Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Investigating the enzymatic mechanisms of the inverting and retaining glycosyltransferases by NMR spectroscopy Chan, Hau Wing 2012

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata

Download

Media
24-ubc_2012_fall_chan_hauwing.pdf [ 13.38MB ]
Metadata
JSON: 24-1.0072879.json
JSON-LD: 24-1.0072879-ld.json
RDF/XML (Pretty): 24-1.0072879-rdf.xml
RDF/JSON: 24-1.0072879-rdf.json
Turtle: 24-1.0072879-turtle.txt
N-Triples: 24-1.0072879-rdf-ntriples.txt
Original Record: 24-1.0072879-source.json
Full Text
24-1.0072879-fulltext.txt
Citation
24-1.0072879.ris

Full Text

INVESTIGATING THE ENZYMATIC MECHANISMS OF THE INVERTING AND RETAINING GLYCOSYLTRANSFERASES BY NMR SPECTROSCOPY  by  Hau Wing Chan  B.Sc., Simon Fraser University, 2006  A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF  DOCTOR OF PHILOSOPHY  in  THE FACULTY OF GRADUATE STUDIES  (Biochemistry and Molecular Biology)  THE UNIVERSITY OF BRITISH COLUMBIA  (Vancouver)  July 2012  © Hau Wing Chan, 2012  Abstract The overall goal of this thesis was to investigate the structures and enzymatic mechanisms of glycosyltransferases using NMR spectroscopy, complemented with X-ray crystallography and enzyme kinetic measurements. The bifunctional sialyltransferase CstII from Campylobacter jejuni and the α-1,4-galactosyltransferase LgtC from Neisseria meningitidis were chosen to be the model inverting and retaining enzymes, respectively. Cell surface glycans are often terminated by sialic acid, which is incorporated onto sugar acceptors by sialyltransferases. The crystal structures of the GT family 42 CstI/CstII provided key insights into the sialyl-transfer mechanism, including tentative identification of His202/His188 as the catalytic base. In support of this hypothesis, the CstII-H188A mutant is able to catalyze sialyl transfer from CMP-Neu5Ac to added anions such as azide and formate, but not to its natural sugar acceptor lactose. By systematically introducing point mutations at the subunit interfaces, two active monomeric variants, CstII-F121D and CstIIY125Q, were obtained and characterized by NMR spectroscopy. In contrast to the wild-type tetramer, the monomeric CstII variants yielded good quality amide 1H/15N-HSQC and methyl-TROSY NMR spectra. However, the absence of signals from approximately one half of the amides in the 1H/15N-HSQC spectra of both monomeric forms suggests that the enzyme undergoes substantial conformational exchange on a msec-sec timescale. The histidine pKa values of CstII-F121D in its apo form were measured by monitoring the pHdependent chemical shifts of [13C1]-histidine, biosynthetically incorporated into the otherwise uniformly deuterated protein. Consistent with its proposed catalytic role, the sitespecific pKa value ~ 6.6 for His188 matches the apparent pKa value ~ 6.5 governing the pHdependence of kcat/Km for CstII towards CMP-Neu5Ac in the presence of saturating acceptor substrate. STD-NMR spectroscopy was also employed to investigate the orientation of bound CMP-3FNeu5Ac in the presence or absence of sugar acceptor lactose. Combined with the crystal structure of CstII binary complex, the STD data indicates that an “open/closed” equilibrium of its active site lid motif shifts upon sugar donor binding. The enzymatic mechanism of the retaining glycosyltransferase LgtC appears to involve a “front-side attack” SNi mechanism or a SNi-like mechanism with a short-lived oxocarbenium-phosphate  ion  pair  intermediate.  Furthermore,  based  upon  X-ray  crystallographic studies, two flexible loops were proposed to become ordered over the active ii  site of LgtC upon sugar donor binding. Accordingly, NMR spectroscopy was used to investigate to the dynamic properties of the enzyme with an emphasis on the roles of these motions in catalysis. The amide 1H/15N-TROSY-HSQC and methyl-TROSY spectra of LgtC were partially assigned using a variety of NMR spectroscopic approaches, combined with mutagenesis of all the isoleucine residues in the protein. More than expected number of methyl signals was observed, indicating that LgtC adopts multiple conformational states. These states, termed "opened" and "closed", are in equilibrium on a seconds time-scale, and their relative populations change upon mutation and substrate binding. Furthermore, relaxation dispersion studies indicated substantial msec-sec time-scale motions of methyl groups both within and distal to the active site in apo and substrate-bound forms of LgtC. Thus LgtC exhibits a range of dynamic behaviours, potentially linked to its catalytic function. They were studied in this thesis using NMR spectroscopy and kinetic studies.  iii  Preface The majority of the work in Chapter 2 has been published (Chan, P. H., Lairson, L. L., Lee, H. J., Wakarchuk, W. W., Strynadka, N. C. Withers, S. G., McIntosh, L. P. (2009) NMR Spectroscopic Characterization of the Sialyltransferase CstII from Campylobacter jejuni: Histidine 188 Is the General Base. Biochemistry 48:163-170). I conducted most of the research work and helped writing the manuscript. The chemical rescue experiments and pHdependent kinetic studies of the enzyme were done by Dr. Luke Lairson. The crystal structure of the monomeric CstII-Y125Q was solved by Dr. Ho-Jun Lee. STD-NMR studies, carried out after this research was published, are included herein. The work in Chapters 3 and 4 is currently being prepared for publication. Again, I carried out the vast majority of the research. The isoleucine to alanine mutants of LgtC were generated and characterized kinetically by Sophie Weissbach. The variants with unnatural amino acids were prepared and characterized by Adrienne Cheung with NMR spectroscopy assistance by Dr. Mark Okon.  iv  Table of Contents  Abstract .................................................................................................................................... ii Preface..................................................................................................................................... iv Table of Contents .................................................................................................................... v List of Tables ......................................................................................................................... xii List of Figures ....................................................................................................................... xiii List of Abbreviations ......................................................................................................... xviii List of Amino Acid Abbreviations ..................................................................................... xxii Acknowledgements ............................................................................................................ xxiii Chapter 1  General introduction ...................................................................................... 1  1.1  GLYCOBIOLOGY ................................................................................................... 2  1.2  GLYCOSIDE HYDROLASES AND GLYCOSYLTRANSFERASES .................. 3  1.2.1  Glycoside hydrolases ............................................................................................ 3  1.2.2  Glycosyltransferases ............................................................................................. 4  1.2.2.1  General structures of GTs ............................................................................. 4  1.2.2.2  Enzymatic mechanisms of GTs .................................................................... 7  1.2.2.3  Proposed alternative SNi or SNi-like mechanisms of retaining GTs ............. 9  1.3  STRUCTURAL DYNAMICS OF GTs .................................................................. 11  1.3.1  Protein structural dynamics................................................................................. 11  1.3.2  Glycosyltransferase dynamics ............................................................................ 11  1.4  SIALYLTRANSFERASE CSTI/II FROM CAMPYLOBACTER JEJUNI ............. 15  1.4.1  CstI and CstII ...................................................................................................... 16  1.4.2  Monofunctional and bifunctional CstII ............................................................... 18  1.4.3  Catalytic His188 of CstII .................................................................................... 18  1.4.4  Substrate binding of CstII ................................................................................... 19  1.4.5  Structural dynamics of CstII ............................................................................... 20  1.5  GALACTOSYLTRANSFERASE LGTC FROM NEISSERIA MENINGITIDIS... 21  v  1.5.1  Structure of LgtC ................................................................................................ 21  1.5.2  Substrate binding of LgtC ................................................................................... 23  1.5.3  Controversial enzymatic mechanism of LgtC..................................................... 23  1.5.4  Structural dynamics of LgtC ............................................................................... 24  1.6  NMR SPECTROSCOPY AND LARGE MACROMOLECULES ........................ 25  1.6.1  Transverse relaxation-optimized spectroscopy ................................................... 28  1.6.2  Methyl-TROSY spectroscopy............................................................................. 29  1.6.2.1  Selective methyl labeling ............................................................................ 32  1.6.2.2  Studies of structural dynamics and conformational exchange using methyl-  TROSY approaches .................................................................................................... 34 1.6.2.2.1 Magnetization exchange spectroscopy ................................................... 34 1.6.2.2.2 Relaxation dispersion ............................................................................. 35 1.6.3 1.7  STD-NMR spectroscopy..................................................................................... 38 OVERVIEW OF THE THESIS .............................................................................. 40  Chapter 2  Investigation of the enzymatic mechanism and structural dynamics of the  bifunctional sialyltransferase CstII from Campylobacter jejuni ....................................... 42 2.1  INTRODUCTION .................................................................................................. 43  2.1.1  Catalytic His188 of CstII .................................................................................... 43  2.1.2  Structural dynamics of CstII ............................................................................... 43  2.1.3  Substrate binding of CstII ................................................................................... 44  2.2  METHODS ............................................................................................................. 45  2.2.1  Site-directed mutagenesis ................................................................................... 45  2.2.2  Protein expression and purification .................................................................... 45  2.2.3  Activity assays .................................................................................................... 46  2.2.4  pH-Dependence of kcat/Km .................................................................................. 46  2.2.5  Anion rescue studies ........................................................................................... 46  2.2.6  Analytical size exclusion chromatography and static light scattering ................ 47  2.2.7  CD spectroscopy ................................................................................................. 47  2.2.8  X-ray crystallographic analysis of CstII-Y125Q ................................................ 47  2.2.9  NMR spectroscopy.............................................................................................. 48  2.2.10  STD-NMR spectroscopy................................................................................. 49 vi  2.3  RESULTS ............................................................................................................... 50  2.3.1  Chemical rescue of CstII-H188A activity........................................................... 50  2.3.2  pH-Dependent activity of CstII ........................................................................... 52  2.3.3  Monomerization of CstII..................................................................................... 53  2.3.3.1  Oligomerization states of the mutants......................................................... 56  2.3.3.2  Kinetic analysis of the monomeric CstII-Y125Q ....................................... 56  2.3.4  Structural characterization of monomeric CstII .................................................. 57  2.3.5  NMR spectroscopic studies of CstII monomeric mutants .................................. 59  2.3.6  Methyl relaxation dispersion of CstII-Y125Q .................................................... 63  2.3.7  NMR determination of the pKa value of His188 ................................................ 64  2.3.7.1 2.3.7.2 2.3.8 2.4  Peak assignment of catalytic His188 by site-directed mutagenesis ............ 64 pKa measurement of His188 by NMR spectroscopy ...................................... 66  Substrate binding mode of CstII ......................................................................... 66 DISCUSSION AND CONCLUSIONS .................................................................. 70  2.4.1  Monomerization of CstII..................................................................................... 70  2.4.2  NMR spectroscopic analysis of monomeric forms of CstII................................ 71  2.4.3  Catalytic His188.................................................................................................. 73  2.4.4  Substrate binding mode and dynamics of the lid motif of CstII ......................... 75  2.4.5  Proposed conformational models of the CstII reaction pathway ........................ 76  Chapter 3  NMR  spectral  assignment  of  lipooligosaccharide  α-1,4-  galactosyltransferase (LgtC) from Neisseria meningitidis ................................................. 79 3.1  INTRODUCTION .................................................................................................. 80  3.1.1  Crystal structures of LgtC ................................................................................... 80  3.1.2  Proposed enzymatic mechanism of LgtC ........................................................... 80  3.1.3  Proposed conformational dynamics of LgtC ...................................................... 81  3.2  METHODS ............................................................................................................. 82  3.2.1  Cloning and site-directed mutagenesis of LgtC .................................................. 82  3.2.2  Protein expression and purification .................................................................... 82  3.2.3  Optimization of LgtC NMR buffer conditions ................................................... 84  3.2.4  CD spectroscopy ................................................................................................. 84  3.2.5  Refolding screen of LgtC .................................................................................... 84 vii  3.2.6  Activity assays .................................................................................................... 87  3.2.7  NMR spectroscopy.............................................................................................. 87  3.3  RESULTS ............................................................................................................... 89  3.3.1  Constructs of LgtC .............................................................................................. 89  3.3.2  Optimization of NMR buffer conditions............................................................. 90  3.3.3  NMR spectroscopy of LgtC ................................................................................ 92  3.3.3.1  1  H/15N-HSQC spectrum of 15N-labeled LgtC ............................................. 92  3.3.3.2  Refolding of LgtC ....................................................................................... 93  3.3.3.2.1 CD spectroscopy of LgtC ....................................................................... 94 3.3.3.2.2 Optimization of LgtC refolding conditions ............................................ 95 3.3.3.2.3 Protium-deuterium exchange of LgtC .................................................... 98 3.3.3.3  1  H/15N-TROSY-HSQC spectra of uniformly deuterated  15  N-labeled LgtC  100 3.3.3.3.1 Sequence-specific  15  N backbone assignment of LgtC by NMR  spectroscopy.......................................................................................................... 103 3.3.3.3.2  15  N backbone assignment of LgtC by site-directed mutagenesis ......... 107  3.3.3.3.3 Enzymatic activities of LgtC IA mutants ............................................. 108 3.3.3.3.4 Selective 15N amino acid labeled LgtC ................................................ 