III.5.2.C. The catalytic Lewis acid Arg300 side-chain switches in and out

of the GS active site in response to the presence of ADP

Mutating Arg300 to Ala decreases GS specific activity 2,600 fold (22). The ADP,
glucose-bound thS structure reveals that Arg300 is close to the phosphate group and
probably acts as a Lewis acid to donate a proton to the ADPGlc phosphate to facilitate its
departure from the glucose moiety (Figure 111.22).

A virtually identical conformation of Arg300 where its side chain is proximal to
the ADP phosphate was seen in almost all GS structures with ADP or an ADP moiety
bound, including the ADP-AtGS complex (20). In contrast, when there is no ligand in the
active site as in the apo-deS, apo-AtGS, and apo-PaGS structures, the Arg300 (Arg299
in AtGS; Arg257 in PaGS) side-chain is flipped out of the active site. It rests on the loop
327-331 and is stabilized by the hydrogen bond from the Ala329 backbone amide and

several Van der Waals contacts from Ala329 and Gly330 (Figure 111.50).

152

Apparently, the Arg300 side chain switches in and out of the GS active site in
response to the presence of the ligand ADP, particularly the distal phosphate. However,
in the ADP-bound E3 77A structure several water molecules occupy the space between
ADP phosphate and Arg300. As a result, the Arg300 side chain flipped away from the
active site and the distance between Arg300 NH2 and phosphate 01 B is approximately

10.4 A (Figure 111.32).

153

 

Figure 111.50 In apo-E.coli deS (cyan, colored by atom), apo-AtGS (1rzv.pdb,
ruby), apo- PaGS (2bis.pdb, orange) and E377A-ADP-HEPPSO- complex
(magenta). The side-chain of Arg300 and its equivalents are out of the GS active site
and are in contact with Ala329 and Gly330 (Lys in PaGS). 1n ADP, Glc-bound
thS (2qzs.pdb, green, atom-wise colored), oligosaccharide-bound E377A (blue),
and ADP-bound AtGS (1rzu.pdb, yellow), Arg300 and its equivalent side-chain are
close to the ADP phosphate and their interaction are shown as dotted lines (black
lines for thS and red lines for AtGS complex). Hydrogen bonds between the
Arg300 side-chain and Gly330 in apo-deS, and the Arg300 side-chain and
maltotriaose in oligosaccharide-bound E3 77A complex are also shown as dotted

lines. Residues are labeled according to the E. coli GS sequence.

154

111.5.2.B. Catalytic Lewis acid residue Lys305

Lys305 is located close to the distal phosphate and is a potential candidate for
protonation of the phosphate to make it a better leaving group (Figure 111.24). Consistent
with this supposition is the fact that mutating Lys305 to Ala resulted in more than a
thousand fold decrease in specific activity (22). Equivalent lysine 305 and arginine
300residues occupy identical positions relative to the phosphate in all available GT-B
retaining enzyme structures (GP (43), MalP(2asv.pdb) , OtsA (40), AGT (44)) (Figure

111.51) that do not require a divalent Mn2+ ion for their enzymatic action. It therefore

seems likely that nature either uses an exogenous divalent metal ion as for GT-A
enzymes, or positively charged side chains in the majority of GT-B enzymes to make

phosphate a better leaving group.

155

PLP

 

 

Hisl61

Figure 111.51 Structural comparison of ADP, glucose binding site and active site residues between

E. coli GS (in yellow and atom colored, thSb), MalP (blue, 2asv.pdb), rabbit R-GP (pink,
1gpa.pdb), OtsA (green, 1uqu.pdb) and AGT (cyan, 1y6f.pdb). MalP Ligand, PO43', ASO (1,5-
anhydrosorbitol) and maltopentaose, occupies the equivalent positions of the ADP distal phosphate
group, glucose and HEPPSO in the thSb structure, respectively. Maltotriaose in E377A complex
was also shown (red sticks). Residues are labeled according to the E. coli GS sequence. Top: the front
view. Bottom: the back view.

156

111.5.2.B. Glu377 is responsible for the —1 sugar positioning and helps

Lys305 maintain positive charge

Glu3 77 deserves special attention because mutation at this position caused the most
deadly effect to GS specific activity (E377A: 10,000 fold decrease; E3770: 25,000 fold
decrease). Glu3 77 is widely conserved in MalP, GP, trehalose phosphorylase, N-
acetylglucosaminyltransferase (GTA), galactosyltransferase (GTB) and OtsA (equivalent
residue Asp) (Figure 111.51) and is the first residue in the E-X7-E motif of some retaining
glycosyltransferases, such as eukaryotic glycogen synthases (GT3), 01-N-
acetylglucosarninyltransferase (GT4), starch synthases, bacterial glycogen synthase
(GT5) and a-mannosyltransferase (GT15). Substitution of E. coli GS Glu3 77 by
Gln/Ala(11), MalP Glu637 by Gln(45) , and GTB Glu303 by Ala (41) all result in a
drastic reduction in enzyme activity. Mutating Glu510 (the E377 equivalent) completely
inactivated the human muscle glycogen synthase (46). Glu377 and its equivalents were
therefore thought to be either the catalytic nucleophile or the general acid/base catalyst
for these glycosyltransferases (11,41,46-48), however we found that Glu377 is located on
the or-face of the glucose to be transferred and is relatively far both from the C1 glucose
position and from the ADP B-phosphate position in our ECGS structures, which is
inconsistent with Glu3 77 playing a direct catalytic role in the reaction. Positively
charged Arg300 and Ly3 05 interact with the substrate phosphate group (Figure 111.22),
and are therefore much more likely to play the role of a proton transfer agent than
Glu377. In fact, the Glu377 carboxyl oxygen atoms are 2.93 A from Lys305 (Figure
111.52). Perhaps Glu3 77 is deprotonated and negatively charged which would help the

neighboring Lys305 maintain its positively charged side chain. More importantly, Glu377