111 3.3.3.3.5  15  N-glutamate selectively labeled LgtC-Q189A and LgtC-Q189E ...... 114  3.3.3.3.6 Summary of 15N amide backbone assignment of LgtC ........................ 116 3.3.4  Methyl-TROSY of selectively 13C-methyl-labeled LgtC ................................. 117  3.3.4.1  Effect of substrate binding on the methyl-TROSY spectrum of LgtC ..... 120  3.3.4.2  Methyl-TROSY spectral assignments of LgtC ......................................... 123  3.3.4.2.1 Assignment by NMR spectroscopy ...................................................... 123 3.3.4.2.2 Assignment of methyl-TROSY spectra by site-directed alanine mutagenesis ........................................................................................................... 128 3.3.4.2.3 Assignment of methyl-TROSY spectra by site-directed isoleucine to valine mutagenesis ................................................................................................ 132 3.3.4.2.4 PRE NMR experiment of [1H/13C] selectively methyl labeled LgtC binary complex...................................................................................................... 141 3.3.4.2.5 Summary of the methyl-TROSY assignments of LgtC ........................ 145 viii  H/13C-HSQC spectra of [13Cε1]-histidine-labeled LgtC................................... 147  1  3.3.5 3.4  DISCUSSION AND CONCLUSION................................................................... 149  3.4.1  General insights concerning LgtC NMR spectra .............................................. 149  3.4.2  Insights into of the mode of substrate binding to LgtC .................................... 149  3.4.3  Implication of the multi-conformational states of LgtC ................................... 149  Chapter 4  Investigation  of  structural  dynamics  and substrate  binding  of  lipooligosaccharide α-1,4-galactosyltransferase (LgtC) from Neisseria meningitdis by NMR spectroscopy .............................................................................................................. 153 4.1  INTRODUCTION ................................................................................................ 154  4.1.1  Substrate binding of LgtC ................................................................................. 154  4.1.2  Multiple conformational equilibria of LgtC ..................................................... 154  4.1.3  Structural dynamics of LgtC ............................................................................. 154  4.2  METHODS ........................................................................................................... 156  4.2.1  Protein expression and purification .................................................................. 156  4.2.2  NMR spectroscopy............................................................................................ 156  4.2.2.1  31  4.2.2.2  19  4.2.2.3  STD-NMR spectroscopy........................................................................... 157  4.2.2.4  1  4.2.2.5  Methyl-TROSY NMR spectroscopy......................................................... 157  P NMR spectroscopy.............................................................................. 156 F NMR spectroscopy.............................................................................. 156  H/15N-TROSY-HSQC NMR spectroscopy ............................................. 157  4.2.2.5.1 NMR chemical shift perturbation in methyl-TROSY spectra .............. 158 4.2.2.5.2 Conformational exchange measurements ............................................. 158 4.2.2.5.3 Methyl relaxation dispersion experiment ............................................. 158 4.2.3  LgtC-Q189E glycosyl-enzyme intermediate trapping by NMR spectroscopy . 159  4.2.4  LgtC-Q189E glycosyl-enzyme intermediate trapping observed by electrospray  ionization mass spectrometry ........................................................................................ 159 4.3  RESULTS ............................................................................................................. 160  4.3.1  Substrate binding by LgtC ................................................................................ 160  4.3.1.1  Simultaneous binding of the metal ion and sugar donor to LgtC ............. 160  4.3.1.2  Direct observation of UDP and UDP-2FGal ............................................ 162  4.3.1.2.1 Substrate binding monitored by 31P NMR spectroscopy...................... 162 ix  4.3.1.2.2 Sugar donor binding monitored by 19F NMR spectroscopy ................. 165 4.3.1.3  STD-NMR spectroscopy of LgtC ............................................................. 166  4.3.1.3.1 STD-NMR spectroscopy of UDP binding with LgtC .......................... 167 4.3.1.3.2 STD-NMR spectroscopy of UDP-2FGal binding with LgtC ............... 169 4.3.1.4  Substrate binding of LgtC monitored by methyl-TROSY spectroscopy .. 169  4.3.1.4.1 Spectral changes upon UDP-2FGal binding ........................................ 170 4.3.1.4.2 Spectral changes upon lactose binding ................................................. 172 4.3.1.4.3 Spectral changes upon UDP binding .................................................... 173 4.3.1.5  Substrate binding to 15N-tyrosine-labeled LgtC ....................................... 175  4.3.1.6  Substrate binding investigated using UAA ............................................... 178  4.3.1.6.1 Substrate binding of CF3Phe-labeled LgtC-F132X .............................. 179 4.3.1.6.2 Substrate binding of O13CH3Phe-labeled LgtC-Y186X ....................... 180 4.3.2  Multiple-conformational states of LgtC............................................................ 182  4.3.2.1  Temperature effect on the methyl-TROSY of apo LgtC .......................... 183  4.3.2.2  Magnetization transfer between conformational states ............................. 185  4.3.2.3  Summary of conformational exchange of LgtC ........................................ 188  4.3.3  Studies of LgtC structural dynamics ................................................................. 188  4.3.3.1 4.3.4  Methyl relaxation dispersion of 13C selectively methyl labeled LgtC ...... 188  Intermediate trapping of LgtC-Q189E .............................................................. 195  4.3.4.1  NMR spectroscopy.................................................................................... 195  4.3.4.2  ESI-MS ..................................................................................................... 198  4.4  DISCUSSION AND CONCLUSION................................................................... 200  4.4.1  Substrate and product binding of LgtC ............................................................. 200  4.4.2  Multi-conformational states of LgtC ................................................................ 201  4.4.3  Multiple conformational equilibria of LgtC ..................................................... 202  4.4.4  Structural dynamics of LgtC ............................................................................. 203  4.4.5  Conformational movements of the catalytic loop and α-helix J of LgtC .......... 204  4.4.6  Proposed conformational models of LgtC reaction pathway ............................ 208  Chapter 5  General conclusions and future work ....................................................... 211  5.1  GENERAL SUMMARY OF THE SIALYLTRANSFERASE CSTII STUDIES 212  5.2  FUTURE DIRECTIONS - SIALYLTRANSFERASES ...................................... 214 x  5.3  GENERAL SUMMARY OF THE GALACTOSYLTRANSFERASE LGTC  STUDIES .......................................................................................................................... 215 5.4  FUTURE DIRECTIONS - GALACTOSYLTRANSFERASES .......................... 218  References ............................................................................................................................ 219 Appendices ........................................................................................................................... 234 Appendix A  Enzymatic synthesis of UDP-2FGal ......................................................... 234  A.1  PREPARATION OF 2FGAL ........................................................................... 234  A.2  EXPRESSION AND PURIFICATION OF FUSION ENZYME GALK/GALT 234  A.3  SYNTHESIS OF UDP-2FGAL ........................................................................ 235  A.4  PURIFICATION OF UDP-2FGAL .................................................................. 235  Appendix B  Characterization of UDPGlcDH ............................................................... 236  Appendix C  1  Appendix D  Methyl-TROSY NMR chemical shift tables of LgtC ............................... 243  H/15H-TROSY-HSQC NMR chemical shift table of apo LgtC .............. 238  xi  List of Tables Table 2.1: Screening of CstII mutants for oligomerization and activity ................................. 55 Table 2.2: Steady state kinetic parameters for wild-type CstII and monomeric CstII-Y125Q57 Table 2.3: Data collection and refinement of CstII-Y125Q in complex with CMP-3FNeu5Ac ............................................................................................................................. 58 Table 3.1: Screening of refolding conditions for LgtC ........................................................... 85 Table 3.2: Results of NMR buffer condition screen for LgtC ................................................ 91 Table 3.3: Steady state kinetic parameters for LgtC IA and VA mutants ............................ 110 Table 3.4: Correlation of the distance between Mn 2+ and isoleucine δ1-methyl groups with PRE data............................................................................................................ 144 Table 4.1: Steady state kinetic parameters for UAA-labeled LgtC. ..................................... 179 Table 4.2: Chemical shift perturbation of residue Ile191 in methyl-TROSY spectra. ......... 207 Table B.1: Steady state kinetic parameters for UDPGlcDH. ................................................ 236 Table C.1: NMR chemical shift table of apo 15N-labeled LgtC. .......................................... 238 Table D.1: NMR chemical shift table of apo LgtC .............................................................. 243 Table D.2: NMR chemical shift table of LgtC binary complex with Mg2+•UDP-2FGal. .... 245 Table D.3: NMR chemical shift table of LgtC ternary complex with Mg2+•UDP-2FGal• lactose................................................................................................................. 246 Table D.4: NMR chemical shift table of LgtC product complex with Mg2+•UDP............... 247  xii  List of Figures Figure 1.1: General enzymatic mechanisms used by inverting and retaining GHs. ................. 4 Figure 1.2: Crystal structures of representative GTs from different GT families. ................... 6 Figure 1.3: Enzymatic reactions catalyzed by inverting and retaining GTs. ............................ 7 Figure 1.4: Proposed reaction mechanism by an inverting GT. ............................................... 8 Figure 1.5: Proposed double SN2 displacement mechanism (top), SNi-like (middle), and SNi reaction mechanisms (bottom) used by retaining GTs. ....................................... 10 Figure 1.6: Crystal structures of the monomer subunits of the tetrameric CstII..................... 12 Figure 1.7: Overlaid crystal structures of LgtC binary and ternary complexes. ..................... 13 Figure 1.8: Structures of LOS of C. jejuni and ganglioside GM1 of human. ......................... 16 Figure 1.9: Crystal structures of the monomer subunits of two tetrameric sialyltransferases from Campylobacter jejuni ................................................................................. 17 Figure 1.10: Sialyltransfer reaction catalyzed by CstII .......................................................... 18 Figure 1.11: Superimposed structures of CstII with bound donor and acceptor analogs. ...... 20 Figure 1.12: Structure of LOS of N. meningitidis................................................................... 21 Figure 1.13: Crystal structure of the LgtC ternary complex. .................................................. 22 Figure 1.14: Relationship between the correlation time for molecular motion and the relaxation rates of excited spins. ......................................................................... 26 Figure 1.15: TROSY effect in NMR signals of small and large proteins. .............................. 27 Figure 1.16: Schematic background of TROSY experiment. ................................................. 29 Figure 1.17: Background of methyl-TROSY effect. .............................................................. 32 Figure 1.18: Examples of precursors used in different selective methyl group labeling strategies.............................................................................................................. 33 Figure 1.19: Schematic spectra showing the NMR measurement of conformational exchange. ............................................................................................................................. 35 Figure 1.20: Schematic diagram illustrating the measurement of conformational dynamics by relaxation dispersion experiment. ....................................................................... 38 Figure 1.21: Schematic diagram showing the principle of STD-NMR spectroscopy. ........... 39 Figure 2.1: Chemical rescue of CstII-H188A by exogenous anions. ..................................... 51 Figure 2.2: The pH dependence of kcat/Km for utilization of CMP-Neu5Ac by CstII. ............ 53 xiii  Figure 2.3: Superimposed structures of monomer of wild-type CstII and CstII-Y125Q complexed with the sugar donor analog CMP-3FNeu5Ac ................................. 54 Figure 2.4: Static light scattering of wild-type CstII (left) and CstII-Y125Q (right). ............ 56 Figure 2.5: 1H/15N-HSQC spectra of 15N-labeled CstII variants. ........................................... 60 Figure 2.6: Overlaid 1H/13C-methyl-TROSY spectra of uniformly deuterated and [ 1H/13C] selectively methyl labeled CstII-Y125Q in its apo form (red) and in the presence of 5 mM CMP-3FNeu5Ac (blue) in D2O buffer. ................................................ 62 Figure 2.7: Methyl relaxation dispersion curves of apo CstII and its binary complex. .......... 64 Figure 2.8: Measurement of the pKa value of His188 by NMR spectroscopy. ....................... 65 Figure 2.9: STD-NMR studies of CMP-3FNeu5Ac binding to CstII-F121D. ....................... 67 Figure 2.10: Surface presentation of CstII binary complex. ................................................... 68 Figure 2.11: Binding of lactose with CstII-F121D-CMP-3FNeu5Ac complex...................... 69 Figure 2.12: Model of substrate binding modes and dynamics of the lid motif of CstII along its reaction pathway. ........................................................................................... 77 Figure 3.1: Constructs of LgtC. .............................................................................................. 90 Figure 3.2: 1H/15N-HSQC spectrum of 15N-labeled LgtC-25. ................................................ 93 Figure 3.3: CD spectra of LgtC-25. ........................................................................................ 94 Figure 3.4: Activity assay to screen refolding of LgtC-25. .................................................... 96 Figure 3.5: Activity assay to screen refolding of LgtC (C128/174S). .................................... 97 Figure 3.6: LgtC undergoes rapid amide hydrogen exchange under alkaline conditions. ...... 99 Figure 3.7: 1H/15N-TROSY-HSQC spectra of LgtC. ............................................................ 102 Figure 3.8: Partially assigned 1H/15N-TROSY-HSQC spectrum of uniformly deuterated 13  C/15N-labeled LgtC......................................................................................... 104  Figure 3.9: Residues with assigned signals in the 1H/15N-TROSY-HSQC spectrum of LgtC are mapped onto the crystal structure of the protein. ........................................ 106 Figure 3.10: 1H/15N-TROSY-HSQC spectra of uniformly 15N-labeled LgtC mutants. ........ 108 Figure 3.11: TLC activity assay of the LgtC IA and VA mutants. ....................................... 109 Figure 3.12: Crystal structure of LgtC ternary complex showing the isoleucine and valine residues nearest the active site. ......................................................................... 111 Figure 3.13: 1H/15N-TROSY-HSQC spectra of selectively 15N-labeled LgtC. .................... 113  xiv  Figure 3.14: 1H/15N-TROSY-HSQC spectra of selectively  15  N-glutamate labeled LgtC  mutants. ............................................................................................................. 116 Figure 3.15: Methyl-TROSY spectrum of apo LgtC. ........................................................... 119 Figure 3.16: MALDI-TOF-MS spectra of wild-type LgtC and LgtC-A249X...................... 120 Figure 3.17: Substrate binding monitored by methyl-TROSY spectra of uniformly deuterated and [1H/13C] selectively methyl labeled LgtC................................................... 123 Figure 3.18: Schematic diagrams of the intraresidue magnetization transfer steps from the 13  CH3-methyl groups of Ile, Leu, and Val to the amide (A) 1HN and (B) 13C’. 124  Figure 3.19: Partial assignment of the methyl-TROSY spectra of uniformly deuterated and [1H/13C] selectively methyl labeled apo LgtC obtained using magnetization transfer experiments. ......................................................................................... 127 Figure 3.20: Methyl-TROSY spectra of uniformly deuterated and [1H/13C] selectively methyl labeled apo (A) wild-type LgtC, (B) LgtC-I76A, (C) LgtC-I79A, (D) LgtCI81A, (E) LgtC-I104A, (F) LgtC-V106A, and (G) LgtC-V133A. .................... 129 Figure 3.21: Methyl-TROSY spectra of uniformly deuterated and [1H/13C] selectively methyl labeled (A) wild-type LgtC, (B) LgtC-I76A, (C) LgtC-I79A, (D) LgtC-I81A, (E) LgtC-I104A, (F) LgtC-V106A, and (G) LgtC-V133A binary complex. .......... 131 Figure 3.22: TLC activity assay of the LgtC IV mutants. .................................................... 133 Figure 3.23: Methyl-TROSY spectra of uniformly deuterated and [1H/13C] selectively methyl labeled LgtC IV mutants in their apo-form. ..................................................... 135 Figure 3.24: Chemical shift perturbations of Val133 caused by the mutation of I143V. ..... 137 Figure 3.25: Methyl-TROSY spectra of uniformly deuterated and [1H/13C] selectively methyl labeled LgtC IV mutants saturated with Mg2+ and UDP-2FGal. ...................... 139 Figure 3.26: PRE NMR experiment of LgtC binary complex .............................................. 143 Figure 3.27: Assigned methyl-TROSY spectra of isoleucine residues of uniformly deuterated and [1H/13C] selectively methyl labeled LgtC along its reaction pathway ....... 146 Figure 3.28: 1H/13C-HSQC spectra of deuterated [13Cε1]-histidine-labeled LgtC ................ 148 Figure 3.29: Proposed different conformational states of LgtC. .......................................... 151 Figure 4.1: Simultaneous binding of Mg2+ and UDP-2FGal by LgtC. ................................. 161 Figure 4.2: UDP binding of LgtC monitored by 31P NMR spectroscopy. ............................ 163 Figure 4.3: UDP-2FGal binding of LgtC monitored by 31P NMR spectroscopy.................. 164 xv  Figure 4.4: Hydrolysis of UDP-2FGal by LgtC monitored by 31P NMR spectroscopy. ...... 165 Figure 4.5: UDP-2FGal binding of LgtC monitored by 19F NMR spectroscopy.................. 166 Figure 4.6: STD-NMR studies of UDP binding to LgtC. ..................................................... 168 Figure 4.7: STD-NMR studies of UDP-2FGal binding to LgtC. .......................................... 169 Figure 4.8: CSP of isoleucine δ1-methyl of uniformly deuterated and [1H/13C] selectively methyl labeled LgtC upon UDP-2FGal binding. .............................................. 171 Figure 4.9: CSP of isoleucine δ1-methyl of uniformly deuterated and [1H/13C] selectively methyl labeled LgtC binary complex upon lactose binding. ............................ 172 Figure 4.10: CSP of isoleucine δ1-methyl of uniformly deuterated and [1H/13C] selectively methyl labeled LgtC upon UDP binding........................................................... 173 Figure 4.11: Depletion of UDP from LgtC•Mg2+•UDP binary complex by PK................... 174 Figure 4.12: Substrate binding monitored by 1H/15N-TROSY-HSQC spectra of 15N-tyrosinelabeled LgtC. ..................................................................................................... 176 Figure 4.13: Locations of tyrosine residues in the crystal structure of the LgtC ternary complex. ............................................................................................................ 177 Figure 4.14: Isotopically labeled UAA incorporated at positions 132 and 186 of LgtC. ..... 178 Figure 4.15: Substrate binding of CF3Phe-labeled LgtC-F132X monitored by  19  F NMR  spectroscopy. ..................................................................................................... 180 Figure 4.16: Substrate binding of O13CH3Phe-labeled LgtC-Y186X monitored by 1H/13CHSQC NMR spectroscopy. ............................................................................... 181 Figure 4.17: Methyl-TROSY spectra of uniformly deuterated and [1H/13C] selectively methyl labeled LgtC as a function of temperature. ....................................................... 185 Figure 4.18: Magnetization exchange experiments reveal the slow interconversion between the two conformational states of apo LgtC. ...................................................... 188 Figure 4.19: Methyl relaxation dispersion curves of apo LgtC. ........................................... 189 Figure 4.20: Methyl relaxation dispersion curves of apo LgtC with different delay times. . 190 Figure 4.21: Methyl relaxation dispersion studies of apo LgtC (40 ms delay). ................... 192 Figure 4.22: Methyl relaxation dispersion studies of apo LgtC (20 ms delay). ................... 193 Figure 4.23: Comparison of methyl relaxation data for apo LgtC and its binary complex. . 194 Figure 4.24: Overlaid 1H/15N-TROSY HSQC spectra of  15  N-labeled wild-type LgtC and  LgtC-Q189E. ..................................................................................................... 195 xvi  Figure 4.25: Overlaid 1H/15N-TROSY HSQC spectra of LgtC-Q189E. .............................. 197 Figure 4.26: ESI-MS spectra of  15  N-labeled LgtC-Q189E and its covalent glycosyl-enzyme  intermediate....................................................................................................... 199 Figure 4.27: Active site α-helix J and loop of LgtC. ............................................................ 205 Figure 4.28: Locations of Ile104 and Ile129 relative to the LgtC active site. ...................... 206 Figure 4.29: Proposed conformational states of LgtC along its reaction pathway. .............. 209 Figure A.1: Reaction scheme of enzymatic synthesis of UDP-2FGal. ................................. 234 Figure B.2: Crystal structure of UDGlcDH ternary complex with UDP-GlcA. ................... 237  xvii  List of Abbreviations Δω  Change in angular frequency  ε280nm  Molar absorptivity at 280 nm  2D  Two-dimensional  3D  Three-dimensional  A340  Absorbance at 340 nm  aa  Amino acid  ATP  Adenosine triphosphate  BSA  Bovine serum albumin  CAZy  Carbohydate active enzymes  CD  Circular dichroism  CGE  Covalent glycosyl-enzyme  CMP  Cytidine-5’-monophosphate  CMP-Neu5Ac CMP-3FNeu5Ac  Cytidine-5’-monophospho-N-acetyl-neuraminic acid CMP-3-fluoro-N-acetyl-neuranminic acid  CMPG  Carr-Purcell-Meiboom-Gill  COSY  Correlation spectroscopy  CSA  Chemical shift anisotropy  CstI  Campylobacter α-2,3-sialyltransferase  CstII  Campylobacter α-2,3/2,8-sialyltransferase  CTP  Cytidine-5’-triphosphate  Da DMSO  Dalton Dimethyl sulfoxide  DNA  Deoxyribonucleic acid  DTT  Dithiothreitol  EC  Enzyme commission (classification number) of the International Union of Biochemistry  EDTA  Ethylenediamine tetraacetic acid  HSQC  Heteronuclear single quantum correlation  ESI-MS  Electrospray ionization mass spectroscopy xviii  FID FPLC  Free induction decay Fast protein liquid chromatography  Gal  Galactose  GBS  Gullain-Barré syndrome  Glc  Glucose  GlcNAc  2-acetamido-2-deoxyglucose  GH  Glycoside hydrolase  GT  Glycosyltransferase  HEPES  2-[4-(2-hydroxyethyl)piperazin-1-yl]ethanesulfonic acid  HMQC  Heteronuclear multiple quantum correlation  HPLC  High-performance liquid chromatography  HSQC  Heteronuclear single quantum correlation  IPTG  Isopropyl β-D-thioglucopyranoside  Ka  Acid dissociation equilibrium constant  kcat  Catalytic rate constant (turnover number)  kcat/Km  Apparent second-order rate constant (specificity constant)  Kd  Dissociation constant  kDa  kiloDalton  Ki  Equilibrium inhibitor binding constant  Km  Michaelis constant  Lac  Lactose  LacNAc LB  N-acetyl-lactosamine Luria Bertoni broth  LDH  Lactate dehydrogenase type III from rabbit muscle  LgtC  Lipopolysaccharide α-1,4-galactosyltransferase C  LOSs  Lipooligosaccharides  LPSs  Lipopolysaccharides  MALDI-TOF  Matrix-assisted laser desorption ionization time-of-flight  MES  2-(N-morpholino)ethanesulfonic acid  MFS  Miller-Fisher syndrome  MW  Molecular weight xix  MWCO  Molecular weight cut-off  NAD+(ox)  Oxidized β-nicotinamide adenine dinucleotide  NADH(red)  Reduced β-nicotinamide adenine dinucleotide  NaOAc  Sodium acetate  NDP  Nucleotide diphosphate  NMR  Nuclear magnetic resonance  NOE  Nuclear Overhauser effect  O.D.600  Optical density at 600 nm  ORF  Open reading frame  PBS  Phosphate buffered saline  PCR  Polymerase chain reaction  PDB  Protein data bank (http://www.pdb.org)  PEG  Polyethylene glycol  PEP  Phosphoenolpyruvate  pH*  The pH meter reading without correction for isotope effects  PK  Pyruvate kinase type III from rabbit muscle  pKa*  The apparent pKa value without correction for isotope effects  ppm  Parts per million  PRE  Paramagnetic relaxation enhancement  Rex  Conformational exchange rate  Rf  The ratio of the distance a spot travels from the point of origin to the distance the solvent travels along a TLC plate  rmsd SDS-PAGE  Root mean square deviation Sodium dodecyl sulfate polyacrylamide gel electrophoresis  STs  Sialyltransferases  T1  Longitudinal relaxation time  T2  Transverse relaxation time  TEV  Tobacco Etch Virus  Tm  Mid-point unfolding temperature  Tris  2-Amino-2-hydroxymethyl-propane-1,3-diol  TROSY  Transverse relaxation optimized spectroscopy xx  TLC  Thin layer chromatography  UDP  Uridine-5’-diphosphate  UDP-2FGal  UDP-2’-deoxy-2’-fluoro-galactose  UDP-Gal  Uridine-5’-diphosphate galactose  UV/Vis  Ultraviolet/visible light  WT  Wild-type  YT  Yeast extract tryptone  xxi  List of Amino Acid Abbreviations A  Ala  Alanine  C  Cys  Cysteine  D  Asp  Aspartic acid  E  Glu  Glutamic acid  F  Phe  Phenylalanine  G  Gly  Glycine  H  His  Histidine  I  Ile  Isoleucine  K  Lys  Lysine  L  Leu  Leucine  M  Met  Methionine  N  Asn  Asparagine  P  Pro  Proline  Q  Gln  Glutamine  R  Arg  Arginine  S  Ser  Serine  T  Thr  Threonine  V  Val  Valine  W  Trp  Tryptophan  X Y  Unnatural amino acid Tyr  Tyrosine  xxii  Acknowledgements Firstly, I want to thank my research supervisors, Drs. Lawrence McIntosh and Stephen Withers, for the opportunity to work on this challenging and interesting research project and for their scientific guidance. They are friendly and helpful all the time and I learned quite a lot from them. They both trained me from a naive undergraduate to become a mature scientist. I am happy and feel proud to have them to be my research co-supervisors. At the same time, I want to give a special thank you to Dr. McIntosh for his encouragements during periods of ups-and-downs in my research progress and personal life throughout entire graduate studies. I also want to thank Lloyd Mackenzie and Dr. David Vocadlo for their recommendations and suggestions to start my graduate studies at UBC. Without them, I would not have pursued a scientific career and have the success of today. Secondly, I want to thank my whole family, especially my Dad and Mom, for all their unlimited support. Their decision to immigrate to Canada, and sacrificing their business in Hong Kong as a result, meant that I am able to have access to such a good educational environment in order to pursue my PhD degree. They represent an excellent parental model that I hope I can strive towards in the future. I also want to thank my brother, Danny, for spending time to do final review on this thesis. More importantly, I want to thank my wife, Annie, for her understanding, support, forgiveness, and care during the last few years of my graduate studies. Especially, during the time of this thesis preparation, her mentally support is essential for me to complete the work. I am grateful to have her as my lifelong partner. Lastly, I want to thank all the present and former members from the McIntosh and Withers laboratories. They all helped me a lot in my scientific research and made my time enjoyable in the laboratories. Specifically, I want to thank Dr. Mark Okon for his help and advice in setting up my NMR experiments, and Dr. Soumya De and Adrienne Cheung for their efforts in reviewing this thesis. I also want to thank Drs. David Poon, Hyun-Seo Kang, and Eric Escobar for their help and encouragement in my personal and scientific life.  xxiii  Chapter 1 General introduction  1  1.1  GLYCOBIOLOGY The central dogma of molecular biology is the sequential transfer of information from  DNA to RNA and from RNA to proteins (Crick 1958; Crick 1970). However, numerous additional steps, including post-translational modifications (PTMs) of both nucleic acids and proteins are required to explain the complexity of life. Some common examples of PTMs are phosphorylation, glycosylation, acetylation, and methylation, all of which are involved in complicated networks critical to cellular development and growth. Glycosylation is a widely studied PTM that is essential for many cellular functions. The resulting glycoconjugates, such as glycolipids, glycoproteins, and proteoglycans, play important roles in processes ranging from host-pathogen interactions to the immune response (Dwek 1996). Therefore, it is essential to understand how glycoconjugates are synthesized and degraded, where they are located, and how they interact with other biomolecules. Glycobiology is the study of carbohydrates in biological events. Alterations in the glycosylation patterns of proteins or cell membrane are linked to serious diseases, such as Alzheimer’s disease and cancer (Yuzwa, Macauley et al. 2008; Cazet, Julien et al. 2010). Bacteria also make use of the mimicry of the host cell membrane oligosaccharides in pathogenesis. That is, they avoid the host immune response by displaying host-like oligosaccharides on their cell wall (Moran, Prendergast et al. 1996; Penner and Aspinall 1997; Tzeng and Stephens 2000; Guerry, Szymanski et al. 2002; Abbott, Macauley et al. 2009; Nothaft and Szymanski 2010). Hence, glycans and the enzymes that synthesize and degrade them are potential therapeutic targets. Carbohydrates are also critical to metabolism, and can be exploited for food and biofuel production (Bardgett, Freeman et al. 2008; Wilson 2009; Chen and Qiu 2010; Liu and Khosla 2010; Yeoman, Han et al. 2010; Zinoviev, Muller-Langer et al. 2010). Extensive research is focused on this source of energy, which can be recycled and is environmentally favorable. The importance of this work is highlighted by the recent Deep Water Horizon oil spill and the Fukushima nuclear catastrophe, showing that conventional energy sources, such as gasoline and nuclear energy, are not only limited but also dangerous.  2  1.2  GLYCOSIDE HYDROLASES AND GLYCOSYLTRANSFERASES Since carbohydrates are important in medicine and biotechnology, it is important to  understand the glycosyltransferases and glycoside hydrolases that catalyze their synthesis and degradation, respectively.  1.2.1  Glycoside hydrolases Glycoside hydrolases (GHs) are the enzymes that cleave the glycosidic bonds in  glycosides, glycans, and glycoconjugates. Based on their amino acid sequence similarities, GHs are currently classified into 130 families according to the Carbohydrate Active Enzymes Database (CAZy) (Cantarel, Coutinho et al. 2009). The families can be further grouped into clans sharing common tertiary structures and catalytic mechanisms. In mechanistic terms, the GH families are classified as inverting or retaining. An inverting GH hydrolyzes the glycosidic bond with inversion of the stereochemical configuration of the anomeric carbon of the glycone sugar (Figure 1.1). This typically results from glycosidic bond cleavage via a single SN2 reaction in which a catalytic general base abstracts the proton from a water molecule, thereby facilitating its nucleophilic attack on the anomeric carbon of the glycoside. In addition, a catalytic general acid donates its proton to the aglycone leaving group. In contrast, a retaining GH commonly uses a double SN2 displacement or “ping-pong” mechanism to cleave the glycoside bond. This usually involves a nucleophilic carboxylate, which attacks the anomeric carbon of the glycoside to form a glycosyl-enzyme intermediate, while a general acid donates a proton to the leaving group. In the second step, the now deprotonated base abstracts the proton from a water molecule and facilitates its nucleophilic attack on the anomeric carbon of the glycoside to form the final glycone product and regenerate the unmodified enzyme. The two sequential Walden inversions yield net retention of glycone anomeric stereochemistry.  3  Figure 1.1: General enzymatic mechanisms used by inverting and retaining GHs. The glycone is shown as a schematic carbohydrate and the aglycone as the R group.  1.2.2  Glycosyltransferases Glycosyltransferases (GTs) are the enzymes that catalyze the incorporation of sugar  residues from an activated donor (such as a nucleotide sugar) into a glycosyl acceptor. They are also classified into families based on amino acid sequence relationships, as well as the stereochemical outcomes of their reactions (Breton, Snajdrova et al. 2006; Cantarel, Coutinho et al. 2009). The 94 GT families known to date can be divided into three classes with the GT-A, GT-B, and GT-C tertiary folds defined by X-ray crystallographic analyses (Figure 1.2) (Breton, Snajdrova et al. 2006). 