157

side-chain forms a short hydrogen bond to the glucose 3-hydroxyl and plays an important
role in positioning and stabilizing the —1 glucose (Figure 111.52). Our E377A-ADP-
HEPPSO and E377A-ADP-oligosaccharide structures both show that the glucose moiety
is missing in their GS active sites. As shown in Figure 111.52, an extensive hydrogen bond
network is present between Glu3 77, catalytic residues Lys305 and Arg300, and glucose.
Each of Glu3 77 carboxyl side-chain oxygen accepts two hydrogen bonds. Replacing
either one of them to nitrogen which has one lone pair and can only accept one hydrogen
bond, would seriously disrupt the whole network, which perfectly explains the dramatic
enzyme activity decrease (25,000 fold) observed in mutant E377Q. To understand why
the enzyme retained the appreciable enzyme activity when Glu3 77 was mutated to Asp,
we modeled Asp at the position 377 in the conformation of the OtsA Asp361 (equivalent
to GS Glu377) and find that Aspl37 is still able to interact with glucose and hold it in
place (Figure 111.52). Due to the shorter side-chain of Asp, the interactions between the
position 377, Gln304, and Lys305 observed in the thS structure are lost, which
probably leads to the 57- fold activity decrease in the mutant E3 77D.

Taken together, Glu377 positions and stabilizes the —1 glucose and is an extremely
important residue to GS catalysis. Although Glu3 77 was proposed to play a direct
catalytic role and our structure does show that Glu377 contributes in the maintaining the
positive charge of catalytic residue Lys305 which makes ADP a better leaving group, its
glucose locator role is dominating and has an overwhelming impact to the enzyme

activity.

158

Arg300

  
 
    

glucose

Ser374

Gly380

    

Ser374 Asp37 Cys379

Figure [11.52 Top: The Glu3 77 environment in the wild-type GS structure. Arrows
indicated the hydrogen bond proton donating direction. Bottom: The interaction between
the modeled Asp377 and neighboring residues.

159

111.5.3. GS Complex Structures Support SNl Mechanism Versus The
Double Displacement Mechanism

III.5.3.A. DGM intermediate

Our three separate high—resolution thS structures (thSb, thSc, thSd) from
12 week incubations with ADPGlc have shown individual glucose at a superimposable
position in the active site. In a crystal harvested in a shorter time, 4 weeks, we see density
for most of the glucose moiety, but do not see density for the 04 glucose oxygen or for
the 01 glucose oxygen. We therefore believe that we have captured some intermediate
form that will subsequently be hydrolyzed to the glucose we see in the above-mentioned
three structures. Interestingly, in the active site of MalP which is also a GT-B retaining
enzyme and shares an almost identical active site as E. coli GS, there exists the density for
a similar glucose-like species lacking a l-hydroxyl group at the superimposable position
of glucose in our thSb, thSc, and thSd. Although that density was assigned to 1,5-
anhydrosorbitol (2ASV.pdb), the oxocarbenium ion DGM would fit the density just as
well and hints that such a species is also formed in the thSa active site.

A DGM intermediate is one essential component of an SN] mechanism that was
first suggested for the GT-retaining enzyme glycogen phosphorylase (GP) about twenty
years ago(49). The observation of DGM and its hydrolyzation product glucose in the
E. coli GS active site as well as DGM in the MalP active site strongly suggest that those
GT-B retaining enzymes, which share active site structure, proceed through an SN]
mechanism. Consistent with our hypothesis, several oxocarbenium cation mimics have
been shown to strongly inhibit GT-B retaining enzymes such as E. coli GS (50),

Schizophyllum commune trehalose phosphorylase (51-53), rabbit muscle glycogen

160

phosphorylase(52,54,55), and starch phosphorylase (26). Moreover, the kinetic behavior
of various oxocarbenium cation mimics on Schizophyllum commune trehalose
phosphorylase suggested that the positive charge of the oxocarbenium ion is focused on

C 1 rather than 0-5, which is consistent with the requirement for GT glucose transfer (52).

III.5.3.B. DGM stabilization in GS

Like all carbocations, DGM is electron deficient and has long been considered as
an unstable, transient species. The estimated half-life of DGM in open solvent is on the
order of pico -seconds, slightly longer than bond vibration times(56,57). The extreme
instability of DGM was the major obstacle for the scientific community to accept an SN]
mechanism in the past. But today, our thS complex structures point out that there are
multiple interactions to stabilize DGM and it is possible for DGM to live much longer
when bound in the active site of GT-B retaining enzymes than alone when exposed to
solvent.

The most prominent electrostatic stabilization of DGM is provided by the substrate
phosphate moiety whose oxygen 03B is 3.26 A and 3.14 A to glucose C1 and 05,
respectively in the thSa structure (Figure 111.32). In fact, a stronger inhibition and
tighter binding was observed when oxocarbenium cation mimics were tested with
phosphate, such as D-gluconic acid 1,5-lactone (GL) and phosphate (26) , and
nojirimycin tetrazole and phosphate (55). This again suggests that the nearby leaving
group phosphate is involved in stabilizing the DGM intermediate.

In addition to the substrate ADP moiety, the protein itself also contributes to

stabilize the intermediate DGM, particularly through residue Hisl61. The Hisl61

161

backbone carbonyl group is only 2.85 A from the positively charged DGM C1,

presumably offering electrostatic stabilization to the intermediate DGM.

III.5.3.C. Effort to verify the existence of carbocation DGM using

NMR

C. K. Ingold and E. H. Huphs first proposed that a carbocation is the key
intermediate in unimolecular nucleophilic substitution (SNl) and unimolecular
elimination reactions (E1). Their detailed kinetic, stereochemical and product
investigations suggest that formation of carbocations is the slow rate-determining step in
those SNI and E1 reactions. Later, the carbocation concept was generalized to many other
organic reactions and its significance in organic reaction mechanisms are well
recognized. However, due to their electron deficient nature, carbocations are extremely
reactive towards surrounding molecules e. g. solvent molecules or negative ions
(nucleophile). The direct observation of a stable, long- lived carbocation was so much of
a challenge that it did not become possible until the highly acidic superacid chemical
system was developed. In 1962,George A. Olah used tertiary butyl fluoride to react with
the superacid antimonpentafluoride (SbFs) at —78 °C and obtained the first stable, long -
lived tert-butyl cation. That tert-butyl cation can stay at —78 °C for many hours, long

enough for both spectroscopic and chemical study.