1.2.2.1  General structures of GTs GT-B enzymes consist of two separate Rossmann folds, which are classic nucleotide-  binding domains, joined with a connecting linker region. The two domains face each other and create an active site in the intramolecular interface (Figure 1.2C&D). Appended loops or helices  4  are usually found to cover the active site and create the full binding site for sugar acceptors. The GT-A enzymes contain two Rossmann folds, yet these are closely associated as a single domain (Figure 1.2). Enzymes of this class frequently contain an Asp-x-Asp (or DXD, x is any amino acid) motif, which coordinates a divalent metal ion, such as Mn 2+ or Mg2+, that is essential for its activity. The GT-A fold also contains a small flexible loop region that covers the active site when the sugar donor is bound, thereby creating a full sugar acceptor binding site. GT-C is a hypothetic fold for hydrophobic integral membrane GTs predicted through iterative BLAST searches (Liu and Mushegian 2003; Rosen, Edman et al. 2004). Since these membrane-bound proteins are difficult to express, purify, and crystallize, only two crystal structures with this GTC fold were solved. They are the soluble C-terminal domain of Pyrococcus furiosius oligosaccharyltransferase STT3 (Igura, Maita et al. 2008) and peptidoglycan synthesizing glycosyltransferase (Lovering, de Castro et al. 2007; Yuan, Barrett et al. 2007).  5  Figure 1.2: Crystal structures of representative GTs from different GT families. The GT catalytic domains are colored in blue (α-helices) and red (β-strands). (A) Rabbit Nacetylglucosaminyltransferase I (GnT1) belongs to family GT13. It has a GT-A fold and uses an inverting mechanism to transfer GlcNAc from UDP-GlcNAc to form β-1,2 linkage with the α1,3-linked arm of the trimannose core of Man2Gn2-β-Asn-containing substrates (Gordon, Sivarajah et al. 2006) (PDB code: 2AM4). (B) Human UDP-GalNAc:polypeptide α-Nacetylgalactosaminyltransferase (ppGalNAcT) belongs to family 27. It also has a GT-A fold and uses a retaining mechanism to transfer GalNAc from UDP-GalNAc to a peptide acceptor. This enzyme also contains a separate lectin-binding domain (cyan and yellow) (Fritz, Raman et al. 2006) (PDB code: 2FFU). (C) Human O-GlcNAc transferase (OGT) belongs to family 41. It has a GT-B fold and uses an inverting mechanism to transfer GlcNAc from UDP-GlcNAc to serine and threonine residues of proteins (Lazarus, Nam et al. 2011). OGT is a multi-domain enzyme and also has a tetratricopeptide repeat (TPR) domain (green) and an intervening domain (cyan and yellow) (PDB code: 3PE3). (D) α-Glucosyltransferase (AGT) from E. coli T4 bacteriophage belongs to family 72. It has a GT-B fold and uses a retaining mechanism to transfer glucose from UDP-glucose to 5-hydroxylmethyl cytosine (5-HMC) bases of duplex DNA (Lariviere, Sommer et al. 2005) (PDB code: 1XV5).  6  1.2.2.2  Enzymatic mechanisms of GTs In mechanistic terms, GTs can be divided into two classes based on the stereochemical  outcome of their reactions (Figure 1.3) (Lairson, Henrissat et al. 2008). As with GHs, a GT that inverts the stereochemistry of the anomeric centre of the sugar donor is called an inverting enzyme, whereas one that retains the stereochemistry of the glycosyl linkage is called a retaining enzyme. It is well established that all inverting enzymes characterized to date use a single displacement SN2-like reaction. The hydroxyl group of the sugar acceptor is deprotonated by a general base in the active site of the enzyme, facilitating direct nucleophilic attack on the anomeric carbon of the sugar donor (Figure 1.4). By analogy to well characterized glycoside hydrolases, a retaining GT would be expected to use a double displacement mechanism with two sequential SN2-like reactions leading to the formation and breakdown of a covalent intermediate (Figure 1.5). However, as discussed below, evidences support an internal return ion pair SNi-like or SNi mechanism (Figure 1.5) (Lairson, Henrissat et al. 2008; Gomez, Polyak et al. 2012).  Figure 1.3: Enzymatic reactions catalyzed by inverting and retaining GTs. Both inverting and retaining GTs uses nucleotide activated sugar donors, such as CMP-Neu5Ac, UDP-Glc, or UDP-Gal, to transfer the sugar moiety onto the sugar acceptor. This figure is adapted from (Lairson, Henrissat et al. 2008).  7  Figure 1.4: Proposed reaction mechanism by an inverting GT. The catalytic general base (B) of the GT abstracts the proton from the sugar acceptor to facilitate its SN2 nucleophilic attack on the anomeric carbon of the sugar donor. The catalytic general acid (A) donates a proton to the leaving group NDP (nucleotide diphosphate). In order to test if a retaining GT uses a double displacement mechanism, researchers have attempted to trap and characterize the postulated covalent glycosyl-enzyme intermediate, for example using substrate analogs or mutant enzymes. To date, there have been two reports of glycosyl-enzyme intermediates covalently trapped with mutant retaining GTs. One of those is LgtC with the mutation of Q189E (Lairson, Chiu et al. 2004). LgtC is lipooligosaccharide α-1,4galactosyltransferase from Neisseria meningiditis that catalyzes the incorporation of UDPgalactose as sugar donor into lactose as sugar acceptor to form the trisaccharide product. From X-ray crystallographic work, it was proposed that Gln189 might be the catalytic nucleophile based on its proximity to the anomeric carbon of the sugar donor (Persson, Ly et al. 2001). Mutation of Q189E was therefore performed with the logic of obtaining a more stable covalent glycosyl-enzyme (CGE) intermediate due to the change from an amide to a carboxylate nucleophile. Indeed, a covalently modified protein was detected by ESI-MS and thus intermediate could be “turned over” to product by addition of a lactose acceptor sugar. However, the galactose in the intermediate was found to be linked to Asp190 and not Glu189 (Lairson, Chiu et al. 2004). This was surprising as Asp190 is 8.3 Å from the anomeric centre of the sugar donor in the wild-type enzyme, and thus a significant conformational change would be required  8  in the mutant species for covalent bond formation. At the same time, the kinetic studies of Q189A and D190A mutants showed residual activity of the enzymes, whereas mutation of a true catalytic nucleophile should be completely inactivating. These findings made the proposed catalytic mechanism of LgtC with Gln189 or Asp190 as a nucleophile controversial. The other mutant retaining GTs in which CGE intermediates were trapped and identified by ESI-MS are α-1,3-N-acetylgalactosaminyltransferase (GTA) and α-1,3-galactosyltransferase (GTB) (Soya, Fang et al. 2010). GTA and GTB synthesize the cell surface oligosaccharides for ABO blood group system. In the two closely related enzymes, the postulated nucleophile Glu303 was mutated to Cys and their CGE intermediates were trapped. As with LgtC, the mutations of GTA and GTB might alter their natural enzymatic mechanisms (Goedl and Nidetzky 2009; Gomez, Polyak et al. 2012). Importantly, a CGE intermediate using a substrate analog with a wild-type retaining GT has yet to be reported.  1.2.2.3  Proposed alternative SNi or SNi-like mechanisms of retaining GTs Detailed kinetic and structural studies of retaining GTs point to a SNi-like mechanism  involving a short-lived ion pair with the donor leaving group and acceptor nucleophile positioned on the same side of the anomeric carbon (Figure 1.5) (Breton, Snajdrova et al. 2006; Lairson, Henrissat et al. 2008). LgtC (Persson, Ly et al. 2001), Kre2 (Lobsanov, Romero et al. 2004), ToxB (Reinert, Jank et al. 2005), Extl2 (Pedersen, Dong et al. 2003), Mgs (Flint, Taylor et al. 2005), WaaG (Martinez-Fleites, Proctor et al. 2006), OtsA (Gibson, Turkenburg et al. 2002), and AGT (Sommer, Depping et al. 2004) are the retaining GTs that have defined structures and appear to use an SNi-like mechanism for glycosyltransfer. However, the SNi and SNi-like mechanisms used by retaining GTs also remain somewhat controversial. More recently, by using quantum mechanics/molecular mechanics (QM/MM) simulations, an S Ni-like mechanism involving a short-lived ion pair was proposed for OtsA (Ardevol and Rovira 2011) and an SNi mechanism involving an oxocarbenium ion-like transition state was proposed for LgtC (Figure 1.5) (Tvaroska 2004; Gomez, Polyak et al. 2012).  9  Figure 1.5: Proposed double SN2 displacement mechanism (top), SNi-like (middle), and SNi reaction mechanisms (bottom) used by retaining GTs. A retaining GTs appear to follow SNi-like or SNi mechanisms rather than double SN2 displacement mechanism. (NDP: nucleotide diphosphate; Nuc: nucleophile)  10  1.3 1.3.1  STRUCTURAL DYNAMICS OF GTs Protein structural dynamics Protein exhibits a wide range of dynamic motions involving bond vibrations, side-chain  rotation, loop motions, and larger domain motions. These range from the psec to sec (or longer) time-scale, respectively (Henzler-Wildman and Kern 2007). Aspects of these dynamics are proposed to be important in enzymatic function (Bennett and Huber 1984; Joseph, Petsko et al. 1990; Yon, Perahia et al. 1998; Todd, Orengo et al. 1999; Kumar, Ma et al. 2000; Zhao 2010). Conformational changes are paramount to protein-protein interactions, protein-ligand interactions, and catalytic mechanisms. For a classic example, hexokinase was shown to change from an “opened” to “closed” conformation upon glucose binding. It was proposed that this prevents unnecessary, futile hydrolysis of the substrate ATP (Steitz, Shoham et al. 1981; Aleshin, Fromm et al. 1998). Conformational dynamics can also help to stabilize the transition state during enzymatic catalysis. For example, theoretical calculations suggested that conformational changes are required to stabilize the transition state in chorismate mutase (Crespo, Scherlis et al. 2003; Strajbl, Shurki et al. 2003). Structural dynamics are also key to allosteric and cooperative effects (Yonetani and Laberge 2008). Hence, it is important to study protein structural dynamics in order to understand enzymatic catalysis.  1.3.2  Glycosyltransferase dynamics In order to fully investigate how GTs catalyze the formation of a glycosidic linkage,  studies of their structural dynamics are mandatory. Crystal structures of several GTs generally indicate that active site loop regions are flexible in the apo state and become more ordered upon substrate binding. For example, the crystal structures of bifunctional CstII, α-2,3/2,8sialyltransferase from Campylobacter jejuni, show a conformational change in a loop domain upon binding its substrate (Figure 1.6) (Chiu, Watts et al. 2004). The electron density of this flexible domain is missing when only product CMP is bound. In contrast, when the substrate analog CMP-3FNeu5Ac is bound, the electron density of this flexible motif (residues 175-187) appears in one of four copies of the enzyme in the crystal asymmetric unit, indicating it becomes more ordered with substrate binding (Figure 1.6).  11  Figure 1.6: Crystal structures of the monomer subunits of the tetrameric CstII. (A) CstII complexed with CMP (PDB code: 1RO8) and (B) CstII complexed with CMP3FNeu5Ac (PDB code: 1RO7). The flexible motif (residues 175-187) is colored in green. CMP and CMP-3FNeu5Ac are shown (carbon, grey; oxygen, red; nitrogen, blue; phosphorus, orange). The crystal structures of the LgtC binary complex with only the sugar donor analog UDP-2FGal and its ternary complex with both the sugar donor analog UDP-2FGal and sugar acceptor analog 4’-deoxylactose also revealed the potential ordering of two flexible loops (residues 75-80 and residues 246-251) upon substrate binding (Figure 1.7) (Persson, Ly et al. 2001). Unfortunately, crystals of the apo form of LgtC could not be obtained, as needed to more fully understand this conformational change. Indeed, the authors proposed that apo LgtC form could not be crystallized because of the dynamic flexibility of the two loops that cover its active site. This hypothesis has been supported through in silico experiments (Snajdrova, Kulhanek et al. 2004; Tvaroska 2004; Gomez, Polyak et al. 2012). Collectively, these studies indicate that the dynamic properties of the flexible loop in LgtC dampen upon substrate binding.  12  Figure 1.7: Overlaid crystal structures of LgtC binary and ternary complexes. Overlaid crystal structures of LgtC binary complex with UDP-2FGal (blue) (PDB code: 1G9R) and ternary complex with UDP-2FGal and 4’-deoxylactose (yellow) (PDB code: 1GA8) shows that the two states of LgtC are essentially the same. The flexible loops are colored in green. UDP-2FGal and 4’-deoxylactose are shown (carbon, cyan and yellow; oxygen, red; nitrogen, blue; phosphorus, orange). There are numerous ways to study enzyme dynamics, of which NMR spectroscopy (Wand 2001) and molecular dynamics simulations (Doniach and Eastman 1999) are generally the most useful. Unfortunately, almost all known 3D structures of GTs have been obtained by Xray crystallography. The only two NMR-derived structures of GTs reported to date are the bipartite  glycosyltransferase  (Alg13)  and  oligosaccharyltransferase  (OST)  (Wang,  Weldeghiorghis et al. 2008; Gayen and Kang 2011). However, the NMR-derived structure of Alg13 just represents the glycosyl donor-binding domain of the enzyme without its catalytic domain, whereas that of OST (Ost4) just represents the transmembrane subunit of the enzyme without its catalytic domain. There are several reasons why it is difficult to determine the structures of GTs by NMR spectroscopy. GTs are typically “large” enzymes (i.e. > 30 kDa and often oligomeric), and thus challenging to study even with the most current NMR technologies. Large proteins tumble slowly and their spins undergo fast transverse relaxation, which causes peak broadening. At the same time, the number of resonances increases with protein size, thus leading to increased spectral complexity. Although many new methods and state-of-the-art 13  technologies in NMR have been developed for studying large proteins, such as selective deuteration and  13  C-methyl labeling combined with TROSY experiments (Kay 2011), dealing  with the GTs remains problematic. Furthermore, GTs are often membrane-associated and thus have limited solubility even when their membrane-associating regions are deleted. GTs appear to exhibit a range of dynamics, especially in the active site, on the msec-μsec time-scale that can lead to conformational exchange broadening of NMR signals. Indeed, the crystal structures of these enzymes are usually solved with sugar donor and sugar acceptor bound. In the apo forms of GTs, the electron density of flexible active site regions is usually missing, indicating significant dynamic motions when there is no substrate bound. This is also one of the reasons why the apo forms of GT are often difficult to crystallize. Fortunately, structure dynamics data on GTs can be obtained by complementary use of NMR spectroscopy and X-ray crystallography.  14  1.4  SIALYLTRANSFERASE CSTI/II FROM CAMPYLOBACTER JEJUNI Sialyltransferases (ST's) in the Golgi transfer the essential 9-carbon sugar sialic acid, or  N-acetyl-neuraminic acid (Neu5Ac), onto various acceptors such as lactose. The resulting products are then displayed on the cell surface in the form of glycolipids and glycoproteins. Sialic acids are involved in many important physiological functions, including cell growth and differentiation, cell-cell interactions (Lloyd and Furukawa 1998), and bacterial and viral pathogenesis (Razi and Varki 1998). Examples include sialic acids within lipo-oligosaccharides (LOS's) on the bacterial cell surface (Griffiss, Schneider et al. 1988), and in the gangliosides of human nerve cells (Figure 1.8) (Vyas and Schnaar 2001; Ang, Jacobs et al. 2004). Campylobacter jejuni is a common human food-borne pathogen that causes diarrhea and can give rise to neurodegenerative autoimmune diseases such as Guillain-Barré and Miller-Fisher syndromes (Endtz, Ang et al. 2000). The autoimmune reaction is directed to gangliosides in nerve tissue as C. jejuni displays sialylated LOS's that mimic the gangliosides on human cells (Yuki, Taki et al. 1993; Yuki and Odaka 2005; Godschalk, Kuijf et al. 2007; Yuki 2010). Thus, ST's are also virulence factors, closely linked to the potential of microorganisms to cause serious disease.  15  Figure 1.8: Structures of LOS of C. jejuni and ganglioside GM1 of human. C. jeiuni uses its LOS to “hide” from the immune system of the human host by mimicking ganglioside GM1. The figure is modified from (Ang, Jacobs et al. 2004). 1.4.1  CstI and CstII Among the ST's identified in C. jejuni, crystal structures of the monofunctional α-2,3-  sialyltransferase (CstI) (Figure 1.9A) and the bifunctional α-2,3/2,8-sialyltransferase (CstII) (Figure 1.9B) have been solved (Chiu, Watts et al. 2004; Chiu, Lairson et al. 2007).  16  Figure 1.9: Crystal structures of the monomer subunits of two tetrameric sialyltransferases from Campylobacter jejuni Ribbon diagrams of (A) CstI (PDB code: 2P2V) and (B) CstII (PDB code: 1RO7) complexed with CMP-3FNeu5Ac (carbon, grey; oxygen, red; nitrogen, blue; phosphorus, orange). The flexible motifs (residues 171-199 for CstI and residues 175-187 for CstII) are identified in green. The proposed catalytic bases His202 and His188 of CstI and CstII, respectively, are colored in red.  Both CstI and CstII belong to GT family 42 and use an inverting mechanism for sialyltransfer, forming an α-2,3-linkage to acceptor sugars containing a terminal galactose residue (Figure 1.10). Although the sequences of CstI and CstII are only 42% identical, they use conserved amino acid residues for substrate binding and catalysis and their overall structures in complex with a inactive donor substrate analog, CMP-3FNeu5Ac, are similar (Chiu, Lairson et al. 2007). Both CstI and CstII exhibit a GT-A variant “Cst” fold, which lacks a DXD motif and differs in the exact connectivity of β-strands (Chiu, Watts et al. 2004; Chiu, Lairson et al. 2007). They are tetrameric, and each monomer subunit contains two domains (Figure 1.9) (Chiu, Watts et al. 2004; Chiu, Lairson et al. 2007). One consists of a mixed α/β Rossmann fold nucleotidebinding domain. The second domain, composed of a long coil and two helices, forms a flexible “lid” over the active site. This lid becomes more ordered in the crystal structures of CstI/II upon binding with CMP-3FNeu5Ac (Chiu, Watts et al. 2004; Chiu, Lairson et al. 2007). Note that CstII follows an iso-ordered bi-bi kinetic mechanism with CMP-Neu5Ac binding before the sugar acceptor (Lairson, Henrissat et al. 2008; Lee, Lairson et al. 2011).  17  Figure 1.10: Sialyltransfer reaction catalyzed by CstII (A) Inverting sialyltransferase reaction catalyzed by CstII. Monofunctional and bifunctional CstII both catalyze formation of an α-2,3 linkage between donor CMP-Neu5Ac and an acceptor. The bifunctional CstII can subsequently add a α-2,8 linked-Neu5Ac to the initial product. (B) Wild-type CstII active site highlighting the side chain imidazole of His188 suitably positioned on the -face of the donor substrate analog CMP-3FNeu5Ac to play the role of base catalyst. Atoms are colored according to type (C grey or yellow, N blue, O red, F cyan, P orange) and the carboxyl (C1) and anomeric (C2) carbons of CMP-3FNeu5Ac are labeled. 1.4.2  Monofunctional and bifunctional CstII There are monofunctional (OH19) and bifunctional (OH4384) variants of CstII (Gilbert,  Karwaski et al. 2002). In addition to α-2,3-sialyltransfer, bifunctional CstII also has α-2,8sialyltransferase activity, adding an additional Neu5Ac to the initially incorporated α-2,3-linked sialoside (Figure 1.10). The sequences of the two CstII variants are 97.3% identical (Gilbert, Brisson et al. 2000), and thus the difference of only eight amino acid residues determines the additional activity in the latter enzyme (Gilbert, Karwaski et al. 2002).  1.4.3  Catalytic His188 of CstII Prominent among the active site residues of CstI and CstII is a histidine (His202 and  His188, respectively) proposed to serve as the general base that abstracts the proton from the  18  nucleophilic hydroxyl group of the sugar acceptor, thereby facilitating attack on CMP-Neu5Ac (Figure 1.10). This is a somewhat unusual residue to serve as the base for glycosyl transfer reactions. In essentially all glycoside hydrolases, and in the majority of inverting nucleotide phospho-sugar GTs, this role is performed by glutamate or aspartate residues (whereas retaining GTs appear to use the departing phosphate moiety (Lairson, Henrissat et al. 2008)). Within the GTs, histidine residues have only been proposed as general bases in GT 42 STs, along with a subgroup of the GT-B fold family GT 1 inverting enzymes (Lairson, Henrissat et al. 2008). Although mutational studies have confirmed their overall importance for catalysis, no definitive evidence for these histidine residues serving as a base had been provided prior to the studies presented herein.  1.4.4  Substrate binding of CstII The crystal structures of bifunctional CstII with the sugar donor analog CMP-3FNeu5Ac,  as well as monofunctional CstI in both its apo form and with this bound analog, provided important insights into the mode of substrate binding and testable hypotheses regarding the mechanism of sialyl transfer (Chiu, Watts et al. 2004; Chiu, Lairson et al. 2007). Recently, the crystal structure of CstII complexed with CMP and the terminal trisaccharide of its natural sugar acceptor (Neu5Ac-α-2,3-Gal-β-1,3-GalNAc) has also been obtained (Figure 1.11) (Lee, Lairson et al. 2011). Asn51, Tyr81, and Arg129 were found to be important residues to interact with the trisaccharide. A comparison with the crystal structure of CstII complexed with CMP-3FNeu5Ac shows no significant global conformational differences. However, part of the lid domain (residues 181-190 including His188) of the CstII complexed with CMP and the trisaccharide is disordered in the crystal structure (i.e. missing electron density). The CMP moiety of the CstII•CMP-3FNeu5Ac complex overlaps completely with the CMP in the ternary complex of CstII with this nucleotide monophosphate and the trisaccharide (Figure 1.11B). The Neu5Ac moiety adopts different positions in the two structures, depending on whether it is present within CMP-3FNeu5Ac or the trisaccharide. This structure can be viewed as the product complex of the α-2,3-linkage reaction of CMP-Neu5Ac with Gal-β-1,3-GalNAc, with the resulting trisaccharide shifted to the acceptor site for a subsequent α-2,8-linkage reaction. In addition, the superimposed structures of both complexes (by ignoring the CMP) can be viewed as the reactant complex of the α-2,8-linkage reaction of CMP-Neu5Ac with Neu5Ac-α-2,3-Gal-β-1,3-GalNAc. 19  Figure 1.11: Superimposed structures of CstII with bound donor and acceptor analogs. (A) Superimposed structures of CstII complexed with sugar donor analog CMP-3FNeu5Ac (yellow and magenta) (PDB code: 1RO7) and CstII complexed with CMP and the terminal trisaccharide of its natural sugar acceptor (Neu5Ac-α-2,3-Gal-β-1,3-GalNAc) (cyan) (PBD code: 2X61). Carbon atoms of CMP-3FNeu5Ac are colored in pink. Carbon atoms of CMP and Neu5Ac-α-2,3-Gal-β-1,3-GalNAc are colored in green. All non-carbon atoms are colored according to type (N blue, O red, F, cyan, P orange). The lid domain (residues 155-188), which is not observed in the X-ray crystallographic structure of CstII complexed with CMP and trisaccharide, is in magenta. (B) Close-up view of the active sites of both structures shows the arrangement of the sugar donor analog and sugar acceptor analog. The CMP moieties superimpose in both structures, whereas the Neu5Ac differs depending on whether it is present in the donor or acceptor. 1.4.5  Structural dynamics of CstII CstII contains two domains (Chiu, Watts et al. 2004). One domain consists of a mixed  α/β fold, which forms the nucleotide-binding domain in the form of a Rossmann fold. The second is composed of a long coil and two helices, which forms a flexible “lid” domain. The electron density of this lid is missing in structures of CstII complexed with CMP (PDB code: 1RO8) and with both CMP and the trisaccharide Neu5Ac-α-2,3-Gal-β-1,3-GalNAc (PDB code: 2X61) and detected crystallographically only after binding of CMP-3FNeuAc (PDB code: 1RO7). Closure of the flexible lid domain, which contains the catalytic base His188, appears required for sialyl-transfer and to protect the active site of the enzyme from bulk solvent, thereby minimizing undesirable reactions such as hydrolysis of the donor substrate. However, beyond the presence or absence of electron density from the lid domain in crystallized CstII, the time-scales and amplitudes of its motions have not been characterized.  20  1.5  GALACTOSYLTRANSFERASE LGTC FROM NEISSERIA MENINGITIDIS Neisseria meningitidis is a Gram-negative bacterium that causes meningitis, an  inflammation of the protective membranes covering the central nervous system. (Tzeng and Stephens 2000) The pathogenesis of N. meningitidis, with 10% fatality levels, is dependent upon the bacterium mimicking human lacto-N-neotetraose via its cell wall LOSs (Figure 1.12) (Zhu, Klutch et al. 2002). Parts of the LOSs are synthesized by GTs encoded from three genetic loci (lgt-1, 2, and 3), in which 7 ORFs (lgtcA, lgtB, lgtC, lgtD, lgtE, lgtH, and lgtZ) are in a single lgt1 locus. By studying the structures and catalytic mechanisms of these GTs, general insights into the therapeutic approaches can be obtained.  Figure 1.12: Structure of LOS of N. meningitidis. LOS of N. meningitidis contains lacto-N-neotetrose structure (Galβ1-4GlcNAcβ1-3Galβ1-4Glc) that mimics the immune response in human host. The figure is modified from (Zhu, Klutch et al. 2002).  1.5.1  Structure of LgtC Lipooligosaccharide α-1,4-galactosyltransferase (LgtC) is responsible for the transfer of  α-galactose from sugar donor UDP-Gal to the LOS terminal sugar acceptor lactose. LgtC belongs to GT family 8 and uses a retaining mechanism for galactosyltransfer. The structures of LgtC with a UDP-2-deoxy-2-fluoro-galactose (UDP-2FGal) sugar donor analog in the presence and absence of the acceptor sugar analog 4’-deoxylactose have been determined using X-ray 21  crystallography (Figure 1.13). (Persson, Ly et al. 2001) The LgtC binary structure (PDB code: 1G9R) and ternary structure (PDB code: 1GA8) are essentially the same. The LgtC used for these studies was modified by deletion of a C-terminal membrane association domain, and by mutation of several cysteine residues to prevent the reversible aggregation of the protein. The crystal structure of monomeric LgtC consists of a large N-terminal mixed α/β domain, containing the active site, and a smaller C-terminal helical domain, which mediates membrane attachment.  4’-deoxylactose  Mn2+ Gln189 UDP-2FGal  Asp190  Figure 1.13: Crystal structure of the LgtC ternary complex. A ribbon diagram of the LgtC ternary complex (PDB code: 1GA8) with the proposed flexible loops (residues 75-80 and 246-251) colored in green. The active site residues Gln189 and Asp190 are identified in blue and red, respectively. Also shown are UDP-2FGal and 4’deoxylactose (carbon, grey; oxygen, red; nitrogen, blue; phosphorus, orange; Mn2+, magenta).  22  1.5.2  Substrate binding of LgtC From kinetic analyses, LgtC follows an ordered bi-bi kinetic mechanism with UDP-Gal  binding before the sugar acceptor (Ly, Lougheed et al. 2002). The full sugar acceptor binding site is proposed to be only formed upon the binding of a sugar donor with the concomitant ordering of two dynamic loop regions. LgtC has four “DXD” motifs, one of which coordinates Mn2+ or Mg2+ as required for enzymatic activity. Mn2+ appears to be the preferred metal ion for LgtC, although its specific activity (k cat/Km) only decreases to ~50% when substituted by Mg2+ (Lougheed 1998). Specificity of the sugar donor binding of LgtC was previously investigated through kinetic analyses (Ly, Lougheed et al. 2002). LgtC was found to bind and use UDP-Gal (kcat/Km = 0.5 μM-1 s-1) more specifically than UDP-Glc (kcat/Km = 0.003 μM-1 s-1). Interestingly, in the absence of sugar acceptor, UDP-Gal could be also hydrolyzed by LgtC (k cat/Km = 0.004 μM-1 s1  ).  1.5.3  Controversial enzymatic mechanism of LgtC As discussed above, based on the structure of LgtC and by analogy with  hexosaminidases, which use their own amide as nucleophile, the side-chain amide of Gln189 was initially and cautiously proposed as the catalytic nucleophile in this GT (Zechel and Withers 2000) (Figure 1.13). However, 3% residual enzyme activity was observed in a Q189A mutant lacking the glutamine side-chain, thus bringing into question the essential role of this residue. In an attempt to trap the putative glycosyl-enzyme intermediate, the LgtC-Q189E mutant was created based on the assumption that the carboxylate side-chain of glutamate would be a better nucleophile than the amide side-chain of glutamine and the glycosyl-enzyme intermediate should be more stable (Lairson, Chiu et al. 2004). ESI-MS was used to analyze a glycosylenzyme intermediate formed when the mutant was reacted with UDP-Gal and MS sequencing of proteolytic digests was done to identify the galactose-linked residue. Surprisingly, the trapped residue was Asp190 and not Glu189. In the crystal structure of the wild-type enzyme, the distance between the carboxylic side-chain of Asp190 and the anomeric carbon is ~ 9 Å. Also, the X-ray crystal structure of LgtC-Q189E with Mn2+•UDP-2FGal confirmed that its active site is essentially identical to that of the wild-type protein (Lairson, Chiu et al. 2004). Thus formation of the covalent bond between the carboxylic side-chain of Asp190 and galactose in LgtC-Q189E 23  would require a dramatic conformational change. Alternatively, the results of this study could be an artifact as the sample was acidified and unfolded/proteolyzed before MS analysis. Based on extensive kinetic studies, as well as recent theoretical calculations, LgtC appears to exploit an SNi-like mechanism (Figure 1.5) (Ly, Lougheed et al. 2002; Lairson, Chiu et al. 2004; Tvaroska 2004; Lairson, Henrissat et al. 2008; Gomez, Polyak et al. 2012). This “front-side attack” mechanism with net retention of stereochemistry would require at least localized conformational and electrostatic changes to stabilize a short-lived oxocarbeniumphosphate ion pair intermediate. This transition state has to be stabilized in a perfectly suitable electrostatic environment by rearrangement of the residues in the active site of LgtC. Hence, the dynamic properties of LgtC and its catalytic mechanism could be highly correlated.  1.5.4  Structural dynamics of LgtC Based on the crystal structure of the LgtC ternary complex, two potentially flexible loops  (residues 75-80 and 246-251) were proposed to become more ordered and cover the active site of the enzyme upon binding of Mn2+ and UDP-2FGal (Figure 1.13). The conformational change could therefore result in formation of the sugar acceptor binding site. However, there is no X-ray crystallographic structural information for apo LgtC to support or refute this hypothesis. MD simulations do suggest that the loop consisting of residues 246-251 is more flexible than that containing residues 75-80 when the sugar donor is absent (Snajdrova, Kulhanek et al. 2004). The two dynamic loops are proposed to be opened and in a different conformation when there is no sugar donor (Persson, Ly et al. 2001). This phenomenon is similar to that observed with CstII in that the density of its flexible loop is absent when there is no sugar donor (Chiu, Watts et al. 2004). In addition, the proposed SNi “front-side attack” mechanism of LgtC requires stabilization of the short-lived oxocarbenium-phosphate ion pair intermediate. The rearrangements and the dynamics of the residues in the active site are thus important in the catalytic mechanism of LgtC.  24  1.6  NMR SPECTROSCOPY AND LARGE MACROMOLECULES Two major problems usually arise when studying the structures of large protein  molecules using conventional NMR spectroscopy. First, the number of resonance signals increases with the size of the molecule. Second, rotational diffusion slows and the correlation time (τc) for tumbling increases with increasing size. This increases the transverse relaxation rates (Rxy = 1/T2) of the NMR signals from larger proteins (Figure 1.14), leading to linebroadening and eventually disappearance of the signals. In order to overcome these problems, several isotope labeling techniques and NMR methods have been developed. Most importantly, the relaxation rates of the amide and methyl signals of large proteins can be slowed by replacing the neighboring carbon-bonded protons with deuterons (Gardner and Kay 1998). In parallel, transverse relaxation-optimized spectroscopy (TROSY) methods can be used to acquire spectra of large, deuterated proteins (Figure 1.15) (Pervushin, Riek et al. 1997; Fernandez and Wider 2003).  25  Figure 1.14: Relationship between the correlation time for molecular motion and the relaxation rates of excited spins. The plot of the rate constants for longitudinal (R z) and transverse (Rxy) relaxation rates due to local, randomly fluctuating fields as a function of correlation time (τc). When the molecule undergoes fast motion (ω0τc << 1, ω0 and τc are the Larmor frequency and correlation time, respectively), Rz and Rxy will be equal and slow. In contrast, when molecular tumbling slows, which is typically the case for large proteins, Rz will reach a maximum at τc = 1/ω0 and decrease again with increasing τc, whereas Rxy will continue to increase. This figure is modified from (Keeler 2010).  26  Figure 1.15: TROSY effect in NMR signals of small and large proteins. Transverse relaxation time (T2 = 1/Rxy) of the NMR signals of (A) GB1 domain of protein G (PDB code: 2J52, MW = 6.2 kDa) is longer than that of (B) LgtC (PDB code: 1GA8, MW = 37 kDa) during the t1 and t2 acquisition times of a HSQC experiment. Thus, after a two-dimensional Fourier transformation (FT), the line-shape of a signal from the larger protein is broadened since its relaxation rate is faster. (C) With TROSY effect, only the slowly relaxing component of the NMR signal of the larger protein (LgtC) is retained, thus yielding a sharper peak in the transformed spectrum. The figure is modified from (Fernandez and Wider 2003).  