162

 

CH -78OC CH
A3 + SbF5 . l 3 + SbFG-
'H/H/ 0+
H30 0113 H3C/ \CH3

Superacids are acids stronger than 100 % sulfuric acid. Superacid is the key to
obtain stable, long-lived carbocations because their extremely low nucleophilicity
suppresses the deprotonation of alkyl cations to olefins. If deprotonation took place, the
carbocation (a strong acid) would immediately react with olefin (a good 1: base) and
generate complex mixtures through a variety of reactions, such as alkylation,

oligomerization, polymerization, and cyclization.

(CH3)3C+ = H++ (CH3)2 C=CH2

Since the remarkable role of superacid in stabilizing carbocations were revealed by
George A. Olah, who later won the Nobel Prize in 1994 for his work on carbocations and
superacids, thousands of carbocations have been investigated and characterized. The
applied physical methods include conductivity, UV-vis spectroscopy, 1R, laser Raman, X-
ray diffraction, NMR (nuclear magnetic resonance), and ESCA (electron spectroscopy for
chemical analysis). Because of the rehybridization from sp3 to sp2 and the effect of

significant positive charge, the carbocation carbon nucleus is greatly de-shielded and

163

generally displays a large chemical shift. Therefore, 13 C NMR, which was used to
characterize the first stable alkyl cation by George A. Olah in 1962, is still one of the
most straightforward techniques to detect the carbocation in solution.

In an effort to verify the existence of DGM in GS, 13‘C and lH-BC 2D NMR
analyses were employed with GS- [1-13 C] ADPGlc-HEPPSO complex. The GS- [l-13 C]
ADPGlc-HEPPSO complex displayed a peak at 177.98 ppm which was not seen in GS,
[l--13 C] ADPGlc, or HEPPSO (Figure 111.52). Although the chemical shifts of most
carbocation carbon center is in the range of 200 -320 ppm, the positive charge on DGM
Cl is partially delocalized to the neighboring 04 which may render its chemical shift
lower than regular carbocations. The 177.98 ppm peak was strong and actually was the
only candidate for the DGM positively charged C1.In the [1-13 C] ADPGlc titration
experiment, the peak at a similar position, 177.74 ppm, rose as the collection time
increased (Figure 111.53). To prove our supposition, we synthesized uniformly l3C—
labeled ADPGlc to explicitly identify the source of the 177.98/177.74 ppm peak through
tracing glucosyl unit carbons in a two-dimensional lH-l3 C spectrum. The NMR scanning
range was purposely set between —10- 400 ppm, covering all possible peak regions. This
choice, however, turns out not to be sensible, as even long collection time as 36 hours
was not able to provide sufficient data and decent resolution. We were also not able to
reproduce the 177.98 ppm peak with the second batch of [1-13C] ADPGlc. The second
batch of [1-13‘ C] ADPGlc was made following exactly the same procedure as the first
batch [l-'3 C] ADPGlc, except that HEPES in the enzymatic synthesis and the Tris in the
HPLC separation replaced bicine and ammonium formate, respectively. Bicine and

ammonium formate exhibit peaks between 160-190 ppm whereas HEPES and Tris '3 C

164

NMR peaks are all between 50 -60 ppm and won’t interfere with the high chemical shifi
region where the DGM cation Cl peak may arise.

Although the NMR experiments we carried out did not offer evidence of long-lived
DGM in the GS complex, the possibility to directly observe the long-lived carbocation
DGM in the GS complex in the future still remains. Dramatically increasing GS solubility
would allow more DGM in one NMR tube and using a more advanced NMR

spectrometer and probe could be the solution.

165

ES
omfimm+mo+2omQ$§n:d

a? w: 3; mi ow~ mm: wwfi owfi wwfi 03 mg

L Fp-——-h-—-PLn—nPb-L+Ll-thF-P—-Pb—F—brrbrhpr—Lbbh—h-

cad: ow.ww~

 

 

m: w: my: mi ow“ lewd #2 o: me 09 me

.,.._ 424:. __ ya. .4.“fi.._fi-4‘1.1_1_l.11.21:1311ijufl _ A 1521:344.‘fi9:4. 3214544.; :4. fl: ,2);
_ _ fi . _

 

: €333.11:__,.:j:,_fiJfiB, _ __ 1;- .:.-.,.1._

  

_
_

Em
motammfiamnzd

6388 Omfimmzol 223-82-: 2a 5388 mo: somerfiai .6 880% ~32 m3: 25H...

166

.352 2% ago-:6 a 5388 a: 23 20.23.62. : 2 :33 .8 «50% 222 «3: 2&2

02 ow~ o2

h-PbPD—FPF+hPL-h—DL-th-bbhbb-

mm. m: 02*

   
 

o: 03 ca

PPthb—uppphpbhh—bbth-P+h—hbn

11.4.-1114...

o2 o2 02
«kg

at ow~ 03

hbh-PPFthP++hhP—+thhPLthtL

NQwE _

2029.. -2 02-: 23am

.5 mm 229.13 2+ OmEmEmor-gomnz

E a swan-«.3 2+ ommmm22+mo+2omn<

2.: m 223% 2+ ommmmm+mo+2oa<

be: 229:8 .3 a
08: 22282—8 .E : OmmmmE+mO+20mQ<

167

.Ammmmv @5888 9:3ch .28 emu-.9859 Ehooqm Soc @896“ Pa «.50on
325 siege“ as... WE EH 66%...-622 : ea 2 :38 Mo 8:82.22 a £023 .6 Exam 222 m3: 2%:

 

 

 

 

 

 

 

 

 

 

 

 

 

 

o 8% com o Ema oom
2.5m
$63 9.3
.5 3 o m
A m a 2 a a e 223%
Baa-20m EonEE< 2.;
o . Ema . m2
modm $65
8% 3232 :35 822m
o Ea m:
3.3 3.3
E25 2 fiaa 8%:
3% K 4

wmdm

168

III.5.3.B. The substrate phosphate group deprontonates the incoming
acceptor nucleophile

The oligosaccharide-bound E3 77A structure shows that the B-phosphate OBB is
2.62 A from the 04 of the glucan chain acceptor, ideally poised to deprontonate the
incoming nucleophile 4-hydroxy group (Figure 111.36). A similarly close distance was
also spotted between 033 and the acceptor analogue HEPPSO hydroxyl-ethyl group 04
(2.76 A) in the thSb structure (Figure 111.26), also indicating that the phosphate group
is involved in the deprotonation of the incoming acceptor nucleophile. This rationale
explains why GS activity increases as the buffer pH goes up (from 6.8 to 7.5) because the
high pH may promote the phosphate to deprontonate the acceptor hydroxyl group and
facilitate the acceptor nucleophile attack (50). Similarly close distances between the
incoming nucleophile hydroxyl group of the acceptor sugar and the phosphate oxygen of
the leaving group is also seen in the GT-B retaining enzymes OtsA (39) and MalP(21).

In addition, the structures of GT-B retaining enzymes including E. coli GS all
display a very similar overall disposition of the ADP/U DP and glucose in their active
site. The phosphate moiety takes the unusual “tucked under” conformation relative to the
ADP /U DP moiety which could provide some ground state destabilization as well as

assisting acid catalysis of glycosyl transfer.
III.5.3.B. Evidence against the double-displacement

The double displacement mechanism was previously proposed for retaining GT
enzymes with a direct reference to GH enzymes(5 8-61). It involves a nucleophilic attack
from the B-face of the sugar donor and formation of a covalent glucosyl — enzyme

intermediate (Figure 1.12). Attempts to verify the double displacement mechanism for

169

GT-B enzymes, mainly to identify the catalytic nucleophile and to trap a covalent
intermediate, have never been successful (62). Here, our thS complex structure
presents some evidence against the double-displacement mechanism.

The catalytic nucleophile is an essential component of the double-displacement
mechanism. For GS, Glu3 77 and Hisl61 had long been proposed as potential candidates
as the catalytic nucleophile in the double-displacement mechanism. However, in our I
thS complex structures Glu3 77 is located on the a-face of the glucose to be transferred
and is relatively far both from the C1 glucose position and from the ADP -phosphate
position, which is inconsistent with Glu3 77 playing a direct catalytic role in the reaction.
Actually, our mutant E377A/ADP/HEPPSO structure reveals that the glucose moiety of
ADPGlc is missing in the enzyme's active site, indicating that Glu377 is in fact critical
for positioning the glucose properly in the active site (Figure 111.52). Moreover, the
Glu377 side chain two carboxyl oxygen are equally 2.93 A to Lys305 and likely play a
role in maintaining the positive charge on the catalytic residue Lys305.

The main chain of E. coli GS Hisl 61 and its structural equivalents in other
retaining GT enzymes, such as His345 in maltodextrin phosphorylase (MalP), are located
at the B-side of the sugar and close to the anomeric interaction center. Although amide
groups are not ideal nucleophiles, such a function is not without precedence (63). To
design an experiment to test whether it is the catalytic nucleophile is difficult because it is
located in the main chain of the protein. Therefore, although the double replacement
mechanism seems not plausible according to the available GT-B retaining enzyme
structures, including our E. coli GS structures, and kinetic studies, it should not be totally

ruled out.

170

[11.5.4 The GS Catalytic Scenario Suggested by E.coli GS Structural

Studies

Based on our structural studies of E. coli GS and its mutants, we propose the GS
catalytic mechanism as follows. Firstly, GS adopts the open form with a wide-open
interdomain cleft where the sugar donor ADPGlc and the sugar acceptor glucan chain
binds to the C-terminal domain side and the N-terminal domain side, respectively. Then,
the concurrence of sugar donor and acceptor stimulates the communication between the
N- and C-domains and the enzyme closes. In the closed conformation of the enzyme, the
basic residues Arg300 and Lys305 promote the phosphate-sugar bond break by
protonating the ADP phosphate or stabilizing negative charge on the substrate ADP
moiety. The Hisl61 main chain (C=O) and the nearby substrate phosphate group stabilize
the resulting DGM intermediate. The phosphate group also deprontonates the incoming
acceptor nucleophile, which in turn attacks the partially positively charged DGM
anomeric carbon C1 from the a-side. Finally, the substrate sugar is transferred to the

acceptor and GS opens up to release the product ADP and prepare for the next cycle.

III-6 Mechanistic Implications for Starch Synthase

III-6.1 Plant Starch Synthases Share A Similar Active Site And

Catalytic Mechanism With Bacterial GS

Like bacterial glycogen synthase, plant starch synthases also use ADPGlc to
elongate a -1,4 -linked glucan chains with retention of configuration. Both of them have

a GT-B fold. Plant starch synthase (GBSS, $811 and 88111) and bacterial GS share 29 -

171

34% sequence identity. We carried out a multiple sequence alignment on GSs, GBSS,
8811 and SSIII from a variety of organisms and identified a number of conserved
residues.

Almost all the residues in the E. coli GS active site have identical counterparts in
SSs (Table 111.7). They include the KTGGL motif, the glucose locator Glu3 77; the
catalytic residues Lys305 and Arg300, DGM intermediate stabilizer Hisl61, residue
Tyr95, Aspl37, and Trp138 that orient and bind to the glucan chain acceptor in the
interdomain clefi. Moreover, Many of the buttressing residues of the active-site residues
are also conserved in all SSs, such as Glu9 (interacting with KTGGLThrl6 and Tyr95),
and Gly400 (interacting with Glu377). Some of the conserved residues in SSs have been
experimentally proven to participate in starch synthase catalysis, such as the KTGGL
element Lysl 5(64), Asp21, Aspl37 and Glu377(65). In addition, previous studies
suggest that one arginine residue is involved in the ADP reaction as the arginine-specific
reagent phenylglyoxal inhibited $8113 but the repression can be recovered with a higher
concentration of ADPGlc(66). Our sequence alignment points out that the arginine
residue is probably the E. coli GS Arg300 equivalent: maize SSIIa Arg551. Taken
together, GT-B retaining bacterial GS and plant starch synthase share a closely similar
active site and are very likely to employ a similar mechanism to catalyze glucan chain

elongation.