27  1.6.1  Transverse relaxation-optimized spectroscopy A two-dimensional  15  N-HSQC (heteronuclear single quantum correlation) spectrum  contains peaks at the 1H and  15  N chemical shifts of each directly bonded 1H-15N pair in a  molecule. Decoupling is usually applied during the t 1 (for  15  N shift) and t2 (acquisition of 1H  signal) time periods in order to obtain a single peak and not a multiplet split in each dimension by the large one-bond J coupling (~92 Hz for an amide) between the two nuclei. However, due to the interplay (cross-correlation) of dipolar and chemical shift anisotropy (CSA) relaxation mechanisms, the transverse lifetimes of each component of the quartet differ significantly for larger molecules (Figure 1.16A & B). Thus, the TROSY approach effectively involves recording a 1H/15N-HSQC without decoupling in order to retain the sharp signal from the mostly slowly relaxing component of the J-coupled quartet by line-selective transfer, while discarding the broader signals from the faster relaxing components (Figure 1.16C). The TROSY effect is most pronounced for larger, slower tumbling proteins and for data recorded at higher magnetic fields, and benefits from the combined use of deuteration to remove the effects of dipolar relaxation due to nearby carbon-bonded protons. The TROSY approach can also be implemented into 1H-15N13  C triple resonance experiments for assigning the NMR signals of a protein, as well as into pulse  sequences to measure relaxation and hence protein dynamics (Zhu, Xia et al. 2000).  28  Figure 1.16: Schematic background of TROSY experiment. (A) Four energy levels of a two-spin system of spins I and S. For example, level 3 is labeled as spin I in β-state and spin S in α-state. The two I-spin transitions (1-3 and 2-4) are in black lines and the two S-spin transitions (1-2 and 3-4) are in blue lines. Due to the cross-correlation of dipolar and CSA relaxation, the relaxation rates of 1-2 and 1-3 transitions are faster than 2-4 and 3-4 transitions. (B) Schematic HSQC spectrum without decoupling shows four peaks with different peak intensities due to different relaxation rates. The cross peak created from fast relaxing components of 1-2 and 1-3 transitions broadened/disappeared. The cross peak created from slowly relaxing components of 2-4 and 3-4 transitions is the most intense. The intensities of the cross peaks created from combinations of fast and slowly relaxing components (either 1-3 and 3-4 or 2-4 and 1-2) are medium. (C) Schematic TROSY spectrum shows only one intense peak by line-selective transfer. The figure is modified from (Keeler 2010).  1.6.2  Methyl-TROSY spectroscopy Amide TROSY-based experiments significantly improve the spectral sensitivity and  resolution for larger proteins. However, when dealing with very large proteins, such as malate synthase G (MSG) (Tugarinov and Kay 2003; Korzhnev, Kloiber et al. 2004), which is an 82 kDa single polypeptide chain enzyme, and the 20S proteasome (Sprangers and Kay 2007) which is a 670 kDa multi-unit complex, this approach still suffers due to problems of spectral overlap, difficulties in obtaining signal assignments, and the need to prepare deuterated samples that are re-protonated at all amides (usually via unfolding and refolding in H2O buffer) (Tugarinov and Kay 2003). Accordingly, the method of methyl-TROSY was introduced as an alternative approach for investigating the structures and dynamics of large proteins (Ollerenshaw, Tugarinov et al. 2003). There are several advantages of using methyl-TROSY methods relative to amide  29  TROSY-based approaches. First, in comparison to the number of the signals from the backbone amides, the number of methyl signals is relatively small. This helps to reduce the complexity of the resulting spectrum. Second, the sensitivity of the methyl-TROSY is superior since there are 3 protons for each methyl group versus 1 for each amide group. Third, experiments can be run in D2O buffers, thus simplifying solvent suppression and removing the need for reprotonating amides. The methyl-TROSY relies on the rapid methyl 3-fold axis rotation. The slowly relaxing proton transitions are converted to slowly relaxing multiple quantum coherences and then back to slowly relaxing proton transitions for detection (Figure 1.17). This coherence transfer pathway is in parallel with the fast relaxing components; thus these pathways remain separated and only the slowly relaxing signals can be detected. This phenomenon is similar to the TROSY effect of the amide spin system described previously. Interestingly, the standard  1  H/13C-HMQC  (heteronuclear multiple quantum coherence) pulse sequence, with decoupling, produces the methyl-TROSY effect (Tugarinov, Hwang et al. 2003).  30  31  Figure 1.17: Background of methyl-TROSY effect. (A) Energy-level diagram for AX3 spin system of a rapidly rotating 13CH3 methyl group indicating the fast and slowly relaxing transitions. The first spin state (orange) of each function corresponds to 13C and the remaining (green) corresponds to 1H of methyl group. The slow and fast coherence transfer pathways remain separated in the 1H/13C-HMQC pulse scheme to produce methyl-TROSY effects. (B) Simplified 1H/13C-HMQC pulse scheme illustrates the coherence transfer of the slowly relaxing 1H transition (blue) to the slowly relaxing 13C transition (green) and back to the slowly relaxing 1H transition (blue) for detection. WALTZ16 is used to decouple the 13C signal. The figure is modified from (Kay 2011). One drawback of the methyl-TROSY approach is that the local fields around the methyl group still cause dipole-dipole and CSA cross relaxation effects. Fortunately, these unwanted relaxation effects can be prevented by using a fully deuterated sample with otherwise specifically labeled isoleucine and/or leucine/valine methyl groups ( 13CH3). Spectral assignments of the methyls are typically obtained through mutagenesis or, when amide assignments are available, through scalar correlation experiments linking amide and methyl groups.  1.6.2.1  Selective methyl labeling When using the methyl-TROSY method, the NMR active  13  CH3 methyl of alanine (β1-  methyl) (Ayala, Sounier et al. 2009), isoleucine (δ1-methyl or γ2-methyl) (Tugarinov and Kay 2004; Ayala, Hamelin et al. 2012), leucine (one of the δ-methyls) (Tugarinov and Kay 2004), methionine (ε1-methyl) (Gelis, Bonvin et al. 2007), and valine (one of the γ-methyls) can be selectively incorporated into an otherwise deuterated protein by using specific precursors (Tugarinov and Kay 2004) (Figure 1.18). Currently, the precursors used to label the methyl groups of isoleucine (δ1), methionine (ε1), leucine (δ), and valine (γ) are commercially available, whereas those for alanine (β1) and isoleucine (γ2) require chemical synthesis with established protocols.  32  Figure 1.18: Examples of precursors used in different selective methyl group labeling strategies. Precursors to 13CH3 label the terminal methyl groups of (A & D) leucine, valine, (B, C, & E) isoleucine, and (F) methionine in a deuterated background with the remainder of the side-chain containing either 12C (blue) or 13C (red) nuclei. The precursors of (C-E) can be used for spectral assignment by linearized scalar coupling NMR experiment (Tugarinov and Kay 2003).  33  1.6.2.2  Studies of structural dynamics and conformational exchange using methylTROSY approaches  1.6.2.2.1  Magnetization exchange spectroscopy  If a protein has two states (A and B) that interconvert slowly, then nuclei sensitive to the conformations of the protein will typically yield NMR signals at two distinct chemical shifts (differing by where kex < ). However, if kex is ~ 0.1 - 10 s-1, then exchange of longitudinal magnetization between the two states may be measureable using a variety of "exchange spectroscopy" approaches (Figure 1.19) (Farrow, Zhang et al. 1994; Sprangers, Velyvis et al. 2007). For example, if a single methyl group has different chemical shifts in two slowly interconverting conformational states of a protein, two distinct signals will be observed in a methyl-TROSY spectrum (Figure 1.19B). However, if interconversion occurs during a mixing period of length Tmix, some of the 13C signal of state A is transferred to the 1H signal of state B, and vice versa (Figure 1.19A). Thus cross peaks corresponding to A to B and B to A transfers will be observed in the resulting spectrum (Figure 1.19C). This not only proves that conformational exchange occurs, but can also provide thermodynamic (equilibrium constant) and kinetic (rate constants) information, as well as the assignment of the signals from the nuclei in each state.  34  Figure 1.19: Schematic spectra showing the NMR measurement of conformational exchange. (A) Schematic diagram illustrating magnetization transfer from the 13C signals of state A (blue) or B (red) to the 1H signals of state B or A during the mixing time delay (Tmix) for a single methyl group. Schematic 1H/13C-HMQC exchange spectra recorded (B) without a transfer delay (Tmix = 0), and (C) with a delay (Tmix > 0). In spectrum (B), only "auto" peaks A and B, corresponding to the 1H/13C-signals of a methyl group in each state, are seen. In spectrum (C), the “auto” peaks have decreased in intensity due to relaxation, whereas "exchange" peaks (purple) appear due to transfer of magnetization between the two states. This figure is adapted from (Sprangers, Velyvis et al. 2007).  1.6.2.2.2  Relaxation dispersion  Conformational exchange on the μsec-msec time-scale can be investigated by  13  C  relaxation dispersion measurement on the methyl-labeled sample (Korzhnev, Kloiber et al. 2004). In this method, the contribution of exchange towards relaxation of  13  C-1H multiple  quantum coherence is measured during a constant time spin-echo delay (T) with variable numbers of CPMG refocusing pulses. The concept is illustrated in Figure 1.20. If there is chemical/conformational exchange broadening, the signal will increase in intensity with increasing frequency of the CMPG refocusing pulses (ν CPMG). If there is no exchange broadening, the signal will remain the same regardless of the number of the pulses. The effective chemical exchange rate (Rex) can be calculated from the peak line-width (R2,eff) as a function of refocusing pulses (νCPMG) by equation 1-1 (Korzhnev, Kloiber et al. 2004; Sprangers, Velyvis et al. 2007):  35  (1-1)  I(vCMPG) is the intensity of a cross-peak recorded with CPMG pulses over a constant time delay T, and I(0) is that of a control spectrum recorded without the delay. The effective field νCPMG = 1/(4τCPMG) with 2τCPMG being the time between the centers of successive  13  C refocusing pulses.  In favorable cases, fitting of the relaxation dispersion profile (R 2,eff versus νCPMG) can yield the exchange rate constant, populations of the interconverting states, and the chemical shift differences between the states (Kleckner and Foster 2012).  36  37  Figure 1.20: Schematic diagram illustrating the measurement of conformational dynamics by relaxation dispersion experiment. (A) Two conformational states are represented by a cat (slow runner) and a leopard (fast runner). In the absence of interconversion, these two populations give rise to two separate peaks in an NMR spectrum. (B) When these two conformational states undergo chemical exchange at a rate comparable to their speed (chemical) difference, the two populations will be mixed and give rise to one broad peak. (C) In the absence of exchange, the position of either state can be refocused by a single CPMG 180° pulse in the middle of a constant delay time (T). This is a spin-echo. (D) However, when a single CPMG pulse is applied and if two states interconvert during T (i.e. the cat and leopard change identities), they will not refocus completely. This results in a broad signal. (E) On the other hand, when multiple CPMG pulses are applied quickly, the two interconverting states cannot diverge as far and a sharper signal is detected. (F) By applying this concept, a dispersion curve is obtained if R2,eff values can be plotted as a function of CPMG pulses (νCPMG) at a constant time delay and thus the Rex rate, which is the function of kex, Δω, and populations, for the conformational dynamics is determined. This figure is adapted from (Sprangers, Velyvis et al. 2007).  1.6.3  STD-NMR spectroscopy Saturation transfer difference (STD) NMR spectroscopy is an approach for identifying  the binding determinants (“epitopes” by analogy to antigen-antibody interactions) of low affinity ligands with proteins (Meyer and Peters 2003). Thus, protein-carbohydrate interactions are widely studied by this approach as their Kd values are typically within the appropriate μM to mM range (Mayer and Meyer 2001; Blume, Berger et al. 2008; Brecker, Schwarz et al. 2008; Blume, Fitzen et al. 2009; Liu, Meng et al. 2009). In this technique, the protein nuclei are selectively saturated by spin diffusion and the saturation is transferred through intermolecular 1H-1H cross relaxation to a bound ligand (Haselhorst, Lamerz et al. 2009) (Figure 1.21). The level of signal saturation depends on the relative distances between the epitopes and the protein, as well as the rates of complex formation and dissociation. The closer a bound ligand proton is to the protein and/or the slower the rate of dissociation, the more its signal will be attenuated due to saturation transfer. Thus, features of the binding mode of the ligand can be approximated. In practice, “offresonance” (i.e. reference without protein saturation) and “on-resonance” (with protein saturation) spectra are acquired for the same protein and ligand sample. In each case, only the signal from the “free” ligand is observed, whereas differences between the two spectra reflect saturation transfer to the ligand while protein-bound. The advantages of STD-NMR spectroscopy are that it requires relatively small amounts of protein, is not limited by the size of the protein, 38  and relies on observation of the relatively simple one-dimensional spectrum of the ligand. However, one should note that high-affinity ligands with long life-time in the bound state (slowexchange) cannot be detected by STD-NMR spectroscopy because this method relies on the “memory” of saturation being transferred from the bound to the free state of the ligand.  Figure 1.21: Schematic diagram showing the principle of STD-NMR spectroscopy. Ligand 1, but not ligand 2, is able to bind to the protein. STD-NMR signals are observed in the spectrum of the free ligand if there is transfer of saturation when it is in a bound state. This diagram is modified from (Haselhorst, Lamerz et al. 2009).  39  1.7  OVERVIEW OF THE THESIS CstII and LgtC were chosen as model enzymes for investigating and comparing the  enzymatic mechanisms of inverting and retaining GTs. Both belong to the same GT-A family, yet CstII exploits an inverting mechanism, whereas LgtC utilizes a retaining mechanism. The crystal structures of both CstII and LgtC in substrate/inhibitor-bound states have been solved (Persson, Ly et al. 2001; Chiu, Watts et al. 2004) and numerous kinetic studies have been carried to investigate their catalytic properties. Chapter 2 describes the study of CstII. The crystal structures of both monofunctional CstI and bifunctional CstII with sugar donor analog CMP-3FNeu5Ac provided important insights into the mode of substrate binding, as well as testable hypotheses regarding the mechanism of sialyl transfer (Chiu, Watts et al. 2004; Chiu, Lairson et al. 2007). Recently, the crystal structure of CstII complexed with CMP and the terminal trisaccharide of its natural sugar acceptor (Neu5Acα-2,3-Gal-β-1,3-GalNAc) has also been obtained (Lee, Lairson et al. 2011). Although the initial and final steps of α-2,3-sialyltransfer and the initial step of α-2,8-sialyltransfer reactions by CstII have been illustrated crystallographically, the substrate binding mode and the motion of the flexible lid domain of this ST during catalysis remain unclear. These questions have been addressed in this project by using STD-NMR spectroscopy. Furthermore, His188 of CstII was proposed as the catalytic base in the mechanism. Mutagenesis and chemical rescue experiments confirmed the importance of this residue in catalysis (Chiu, Watts et al. 2004; Chan, Lairson et al. 2009). Complementary NMR-monitored pH-titration studies were performed to demonstrate that His188 has the same pKa as the general base. This work first required the monomerization of the tetrameric enzyme in order to reduce its size to a range amenable for NMR spectroscopic characterization. Most of the results described in Chapter 2 have been published (Chan, Lairson et al. 2009). Chapters 3 and 4 present an analysis of LgtC. In the first of these two chapters, approaches to assign the NMR spectra of this 34-kDa protein are summarized. Advanced NMR techniques, such as 1H/15N-TROSY-HSQC and methyl-TROSY, selectively isotope labeled proteins, mutagenesis, and extensive optimization of sample conditions were required for this research (Tugarinov and Kay 2003; Sprangers and Kay 2007; Sprangers and Kay 2007; Sprangers, Li et al. 2008). Chapter 4 describes the investigation of substrate binding and dynamics of LgtC by NMR methods. The key results of these studies are demonstration of the 40  multiple conformational states in equilibrium and the changes in the associated dynamics upon substrate binding. A discussion of the possible linkage between the enzymatic mechanism (S Ni) of LgtC and its multiple conformational states is provided. The work is currently being prepared for publication.  