172

Table 111.7 Selective conserved residues in bacterial GS and plant starch synthases.
Those residues critical to GS activity and their equivalents in 885 are colored pink.
Multiple sequence alignment was conducted with DNASTAR. GS sequences used were:
Escherichia. coli GS (POA6U8), Agrobacterium tumefaciens GS (AADO3474),
Pyrococcus abyssi GS (N P_125769), Sequences of granule-bound starch synthases from
various organisms were those of barley (AAL77109), maize (PO4713), potato
(CAA41359), and rice (P19395), Other starch synthases used were those of maize SSI
(AAB99957), potato SSI (P93568), wheat SSI (Q43654), wheat SSIIa (BAE48798),
maize SSlIa (AAS77569), potato SSII (CAA61241), potato SSIII (Q43846), wheat SSllI

(AAF87999) and Chlamydomonas reinhardtii SS (AAC17971).

173

Table III-7

 

 

 

 

 

 

 

   

 

H N 246 G E.coli GS
Q N 246 G AtGS
H R 217 G PaGS
H N 346 G barley GBSS
H N 347 G maize GBSS
H N 351 G potato GBSSI
H N 353 G rice GBSS
H N 401 G maize SSI
H N 401 G potato SSI
H N 408 G wheat SSI
H N 562 G wheat SSlla
H N 492 G maize SSlla
H N 530 G potato $811
8 N 986 G potato SSIII
S N 1385 G wheat SSIII
H N 416 G Chlamydomonas
reinhardtii SS
{"1
30016 i L 373 PSRF E PCGLTQL 398 TG G L E.coli 68
30018 1 L 372 PSRF E1PCGLTQL 397 TG 6 L AtGS
25732 F 335 PSYF1EL1PFGLTQL 360VG Gs PaGS
40112;", L 475 TSRF s1PCGL IQL 500 TG 1611 barley GBSS
40213 L 477 TSRF E; PCGL IQL 502 TG G L maize GBSS
40611}. L 480 PSRF E PCGL IQL 505 T6 G L potato GBSSI
408R L 481 PSRF-E PCGL IQL 506 T6 G rice GBSS
45513. L . 528 PSRFE PCGLNQL 553 TG 165" L maize SSI
4551? L 510F 528 PSRFE PCGLNQL 553 T6 1GifL potato SSI
46212. L 517 F' 535 PSRF1L1E PCGLNQL 561 TG 6 L wheat SSI
621R L 676 F 694 PSRF1E1 PCGLNQL 719VGG L wheat SSlla
5511? L 606 F 624 PSRFvE PCGLNQL 649VG 6 L maize SSIla
5891-? L ~ 644F' 662 PSRF1E1PCGLNQL 687VG G L potato SSII
10401}? L 104401K G 11001Y1 1118 PS IF E PCGLTQL 1143TG G: L potato SSIII
14391? L 14430 G 14991Y§ 1517 PS IF EiPCGLTQL 1542TG:61 L wheat SSIII
4703. L 4740 G 526M 544 PSMFE, PCGLTQL 569TG {31L Chlamydomonas
reinhardtii SS
III-6.2. Glucan Chain Binding In Starch Synthases

Multiple sequence alignment reveals that the glucan chain binding residues in the

interdomain catalytic cleft are strictly conserved in SS, such as Tyr95, Aspl37, and

174

Phe138. None of the surface binding residues especially in the Gb binding site is
conserved in SSS. However five out of eight binding residues at the Ge site used their
backbone amide group to lock the oligosaccharide. The binding residue, Lysl99, was not
conserved throughout the SSS, but was conserved in one SS subfamily, GBSS. Taken
together, we believe that the glucan chain may not bind to SSS on the surface in exactly

the same way as to GS.

III.6.3. Insights into GBSS and SS Specificities

Granule-bound starch synthase (GBSS) and soluble starch synthase SS (SSI, SSII
and SSIII) are two major isoforms of starch synthase. GBSS preferably catalyzes
transglycosylation reactions with long, linear amylose as the product, whereas the
homologous soluble SS (SSI, S811 and SSIII) produces short, more branched
amylopectin(67,68). One interesting feature of GBSS is that it can be activated by
amylopectin but soluble SS cannot (69). In addition, GBSS and soluble SS exhibit quite
a few different properties, such as affinities for ADPGlc and glucan substrate,
thermosensitivity and the processivity of glucan chain extension.

In an effort to find out the determinants of GBSS and SS specificities, a series of
potato GBSS and 8811 chimeric and truncated proteins were tested(70). It turns out that
the C-terminal region of GBSS confers most of the specific properties of GBSS. Based
on our GS structure, this C-terminal region (a13- (118) spans the two domains of the
enzyme and defines a surface of the enzyme that is quite far from the active site (Figure
111.56). AS we mentioned before, SS and GS share a similar domain — Spanning active

Site. This C-terminal region may bind to amylopectin and promote reorganization of two

175

domain relative position, in turn affecting catalysis. This assumption is now under

examination in Dr. Geiger’s lab.

 

Figure 111.56 Modeled potato GBSS structure based on E. coli GS structure.
The region ranging from helix (113 to (118 is colored red which is the
determinant of most specific properties of GBSS and may play a role in
tuning the relative orientation of the two domains.