41  Chapter 2 Investigation  of  the  enzymatic  mechanism  and  structural dynamics of the bifunctional sialyltransferase CstII from Campylobacter jejuni  Cell surface glycans are often terminated by sialic acid, which is incorporated onto sugar acceptors by sialyltransferases. The crystal structure of the GT family 42 Campylobacter jejuni α-2,3/2,8-sialyltransferase (CstII) provided key insights into the sialyl-transfer mechanism, including tentative identification of His188 as the catalytic base. In support of this hypothesis, the CstII-H188A mutant is able to catalyze sialyl transfer from CMP-Neu5Ac to added anions such as azide and formate, but not to its natural sugar acceptor lactose. Complementing this work, NMR spectroscopy was used to investigate the structure and dynamics of CstII and to measure the site-specific pKa value of His188 for comparison with the apparent pKa governing the pH-dependent kcat/Km of enzyme. By systematically introducing point mutations at the subunit interfaces, two active monomeric variants, CstII-F121D and CstII-Y125Q, were obtained and characterized. In contrast to the wild-type tetramer, the monomeric CstII variants yielded good quality 1H/15N-HSQC and methyl-TROSY NMR spectra. However, the absence of signals from approximately one half of the amides in the 1H/15N-HSQC spectra of both monomeric forms suggests that the enzyme undergoes substantial conformational exchange on a msec-sec timescale. The histidine pKa values of CstII-F121D in its apo form were measured by monitoring the pH-dependent chemical shifts of [ 13C1]-histidine, biosynthetically incorporated into the otherwise uniformly deuterated protein. Consistent with its proposed catalytic role, the sitespecific pKa value ~ 6.6 of His188 matches the apparent pKa value ~ 6.5 governing the pHdependence of kcat/Km for CstII towards CMP-Neu5Ac in the presence of saturating acceptor substrate. At the same time, STD-NMR spectroscopy was employed to investigate the sugar donor and acceptor binding of CstII.  42  2.1  INTRODUCTION  2.1.1 Catalytic His188 of CstII Prominent among the active site residues of CstI and CstII is a histidine (His202 and His188, respectively) proposed to serve as the general base that abstracts the proton from the nucleophilic hydroxyl group of the sugar acceptor, thereby facilitating attack on CMP-Neu5Ac (Figure 1.10). This is a somewhat unusual residue to serve as the base for glycosyl transfer reactions. Although mutational studies have confirmed their overall importance for catalysis, no definitive evidence for these histidine residues serving as a base has been provided. Thus, two avenues were followed within this research to test this proposal. First, "chemical rescue" experiments, carried out by Dr. L. L. Lairson, verified that an otherwise inactive H188A mutant of CstII can still catalyze the reaction of CMP-Neu5Ac with anions including azide and formate to yield a sialyl azide product of inverting anomeric configuration. Second, NMR spectroscopy was used to measure the site-specific pKa value of His188 for comparison with that determined by kinetic analysis. Preliminary NMR studies were foiled by the high molecular weight (~128 kDa) of the tetrameric enzyme (Chiu C., PhD thesis, Univ. of British Columbia, 2007). Fortunately, the active sites of CstI and CstII are contained entirely within a monomer subunit. By introducing point mutations at the subunit interfaces, two active monomeric variants of CstII were obtained, thereby enabling these NMR measurements. The X-ray crystal structure of the monomeric CstII-Y125Q, solved by Dr. H. J. Lee, showed that monomerization resulted in no significant conformational changes relative to wild-type CstII. In support of its proposed role as a catalytic general base, the pKa value of ~6.6 determined by NMR-monitored titrations of monomeric CstII indeed closely matches that of pKa value ~6.5 governing its pH-dependent activity (kcat/Km).  2.1.2 Structural dynamics of CstII The electron density of the lid domain is missing in structures of CstII complexed with CMP (PDB code: 1RO8) and with CMP and the trisaccharide Neu5Ac-α-2,3-Gal-β-1,3-GalNAc (PDB code: 2X61) and detected crystallographically only after binding of CMP-3FNeuAc (PDB code: 1RO7). Closure of the flexible lid motif, which contains the catalytic base His188, appears required for sialyl-transfer and may serve to protect the active site of the enzyme from bulk solvent, thereby minimizing undesirable reactions such as hydrolysis of the donor substrate. 43  However, beyond the presence or absence of electron density from the lid domain in crystallized CstII, the time-scale and amplitudes of its motions have not been characterized. In order to investigate the conformational dynamics on a msec-μsec time-scale of the enzyme, methyl relaxation dispersion measurements were carried out.  2.1.3 Substrate binding of CstII Although the initial and final steps of α-2,3-sialyl-transfer and the initial step of α-2,8sialyl-transfer reaction by CstII have been illustrated crystallographically, the substrate binding mode and the motion of the flexible lid domain of this sialyltransferase during catalysis remain unclear. Attempts to address these questions are made in this project by using STD-NMR spectroscopy.  44  2.2  METHODS  2.2.1 Site-directed mutagenesis The previously described gene encoding "wild-type" CstII, with the mutation I53S and a deletion of the C-terminal 32 residue membrane association sequence to improve solubility, was cloned into the pET28a vector for expression as a His6-tagged construct (Chiu, Watts et al. 2004; Chiu, Lairson et al. 2007). Additional mutations were introduced sequentially using the QuikChange site-directed mutagenesis kit (Stratagene).  2.2.2 Protein expression and purification The His6-tagged proteins were expressed in E. coli BL21 (DE3) cells according to previously published methods (Chiu, Watts et al. 2004; Chiu, Lairson et al. 2007). The cells were grown at 37 °C to an OD600 of 0.8, and then induced with 1 mM IPTG. After further growth at 30 °C for 16 hours, the cells were harvested by centrifugation and lysed by sonication in the presence of 50 mg/L lysozyme (Sigma). The cell debris was removed by centrifuging at 15,000 rpm and CstII was isolated from the supernatant using a HisTrap HP column (GE Healthcare). After thrombin digestion, HisTrap HP and HiTrap Benzamidine FF columns (GE Healthcare) were used to remove the His6-tag, uncleaved His6-tagged proteins, and the protease. The Nterminal tripeptide Gly-Ser-His, or in the case of samples for histidine titrations, Gly-Ser-Gly, remained after cleavage of the tag. containing 1 g/L and Val-[13CH3,  15  N-labeled proteins were expressed in M9 minimal media  NH4Cl. Proteins selectively labeled with Ile δ1-[1H/13C], Leu-[13CH3,  15  12  CD3]  12  CD3] in an otherwise deuterated background were expressed in M9 media  containing 70 mg/L of 4-[13C,1H]-3,3-2H-α-ketobutyrate, 120 mg/L of 2-keto-3-methyl-d3-3-d14-13C-butyrate (α-ketoisovalerate deuterated at the β-position and with one of the two methyl groups  13  CH3 and the other  12  CD3), 1 g/L  15  NH4Cl and 3 g/L D-[2H]-glucose in 99% D2O  (Cambridge Isotope Laboratories) (Tugarinov and Kay 2004; Sprangers and Kay 2007; Sprangers and Kay 2007). Proteins selectively labeled with [13Cɛ1]-histidine in an otherwise deuterated background were expressed in the histidine auxotrophic strain BL21 (DE3) hisG, grown in M9 media containing 50 mg/L [ 13Cɛ1]-L-histidine (Icon Isotopes), 1 g/L NH4Cl and 10 g/L D-glucose in 99% D2O (Waugh 1996; Venter, Ashcroft et al. 2002). Unless stated otherwise, the purified proteins were exchanged into 20 mM Tris, 150 mM NaCl, 5 mM DTT, pH 7.5 in H2O or D2O, using an Amicon Ultra-15 Centrifugal Filter Device. Protein concentrations were 45  determined by UV absorbance using predicted 280 values of 34,270 M-1cm-1 for CstII-Y125Q and 35,760 M-1cm-1 for both wild-type CstII and CstII-F121D (Wilkins, Gasteiger et al. 1999).  2.2.3 Activity assays The activities of unpurified mutants in cell lysates (induced with 1 mM IPTG and expressed overnight at 30 °C) were determined qualitatively by a TLC fluorescent assay using CMP-Neu5Ac and bodipy-lactose (Aharoni, Thieme et al. 2006). Steady-state kinetic parameters for purified CstII and CstII-Y125Q were determined by a continuous coupled enzyme assay (Gosselin, Alhussaini et al. 1994; Lee, Lairson et al. 2011).  2.2.4 pH-Dependence of kcat/Km The pH dependence of kcat/Km for CstII was determined using the method of substrate depletion at a low concentration ([S] << Km) of donor substrate (50 μM CMP-Neu5Ac) and a saturating concentration of the acceptor substrate (15 mM 3’-sialyl lactose). Rates of reaction were measured using a continuous coupled assay (Gosselin, Alhussaini et al. 1994). Reactions were performed using 20 mM buffer with coupling components at 37 °C and initiated by the addition of 23 μg/mL purified enzyme (Gosselin, Alhussaini et al. 1994). Assays were carried out at pH values ranging from 5.5 to 9.0 using 20 mM buffers as follows: sodium citrate for pH 5.5 to 6.0, MES for pH 6 to 6.5, sodium phosphate for pH 6.0 to 7.5, HEPES for pH 7.0 to 8.0, and Tris for pH 7.5 to 9.0. Overlapping data points indicated no significant dependence of activity on the specific buffer identity. A linear relationship between observed rates and CstII concentrations was observed at all pH values. Release of CMP, coupled to the oxidation of NADH (λ = 340 nm, ε = 6.22 mM-1 cm-1), was monitored continuously at 340 nm for a minimum of 10 minutes and the data were fit to the first order rate equation using GraFit 5.0 (Erithacus Software). The resulting rate constant, adjusted for enzyme concentration, gives the kcat/Km for a given pH value. Data for kcat/Km dependence on pH were fit to a single ionization equilibrium curve using GraFit 5.0 to determine an apparent pKa.  2.2.5 Anion rescue studies Chemical rescue of the CstII-H188A mutant was monitored using the continuous coupled assay (Gosselin, Alhussaini et al. 1994). Initial rates for the CstII-H188A (16 μM) and wild-type 46  enzyme (0.8 μM in monomer subunits) were determined at 0.5 mM CMP-Neu5Ac with varying concentrations of sodium formate (0 – 1 M) or sodium azide (0 – 0.25 M). Formation of an linked sialyl azide derivative was confirmed by negative ion mode ESI-MS and by the observed ability of Clostridium perfringens sialidase to convert the putative species to sialic acid, as determined by TLC analysis (4: 2: 1: 0.1 EtOAc: MeOH: H2O: AcOH).  2.2.6 Analytical size exclusion chromatography and static light scattering The oligomerization states of CstII mutants were determined using a Sephacryl S-200 high resolution XK16/60 gel filtration column (GE Healthcare), equilibrated with buffer (5 mM PBS, pH 7.4) connected to the miniDAWN light-scattering equipment coupled to an interferometric refractometer (Wyatt Technology) at 22 °C. Data analysis was performed in real time using ASTRA software (Wyatt Technology).  2.2.7 CD spectroscopy CD spectra of ~10 μM protein in a 0.2 cm path length cuvette were recorded on a JascoJ810 spectropolarimeter. Four scans at 50 nm/min were averaged, followed by subtraction of a buffer blank spectrum. Thermal denaturation measurements were monitored at 220 nm with heating at a rate of 1 °C/min. Mid-point unfolding temperatures (Tm) were determined by fitting to a standard two-state model (Pace 1990).  2.2.8 X-ray crystallographic analysis of CstII-Y125Q Crystals of CstII-Y125Q mutant (6.4 mg/mL) were grown by the hanging-drop vapour diffusion technique at 21 °C with 10 mM CMP-3FNeu5Ac in a mother liquor containing 75 mM tribasic ammonium citrate, pH 8.0, and 14% (w/v) PEG-3350. X-ray diffraction data were collected with 15% ethylene glycol as a cryoprotectant at 100 K under a nitrogen stream using an in-house Cu-Kα rotating anode X-ray generator coupled to a Mar345 detector. The crystal belongs to the space group I4 with unit cell dimensions a=116.63, b=116.63, c=45.39 Å, and contains one molecule in the asymmetric unit. Collected data were processed by MOSFLM (Leslie 2006) and the CCP4 suite of programs (Collaborative Computational Project 1994). All statistics for data collection and refinement are summarized in Table 2.3.  47  The structure of CstII-Y125Q was solved by molecular replacement with PHASER (McCoy 2007) using the monomer of wild-type CstII (PDB accession code: 1RO7) as the starting model. Subsequent model building was performed manually by COOT (Emsley and Cowtan 2004) and refinements were carried out by REFMAC (Murshudov, Vagin et al. 1997), excluding 5% of reflections for the R-free calculation. The quality of the final model was validated using MolProbity (Davis, Leaver-Fay et al. 2007). The CMP-3FNeu5Ac ligand model and its refinement restraints were generated from the PRODRG server (Schuttelkopf and van Aalten 2004). All structural figures were generated using PyMOL (DeLano and Lam 2005). The co-ordinates of CstII-Y125Q have been deposited in the Protein Data Bank under the accession code 2WQQ.  2.2.9 NMR spectroscopy NMR spectra were acquired at 25 °C on a Varian Inova 600 MHz spectrometer equipped with a cryogenic 1H/2H/13C/15N probe. Spectra were processed using NMRPipe (Delaglio, Grzesiek et al. 1995) and analyzed with SPARKY 3 (Goddard 1999). One-bond sensitivityenhanced 1H/15N-HSQC spectra (Kay, Keifer et al. 1992) were recorded for CstII-F121D, and CstII-Y125Q.  1  15  N-labeled CstII,  H/13C-methyl-TROSY and relaxation dispersion methyl-  TROSY experiments (Tugarinov and Kay 2003; Sprangers and Kay 2007; Sprangers and Kay 2007; Sprangers, Li et al. 2008) were recorded with [1H/13C]-methyl labeled deuterated CstIIY125Q (0.1 mM) in D2O sample buffer with or without 5 mM CMP-3FNeu5Ac. One-bond sensitivity-enhanced 1H/13C-HSQC spectra with a CPMG pulse train (Mulder, Spronk et al. 1996) were recorded as a function of pH* with [13Cɛ1]-histidine labeled CstII-F121D and CstIIF121D-H188A in D2O sample buffer. The sample pH* values were adjusted by addition of L amounts of 0.05 M DCl and 0.05 M NaOD and measured using a NMR Tube Micro Probe Electrode (IQ Scientific Instruments). pKa* values were determined by simultaneously fitting the resulting pH-dependent 1Hɛ1 and 13Cɛ1 chemical shifts to standard equations for the ionization of a single group using Matlab (MathWorks). The standard deviations of fit pKa values were determined with a Monte Carlo approach using estimated errors of 0.05 in pH*, 0.04 ppm in 1H, and 0.2 ppm in 13C. The observed pH meter readings, denoted as pH*, were not corrected as the apparent pKa* values determined in D2O are approximately the same as those determined in H2O  48  due to compensating isotope effects on the acid dissociation equilibrium and the glass electrode (Bundi and Wuthrich 1979).  2.2.10 STD-NMR spectroscopy CstII-F121D (10 μM protein, 20 mM d11-Tris, 150 mM NaCl, 5 mM DTT, pH 7.5 in D2O) was used for STD-NMR spectroscopy (Angulo, Rademacher et al. 2006). Either 1 mM CMP-3FNeu5Ac or 1 mM CMP-3FNeu5Ac and 1 mM lactose were added to the NMR tube containing the protein. STD-NMR spectra were acquired at 25 °C using 600 MHz spectrometer. On-resonance irradiation was performed at -1 ppm and off-resonance at 30 ppm. Irradiation was performed using Waltz16 pulses and 70 ms duration to give a total saturation time of 2.0 s (Angulo, Rademacher et al. 2006). STD NMR spectra were acquired with a total of 2048 scans. Reference spectra were acquired with a total of 512 scans. STD effect was calculated as the ratio of peak intensity in the presence (ISTD) and absence (IRef) of protein saturation using the following equation: (2-1) The STD effects were the normalized by assigning a value of 100% to the ligand epitope showing the largest ASTD value.  49  2.3  RESULTS  2.3.1 Chemical rescue of CstII-H188A activity Chemical rescue is a powerful approach for confirming the locations and identities of residues involved in acid/base or nucleophilic catalysis by glycoside hydrolases. It involves the demonstration of rescue of steady-state activity by mutants modified at the position in question upon the addition of exogenous anions such as azide (MacLeod, Lindhorst et al. 1994; Viladot, de Ramon et al. 1998; Zechel and Withers 2000) (Figure 2.1). This approach has also been applied to probe nucleophiles in retaining GT's (Monegal and Planas 2006), but not previously to acid/base residues in this class of enzymes.  50  Figure 2.1: Chemical rescue of CstII-H188A by exogenous anions. (A) Schematic diagram of the chemical rescue experiment. Rates were determined using 16 µM (monomer subunits) CstII-H188A mutant or 0.8 µM wild-type enzyme, 500 µM CMP-Neu5Ac, 160 mM lactose, and standard CstII assay conditions. Increasing concentrations of formate and azide rescue the activity of CstII-H188A (B, D) without substantially influencing wild-type CstII (C, E). Mutation of His188 to alanine severely impairs CstII activity (Chiu, Watts et al. 2004). Addition of either formate or azide to the CstII-H188A mutant resulted in significant concentration-dependent increases in its ability to utilize CMP-Neu5Ac as a substrate (Figure 2.1B & D). In contrast, wild-type CstII is relatively insensitive to these exogenous anions (Figure 2.1C & E). This behaviour is completely consistent with anion rescue occurring exclusively with  51  the mutant. Importantly, rates in the presence of lactose were virtually identical to those in the absence of this acceptor at all concentrations of exogenous nucleophiles. This indicates that the observed chemical rescue is not due to transferase activity involving lactose, but rather arises through a diversionary pathway in which a sialic acid derivative is formed. Confirmation of this pathway was obtained by TLC and ESI-MS analyses of reaction mixtures, demonstrating the formation of a sialyl azide derivative by the H188A reaction mixture. Thus, removal of the sidechain of His188 allows rescue by azide, thereby indicating that His188 is located close to the anomeric carbon of CMP-Neu5Ac and the 3’-OH group of lactose in the CstII active site. The distance and proximity are consistent with the proposed catalytic role of His188 (Figure 2.3).  