176

“1.7. Conclusions

We have determined a series of E. coli GS and GS complex structures and they
provide a comprehensive understanding of the structure and functional relationship of
bacterial GS. All ligand-bound E. coli GS structures exhibit virtually identical molecular
organization and a deeply buried active site in the narrow interdomain cleft. In contrast,
the apo-deS exhibits a wide-open interdomain clefi and the N-domain and C-domain
are spread relatively far apart from each other, indicating that ligand binding induces a
considerably large domain-domain closure. The catalytically active closed form seen in
the ligand — bound GS structure is due to the concurrence of both the substrate ADP
moiety and the oligosaccharide acceptor /acceptor analogue HEPPSO. The structures of
GS in complex with ADP, glucose /DGMl, and HEPPSO indicate the presence of
intermediate DGM in GS catalysis and reveals Hislol and the phosphate are likely to
participate in the stabilization of the intermediate DGM. In addition, Arg300 and Lys305
are close to the phosphate and probably act as a Lewis acid and work electrostatically to
make the phosphate a better leaving group. The ADPGlc-incubated E3 77A crystal misses
the glucose moiety of substrate in the active site, which evidently suggests that Glu3 77
plays a critical role in fixating the glucose to be transferred. In addition, Glu377 may help
the catalytic residue Lys305 maintain its negative charge through two close hydrogen
bonds. Our finding of Glu377’s dual role in GS catalysis perfectly explains the observed
dramatic decrease in enzyme activity upon mutation on that position. In our
oligosaccharide-bound E3 77A and HEPPSO-bound thS structures, maltotriaose and
HEPPSO occupy a similar position in the interdomain clefi and Aspl37 plays an

important role in positioning these acceptor nucleophile analogues. Information gained

177

from our GS structures is in agreement with previous biochemical and mutagenesis
studies. Taken together, our E. coli GS structural studies allows us to confidently propose
an SN] mechanism over the double displacement mechanism for GS and other GT-B
retaining enzymes such as 885, MalP, OtsA and GP, which all share a virtually identical
active site. In addition, the oligosaccharide-bound GS structure suggests that a long
glucan chain binds only to the N-terminal domain allowing frequent opening and closing
of the domains without dissociation of glucan on each cycle.

Finally, based on our E. coli GS structures and the previously reported chimeric
studies of GBSS and 8811, we speculate that the region (0:13 - (x18) in starch synthase is
involved in positioning two domains and that interaction of this region of GBSSI with

amylopectin possibly orients the domains for catalysis.

178

 

References

Edwards, A., Borthakur, A., Bomemann, S., Venail, J ., Denyer, K., Waite, D.,
Fulton, D., Smith, A.M. and Martin, C. (1999) Specificity of starch synthase
isoforms from potato European journal of biochemistry / FEBS, 266, 724-736

Notredame, C., Higgins, D. and Heringa, J. (2000) T-coffee: A novel method for
multiple sequence alignments J. Mol. Biol, 302, 205-217

Drenth, J. (1994) Principles of protein X-ray crystallography (Springer, Ed.), New
York

Vagin, A. and Teplyakov, A. (1997) Molrep: An automated program for
molecular replacement Journal of Applied Crystallography, 30, 1022-1025

Jones, T.A. (1978) A graphics model building and refinement system for
macromolecules. F rodo: A graphics fitting program for macromolecules. Journal

of Applied Crystallography, 11, 268—272

Murshudov, G.N., A.Vagin, A. and Dodson, E.J. (1997) Refinement of
macromolecular structures by the maximum-likelihood method A cta Crystallogn,
Sect. D: Biol. Crystallogr. , 53, 240-255

Brfinger, A.T. (1992) Free R value: A novel statistical quantity for assessing the
accuracy of crystal structures Nature, 355, 472-475

179

10.

11.

12.

13.

14.

15.

16.

Kleywegt, G.J. and Brunger, AT. (1996) Checking your imagination:
Applications of the free R value. Structure, 4, 897-904

Ramakrishnan, C. and Ramachandran, G.N. (1965) Stereochemical criteria for
polypeptide and protein chain conformations. II. Allowed conformations for a pair

of peptide units. Biophysical Journal, 5, 909-933

Kleywegt, G.J. and Jones, T.A. (1996) Phi/ psi-chology: Ramachandran revisited
Structure, 15, 1395—1400

Yep, A., Ballicora, M.A., Sivak, MN. and Preiss, J. (2004) Identification and
characterization of a critical region in the glycogen synthase from Escherichia
coli J. Biol. Chem, 279, 8359-8367

Bergfors, TM. (1999) Protein crystallization techniques, strategies, and tips,
International University Line, La Jolla, CA

Otwinowski, Z.M.W. (1997) Scalepack Methods Enzymol, 276, 307-326

Yep, A., Bejar, C.M., Ballicora, M.A., Dubay, J .R., Iglesias, AA. and Preiss, J.
(2004) An assay for adenosine 5'-diphosphate (ADP)-glucose cpyrophosphorylase
that measures the synthesis of radioactive ADP-glucose with glycogen synthase

Analytical Biochemistry, 324, 52-59

Morell, SA. and Bock, RM. (1954) Ultraviolet absorption spectra of 5'-
ribonucleotides in American Chemical Society 126th meeting (Chemistry). 44,
New York

SDBSweb http://riodbOl .Ibase.Aist.Go.Jp/sdbs/ (national institute of advanced

 

industrial science and technology, date of access)

180

17.

18.

19.

20.

21.

22.

23.

24.

Collaborative Computational Project, N. (1994) The CCP4 suite: Programs for
protein crystallography Acta Crystallogr., Sect. D: Biol. Crystallogr. , 50, 760-
763

Mcguffm, L.J., Bryson, K. and Jones, D.T. (2000) The PSIPRED protein structure
prediction server. Bioinformatics, 16, 404-405 ‘

Cristina, H., Guinovart, J .J ., Fita, I. and Ferret, J .C. (2006) Crystal structure of an
archaeal glycogen synthase: Insights into oligomerization and substrate binding of
eukaryotic glycogen synthases J. Biol. Chem, 281, 2923-2931

Buschiazzo, A., Ugalde, J .E., Guerin, M.E., Shepard, W., Ugalde, RA. and
Alzari, PM. (2004) Crystal structure of glycogen synthase: Homologous enzymes
catalyze glycogen synthesis and degradation EMBO J., 23, 3196-3205

Watson, K.A., Mccleverty, C., Geremia, S., Cottaz, S., Driguez, H. and Johnson,
L.N. (1999) Phosphorylase recognition and phosphorolysis of its oligosaccharide
substrate: Answer to along outstanding question EMBO J., 18, 4619-4632

Yep, A., Ballicora, MA. and Preiss, J. (2004) The active site of the Escherichia
coli glycogen synthase is similar to the active site of retaining GT-b

glycosyltransferases Biochem. Biophys. Res. Commun., 316, 960-966

Geremia, S., Carnpagnolo, M., Schinzel, R. and Johnson, L.N. (2002) Enzymatic

catalysis in crystals of Escherichia coli maltodextrin phosphorylase J. Mol. Biol,
322, 413-423

Rao, V.V.R., Qasba, P.K., Balaji, RV. and Chandrasekaran, R. (1998)

Conformation of carbohydrates, Harwood Academic, Australia

181

25.

26.

27.

28.

29.

30.

31.

Appell, M., Strati, G., Willett, J .L. and Momany, RA. (2004) B3LYP/6—
311++G** study of alpha- and beta-D-glucopyranose and 1,5-anhydro-d-glucitol:
4C1 and 1C4 chairs, 30B and B30 boats, and skew-boat conformations.
Carbohydrate Research, 339, 537-551

Schwarz, A., Pierfederici, EM. and Nidetzky, B. (2005) Catalytic mechanism of
a-retaining glucosyl transfer by corynebacterium callunae starch phosphorylase:

The role of histidine-334 examined through kinetic characterization of site-
directed mutants Biochem. J. , 387, 43 7-445

Ryuta, K., Keiko, H., Toshihiko, A., Kunio, Y. and Kazuaki, H. (2006) Role of
trp140 at subsite —6 on the maltohexaose production of maltohexaose-producing
amylase from alkalophilic Bacillus sp.707 Protein Sci. , 15, 468-477

Alzari, P.M., Souchon, H.N. and Dominguez, R. (1996) The crystal structure of
endoglucanase CelA, a family 8 glycosyl hydrolase from clostridium

thermocellum Structure, 4, 265-275

Dutzler, R., Wang, Y.-F., Rizkallah, P.J., Rosenbusch, J .P. and Schirmer, T.
(1996) Crystal structures of various maltooligosaccharides bound to maltoporin

reveal a specific sugar translocation pathway Structure, 4, 127-134

Claude, J .-B., Suhre, K., Notredame, C., Claverie, J .-M. and Abergel, C. (2004)
Caspr: A web server for automated molecular replacement using homology

modeling Nucleic Acids Res. , 32(Web Server), W606-W609

Sundaralingam, M. (1968) Some aspects of stereochemistry and hydrogen
bonding of carbohydrates related to polysaccharide conformations. Biopolymers,
6, 189-213

182

32.

33.

34.

35.

36.

37.

38.

39.

Goldsmith, E., Sprang, S. and Fletterick, R. (1982) Structure of maltoheptaose by
difference fourier methods and a model for glycogen. J. Mol Biol, 156, 411-427

Dowda, M.K., Zenga, J ., Frenchb, AD. and Reillya, R]. (1992) Conformational
analysis of the anomeric forms of kojibiose, nigerose, and maltose using MM3

Carbohydrate Research, 230, 223-244

Preiss, J. (1984) Bacterial glycogen synthase and its regulation Annu. Rev.
Microbiol, 38, 419-458

Qi, G., Lee, R. and Hayward, S. (2005) A comprehensive and non-redundant

database of protein domain movements Bioinformatics, 21, 2832-2838

Furukawa, K., Tagaya, M., Inouye, M., Preiss, J. and Fukui, T. (1990)
Identification of lysine 15 at the active site in Escherichia coli glycogen synthase,
conservation of a Lys-X-Gly-Gly sequence in the bacterial and mammalian
enzymes J. Biol. Chem, 265, 2086-2090

F urukawa, K., Tagaya, M., Tanizawa, K. and Fukui, T. (1993) Role of the
conserved Lys-X-Gly-Gly sequence at the ADP-glucose-binding site in
Escherichia coli glycogen synthase J. Biol. Chem, 268, 23 837-23 842

Johnson, L.N., Acharya, K.R., Jordan, MD. and Mclaughlin, R]. (1990) Refined
crystal structure of the phosphorylase-heptulose 2-phosphate-oligosaccharide-
AMP complex. J. Mol. Biol, 211, 645-661

Gibson, R.P., Turkenburg, J .P., Charnock, S.J., Lloyd, R. and Davies, G.J. (2002)
Insights into trehalose synthesis provided by the structure of the retaining
glucosyltransferase OtsA Chemistry & Biology, 9, 1337-1346

183

40.

41.

42.

43.

44.

45.

46.

47.

Gibson, R.P., Tarling, C.A., Roberts, 8., Withers, S.G. and Davies, G.J. (2004)
The donor subsite of trehalose-6-phosphate synthase: Binary complexes with

UDP-glucose and UDP-2-deoxy-2-fluoro-glucose at 2a resolution J. Biol. Chem,
279, 1950-1955

Patenaude, S.I., Seto, N.O.L., Borisova, S.N., Szpacenko, A., Marcus, S.L.,
Palcic, M.M. and Evans, S.V. (2002) The structural basis for specificity in human
ABO(h) blood group biosynthesis Nature Structural Biology, 9, 685 - 690

Davies, G., Withers, S.G. and Sinnott, ML. (1997) in Comprehensive biological
catalysis. 119-209, Academic Press, London

Barford, D., Hu, SH. and Johnson, L.N. (1991) Structural mechanism for
glycogen phosphorylase control by phosphorylation and amp J. Mol. Biol, 218,
233-260

Lariviere, L., Sommer, N. and Morera, S. (2005) Structural evidence of a passive
base-flipping mechanism for AGT, an unusual GT-B glycosyltransferase J. Mol.
Biol, 352, 139-150

Schinzel, R. and Palm, D. (1990) Escherichia coli maltodextrin phosphorylase:
Contribution of active site residues glutamate-637 and tyrosine-538 to the
phosphorolytic cleavage of alpha-glucans Biochemistry, 29, 9956-9962

Cid, E., Gomis, R.R., Geremia, R.A., Guinovart, J. and F errer, J .C. (2000)
Identification of two essential glutamic acid residues in glycogen synthase J. Biol.

Chem, 275, 33614-33621

Sinnott, ML. (1990) Catalytic mechanisms of enzymatic glycosyl transfer Chem.
Rev. (Washington, DC, U. S. ), 90, 1171-1202

184

48.

49.

50.

51.

52.

53.

54.

Boix, E., Yingnan, Z., Swaminathan, G.J., Brew, K. and Acharya, KR. (2002)
Structure basis of ordered binding of donor and acceptor substrate to the retaining

glycosyltransferase,a-l,3-galactosyltransferase J. Biol. Chem, 277, 28310-28318

Sprang, S., Goldsmith, E. and Fletterick, R. (1987) Structure of the nucleotide
activation switch in glycogen phosphorylase a. Science, 237, 1012-1019

Fox, J ., Kawaguchi, K., Greenberg, E. and Preiss, J. (1976) Biosynthesis of
bacterial glycogen. Purification and properties of the Escherichia coli b

ADPglucose:1,4-a-D-glucan 4-a-glucosyltransferase Biochemistry, 15, 849-847

Goedl, C., Griessler, R., Schwarz, A. and Nidetzky, B. (2006) Structure—function
relationships for schizophyllum commune trehalose phosphorylase and their

implications for the catalytic mechanism of family GT-4 glycosyltransferases
Biochem. J. , 397, 491—500

Nidetzky, B. and Eis, C. (2001) A-retaining glucosyl transfer catalysed by
trehalose phosphorylase from schizophyllum commune: Mechanistic evidence

obtained from steady-state kinetic studies with substrate analogues and inhibitors
Biochem. J. , 360, 727-736

Eis, C., Watkins, M., Prohaska, T. and Nidetzky, B. (2001) Fungal trehalose
phosphorylase: Kinetic mechanism, pH-dependence of the reaction and some

structural properties of the enzyme from schizophyllum commune Biochem. J. ,
356, 757—767

Papageorgiou, A.C., Oikonomakos, N.G., Leonidas, D.D., Bemet, B., Beer, D.
and Vasella, A. (1991) The binding of D-gluconohydroximo-l, 5-lactone to
glycogen phosphorylase. Kinetic, ultracentrifugation and crystallographic studies.
Biochem. J. , 274

185

55.

56.

57.

58.

59.

60.

61.

62.

Mitchell, E.P., Withers, S.G., Ermert, P., Vasella, A.T., Garman, E.F.,
Oikonomakos, N.G. and Johnson, L.N. (1996) Ternary complex crystal structures
of glycogen phosphorylase with the transition state analog noj irimycin tetrazole

and phosphate in the T and R states. Biochemistry, 35, 7341-7355

Banait, NS. and Jencks, WP. (1991) Reaction of anionic nucleophiles with a—D-

glucopyranosyl fluoride in aqueous solution through a concerted, AnDn(Sn2)
mechanism J. Am. Chem. Soc., 113, 7951-7958

Banait, NS. and Jencks, WP. (1991) General-acid and general -base catalysis of
the cleavage of a-D-glucopyranosyl fluoride J. Am. Chem. Soc, 113, 795 8-7963

Bowles, D., Isayenkova, J ., Lim, E.-K. and Poppenberger, B. (2005)
Glycosyltransferases: Managers of small molecules Curr. Opin. Plant Biol, 8,
254-263

Breton, C., Najdrova, L., Jeanneau, C., Koa, J. and Imberty, A. (2006) Structures
and mechanisms of glycosyltransferases Glycobiology, 16, 29-37

Davies, G.J. (2001) Sweet secrets of synthesis Nature Structural Biology, 8, 98-
100

Unligil, U.M. and Rini, J .M. (2000) Glycosyltransferase structure and mechanism
Curr. Opin. Struct. Biol. , 10, 510-517

Lairson, LL. and Withers, S.G. (2004) Mechanistic analogies amongst
carbohydrate modifying enzymes Chem. Commun. (Cambridge, U. K. ), 20, 2243-
2248

186

 

63.

64.

65.

66.

67.

68.

69.

70.

Persson, K., Ly, H.D., Dieckelmann, M., Wakarchuk, W.W., Withers, S.G. and
Strynadka, N.C.J. (2001) Crystal structure of the retaining galactosyltransferase
LgtC from neisseria meningitidis in complex with donor and acceptor sugar

analogs Nature Structural Biology, 8, 166-175

Gao, Z., Keeling, P., Shibles, R. and Guan, H. (2004) Involvement of lysine-193
of the conserved "K-T-G-G" motif in the catalysis of maize starch synthase Ila
Arch. Biochem. Biophys. , 427, 1-7

Nichols, D.J., Keeling, P.L., Spalding, M. and Guan, H. (2000) Involvement of
conserved aspartate and glutamate residues in the catalysis and substrate binding

of maize starch synthase. Biochemistry, 39, 7820-7825

Imparl-Radosevich, J .M., Keeling, PL. and Guan, H. (1999) Essential arginine
residues in maize starch synthase IIa are involved in both ADP-glucose and
primer binding FEBS Lett., 457, 357-362

Manners, DJ. (1989) Recent developments in our understanding of amylopectin

structure Carbohydrate Polymers, 11, 87-112

Slattery, J .C., Kavakli, 1.H. and Okita, W.T. (2000) Engineering starch for
increased quantity and quality Trends Plant Sci. , 291-298

Denyer, K., Waite, D., Edwards, A., Martin, C. and Smith, AM. (1999)
Interaction with amylopectin influences the ability of granule-bound starch

synthase I to elongate malto-oligosaccharides Biochem. J. , 342, 647-653
Edwards, A., Borthakur, A., Bomemann, S., Venail, J ., Denyer, K., Waite, D.,

Fulton, D., Smith, AM. and Martin, C. (1999) Specificity of starch synthase
isoforms from potato European Journal of Biochemistry, 266, 724-736

187

IIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIII

WWWWWWW