2.3.2 pH-Dependent activity of CstII In order to assess the pKa values of catalytic residues involved in essential proton transfer steps, the kcat/Km value for CstII-catalyzed reaction of CMP-Neu5Ac in the presence of saturating concentrations of sialyl-lactose was measured as a function of sample pH value. As shown in Figure 2.2, the activity of CstII increases with pH, reaching a maximal plateau value near pH 8. The data fit a single ionization event with an apparent pKa value of 6.5 ± 0.1. This kinetic pKa likely reflects the deprotonation of the postulated general base, His188. To test this hypothesis, we used NMR spectroscopy to independently measure the pKa value of His188 in a monomerized CstII variant.  52  Figure 2.2: The pH dependence of kcat/Km for utilization of CMP-Neu5Ac by CstII. The pH dependence of kcat/Km for utilization of CMP-Neu5Ac by CstII, saturated with sialyllactose, at 37 °C. The apparent pKa value (and standard error) of 6.5 ± 0.1 was determined by best fit of the data to a single ionization equilibrium. 2.3.3 Monomerization of CstII The active site of tetrameric wild-type CstII is contained entirely within a monomer unit, and thus it was hypothesized that mutations at the subunit interface would not dramatically affect the activity of the enzyme. Based on inter-monomer contacts and accessible surface areas, determined with the PROTORP server (Reynolds, Damerell et al. 2009), key interfacial residues were identified as Tyr60’, Lys63’, Asp97’, Tyr98’, Pro100’, Asp101’, and Lys253’ on one subunit, and Ala84, His85, Lys113, Ala117, Phe121, Ile124, Tyr125, and Phe126 on the other (Figure 2.3B). As summarized in Table 2.1, several of these were mutated individually or in various combinations to charged or polar residues. Fortunately, many of the resulting proteins were active based on a qualitative TLC assay. The remainder, particularly those with multiple mutations, either expressed poorly in E. coli or were inactive.  53  Figure 2.3: Superimposed structures of monomer of wild-type CstII and CstII-Y125Q complexed with the sugar donor analog CMP-3FNeu5Ac (A) Superimposed structures of wild-type CstII (blue, red) and CstII-Y125Q (green) complexed with the sugar donor analog CMP-3FNeu5Ac. Tyr125 in wild-type CstII and Gln125 in CstIIY125Q are shown as stick models. Carbon atoms are in yellow and orange for CMP-3FNeu5Ac bound to wild-type CstII and CstII-Y125Q mutant, respectively. The lid domain (residues 155188) of the wild-type CstII is identified in red. All non-carbon atoms are colored according to type (N blue, O red, F, cyan, P orange). (B) Structural alignment of the interface between two monomer subunits of the wild-type CstII and CstII-Y125Q. The core interfacial residues are shown as stick models. Residues (primed) from adjacent monomers of CstII-Y125Q are colored in blue and green for carbon atoms, respectively. Corresponding residues of wild-type CstII are displayed in yellow for carbon atoms, with the exception of Gln125 in cyan. Non-carbons are identified according to atom type (N blue, O red).  54  Table 2.1: Screening of CstII mutants for oligomerization and activity Mutation  Expression a  Oligomerization State b  Activity c  Tm d (°C)  Wild-type  Yes  Tetramer  Yes  62  K63D  Yes  Tetramer  Yes  -  D97A  Yes  Tetramer  Yes  -  F121D  Yes  Monomer  Yes  53  Y125Q  Yes  Monomer  Yes  49  K253Q  Yes  Monomer/Dimer  Yes  -  F121D/Y125Q  Yes  Monomer  No  -  K63D/F121D  No  -  -  -  D97A/Y125Q  Yes  Monomer  No  -  Y125Q/K253Q  Yes  Monomer  No  -  K63D/D97A/F121D  No  -  -  -  K63D/F121D/K253Q  No  -  -  -  D97A/Y125Q/K253Q  Yes  Monomer  No  -  K63D/F121D/Y125Q  No  -  -  -  K63D/D97A/F121D/Y125Q  No  -  -  -  F121D/I124D/Y125D/F126D  Yes  Monomer  No  -  K63D/D97A/F121D/Y125Q/K253Q  No  -  -  -  a  Expression measured by SDS-PAGE of whole cell lysates after 16 hrs of induction at 30 °C.  b  Oligomerization state determined by analytical gel filtration chromatography (22 °C) in pH 7.5  sample buffer. c  Activity of unpurified cell lysates measured qualitatively by a TLC fluorescence assay.  d  Tm values determined from thermal unfolding curves monitored by CD spectroscopy at λ = 220  nm (20 mM Tris, 150 mM NaCl, 5 mM DTT, pH 7.5). CstII does not refold reversibly under these conditions.  55  2.3.3.1 Oligomerization states of the mutants The oligomerization states of the active mutants were determined by analytical gel filtration chromatography. CstII-K63D and CstII-D97A remained tetrameric, whereas CstIIK253Q appeared as a mixture of dimeric and monomeric forms. Importantly, CstII-F121D and CstII-Y125Q had longer retention times, thus were identified as monomers. Confirming this conclusion, CstII-Y125Q was found by static light scattering to have a molecular mass of ~36 kDa, whereas that of wild-type CstII was ~140 kDa (Figure 2.4).  Figure 2.4: Static light scattering of wild-type CstII (left) and CstII-Y125Q (right). The calculated molecular mass of CstII-Y125Q (~36 kDa) from the molar mass distribution plots on the sample elution profiles indicates that it is monomeric, whereas wild-type CstII (~140 kDa) is a tetramer. Both species (100 μL of 50 μM protein in 5 mM PBS buffer) yielded monodisperse elution profiles from a gel filtration column connected to the miniDAWN light scattering detector (Wyatt Technology) at 25°C.  2.3.3.2 Kinetic analysis of the monomeric CstII-Y125Q The kinetic properties of wild-type CstII and CstII-Y125Q were compared using a coupled assay (Gosselin, Alhussaini et al. 1994) (Table 2.2). The kcat and Km values for CMPNeu5Ac in the presence of saturating lactose were lower for CstII-Y125Q than for the wild-type enzyme, with the resultant kcat/Km being reduced by ~3-fold. These are relatively small changes, indicating that any inter-subunit interactions are not critical for the activity of the enzyme. Further evidence that the monomer provides a good model of the wild-type tetramer was sought through X-ray crystallographic studies.  56  Table 2.2: Steady state kinetic parameters for wild-type CstII and monomeric CstII-Y125Q Km (μM)  kcat (min-1)  kcat / Km (μM-1 min-1)  WT CstII  440  60  0.14  CstII-Y125Q  170  8  0.05  Data for sugar donor CMP-Neu5Ac in the presence of 160 mM acceptor lactose at pH 7.5 and 37 °C.  2.3.4 Structural characterization of monomeric CstII CD spectroscopy and X-ray crystallography were used to characterize CstII-F121D and CstII-Y125Q. The CD spectra of wild-type CstII, CstII-F121D, and CstII-Y125Q revealed that the secondary structures of the mutants were not significantly altered by the mutations (data not shown). However, the midpoint unfolding temperatures (Tm) of both mutants were substantially lower than that of the wild-type protein, indicating a loss of stability, as would be expected for a monomer compared to its native tetrameric form. As summarized in Table 2.1, CstII-F121D was both slightly more stable and, in practical terms, better behaved (i.e. more soluble and less prone to aggregation during NMR measurements) than CstII-Y125Q. This is likely a result of the additional surface charge introduced by this mutation. CstII-Y125Q crystallized with one molecule in the asymmetric unit. However, the application of four-fold crystallographic symmetry reconstitutes a tetramer very similar to that observed for wild-type CstII (0.68 Å r.m.s. deviation for 699 Cα atoms). Although monomeric in solution as confirmed by analytical gel filtration chromatography and static light scattering experiments, this suggests that the CstII-Y125Q mutant still has the ability to form a tetramer in the crystal lattice under specific crystallization conditions. In comparison to wild-type CstII, there was no significant change in secondary, tertiary, or crystallographically-generated quaternary structure due to the mutation (Figure 2.3A). Indeed, even the α-helix in which the residue 125 is located was not altered in any appreciable way. Interestingly, although the sugar donor analog CMP-3FNeu5Ac is observed in the active site of CstII-Y125Q, the electron density of part of the lid domain (residues 157-161 and 180-186) was missing. This is indicative of conformational disorder within a portion of the CstII active site.  57  Table 2.3: Data collection and refinement of CstII-Y125Q in complex with CMP-3FNeu5Ac Data collection Space group  I4  a (Å)  116.63  b (Å)  116.63  c (Å)  45.39  α, β, γ (°)  90, 90, 90  Wavelength (Å)  1.5418  Resolution (Å)a  34–2.25 (2.37–2.25)  Rsym (%)b  5.9 (30.7)  I/σ(I)  18.4 (3.8)  Completeness (%)  100 (100)  Unique reflections  14,714 (2,137)  Redundancy  4.1 (4.0)  Refinement statistics Average B-factors (Å2) protein  20.6  substrate  48.4  water  25.3  Ramachandran statistics favored regions (%)  95.8  allowed regions (%)  4.2  Rwork (%)c  17.7  Rfree (%)c  21.9  r.m.s.d bonds (Å)  0.013  r.m.s. angles (°)  1.366  a  Values in parentheses represent the highest resolution shell.  b  Rsym = ∑|I(hkl) – <I>| / ∑I(hkl), where I(hkl) is the measured intensity of a given reflection and <I>  is the average intensity of all symmetry equivalent measurements.  58  c  Rwork = ∑|Fo – Fc| / ∑ Fo, where Fo and Fc are observed and calculated structure factors,  respectively. 5% of total reflections were excluded from the refinement to calculate Rfree. d  r.m.s., root mean square  2.3.5 NMR spectroscopic studies of CstII monomeric mutants The 1H/15N-HSQC NMR spectra of wild-type CstII, CstII-F121D, and CstII-Y125Q were obtained (Figure 2.5). Ideally, each of these spectra should contain ~300 peaks arising from the backbone and side-chain amide/indole 1H-15N pairs within the 269 residue protein. However, as expected from its high molecular mass, the spectrum of tetrameric wild-type CstII only showed a small number of broad signals, most of which arise from flexible asparagine and glutamine 15  NH2 groups (Figure 2.5A). In contrast, the spectra of CstII-F121D and CstII-Y125Q each  contained ~150 signals (Figure 2.5B & C). This is consistent with complementary experiments demonstrating that the two mutants are predominantly monomeric. However, the absence of approximately half of the expected signals is suggestive of extensive conformational exchange broadening, either due to msec-sec timescale motions within the monomers or due to an equilibrium with higher order oligomeric forms. Parenthetically, the peaks in the spectrum of CstII-F121D were generally more intense and sharper than those in the spectrum of CstIIY125Q, which might correlate with the higher stability and solubility of the former species. Although CstII-F121D was found to behave better, CstII-Y125Q was generated first and thus used for initial NMR analyses. The 1H/15N-HSQC NMR spectrum of CstII-Y125Q with saturated CMP-3FNeu5Ac was also recorded in the hope that ligand binding would improve its spectra (not shown). Unfortunately, there was no change in the overall number of signals, and small chemical shift perturbations relative to the apo species were observed for only a limited number of resonances.  59  Figure 2.5: 1H/15N-HSQC spectra of 15N-labeled CstII variants. 1  H/15N-HSQC spectra of 15N-labeled (A) tetrameric His6-tagged CstII, (B) monomeric His6tagged CstII-F121D, and (C) monomeric CstII-Y125Q in pH 7.5 at 25 °C. Of ~300 expected 1 N 15 H - N cross peaks, only ~150 are observed in the spectra of the monomeric species, thus suggestive of extensive conformational exchange broadening. Addition of CMP-3FNeu5Ac only causes minor spectral perturbations (not shown).  60  Methyl-TROSY spectra of selectively leucine, valine and isoleucine [ 2H/13C]-methyllabeled CstII-Y125Q with and without substrates were also obtained (Figure 2.6). There are 19 isoleucine, 25 leucine, and 7 valine residues in the protein, thus in the most favorable case, 83 peaks should be observable. In the spectrum of the apo protein, only ~70 peaks were observed. Although this may in part be due to chemical shift degeneracy, the absence of ~15% of the expected signals is again indicative of some conformational exchange broadening. In contrast, ~80 peaks are observed when CstII-Y125Q is bound to CMP-3FNeu5Ac, suggesting that inhibitor binding may increase the overall chemical shift dispersion due to conformational perturbations or may dampen motions of the enzyme leading to the postulated exchange broadening. Addition of acceptor substrate lactose did not lead to any further measurable changes (not shown). Unfortunately, the lack of observable 1H-15N signals from many backbone amides precluded the assignment of these methyl signals by established main-chain directed methods (Tugarinov and Kay 2003).  61  Figure 2.6: Overlaid 1H/13C-methyl-TROSY spectra of uniformly deuterated and [ 1H/13C] selectively methyl labeled CstII-Y125Q in its apo form (red) and in the presence of 5 mM CMP-3FNeu5Ac (blue) in D2O buffer. Only signals from the methyl groups of leucine (δ1 and δ2), valine (γ1 and γ2) and isoleucine (δ1; most upfield in 13C) are detected. In CstII-Y125Q, there are 19 isoleucine, 25 leucine, and 7 valine residues, and thus ideally, 83 peaks should be observed in each spectrum. In the apo spectrum, ~70 peaks are observed, whereas ~80 peaks are observed in the spectrum of CMP3FNeu5Ac bound CstII. Addition of 160 mM lactose did not cause any further spectral changes (not shown).  62  2.3.6 Methyl relaxation dispersion of CstII-Y125Q Methyl relaxation dispersion experiments were used to study the dynamics of CstII. In apo CstII-Y125Q, 18 methyls exhibited measurable conformational exchange broadening, whereas 47 methyls showed this effect when CMP-3FNeu5Ac was present (Figure 2.7). Some of these methyls were common to both forms of the protein, and others were distinct to one form or the other. Unfortunately, without spectral assignments, interpreting these data is difficult. Nevertheless, in the case of the apo enzyme, the observation that many methyl groups showed exchange broadening clearly demonstrates that CstII undergoes significant msec-μsec timescale motions. We speculate that this might reflect the dynamics of the lid domain. In the case of the CMP-3FNeu5Ac bound enzyme, the increased number of exchange-broadened methyls resulted from the association/dissociation of the inert sugar donor. The Ki of the CMP-3FNeuAc is 660 μM (Chiu, Watts et al. 2004), which indicates the inert sugar donor is not tightly bound to the active site of CstII. Furthermore, the enzyme was only 90% saturated under the conditions used for these measurements. If binding is accompanied by a substantial conformational change, then the association/dissociation might occur on a msec-μsec time-scale detectable by relaxation dispersion techniques.  63  Figure 2.7: Methyl relaxation dispersion curves of apo CstII and its binary complex. Selected relaxation dispersion curves from the signals of δ1-methyl groups of isoleucine residues (A & C) and methyl-groups of isoleucine/valine residues (B & D) in the methyl-TROSY spectra of uniformly deuterated and [1H/13C] selectively methyl labeled (E) apo CstII and (F) its binary complex with CMP-3FNeu5Ac. The relaxation dispersions were done with constant delay time of 40 ms at 25 °C using 600 MHz spectrometer.  2.3.7 NMR determination of the pKa value of His188 2.3.7.1 Peak assignment of catalytic His188 by site-directed mutagenesis To directly measure the pKa value of His188, a sample of CstII-F121D selectively labeled with [13Cɛ1]-histidine was prepared (Schubert, Poon et al. 2007). This protein was also deuterated and measurements carried out in D2O buffer to minimize possible adverse relaxation from background protons. CstII-F121D contains 6 histidines, yet only 5 signals were detected in its 1  H/13C-HSQC spectrum (Figure 2.8A). Fortunately, one of these was assigned to His188 due to  its absence in the corresponding spectrum of CstII-F121D-H188A (Figure 2.8A). The remaining signals, labeled as peaks A-D, remain unassigned.  64  Figure 2.8: Measurement of the pKa value of His188 by NMR spectroscopy. (A) 1H/13C-HSQC spectra of [13Cɛ1]-histidine labeled monomeric CstII-F121D at pH* 7.39 (left) and CstII-F121D-H188A at pH* 7.32 (right) are shown. Five peaks are observed in the spectrum of CstII-F121D, whereas only four peaks are observed with CstII-F121D-H188A. The missing signal corresponds to His188, and the remaining peaks are unassigned. Plots of 1Hɛ1 chemical shifts against pH* for His188, Peak A, and Peak B are shown in (B), (D), and (F), respectively. The corresponding plots of 13Cɛ1 chemical shifts against pH* are shown in (C), (E) and (G), respectively. The lines show fit curves with pKa* values of 6.6 ± 0.1 for His188, 6.9 ± 0.1 for peak A, and 6.7 ± 0.1 for peak B.  65  2.3.7.2 pKa measurement of His188 by NMR spectroscopy The 1H/13C-HSQC spectra of CstII-F121D and CstII-F121D-H188A were recorded as a function of pH. Over the course of these titrations, the signals from His188, peak A, and peak B could be followed, whereas those from peaks C and D could not be detected at pH* values less than ~7. From the pH-dependence of their  Cɛ1 and 1Hɛ1 chemical shifts, the fit pKa* values  13  were 6.6 ± 0.1 for His188, 6.9 ± 0.1 for peak A, and 6.7 ± 0.1 for peak B (Figure 2.8B-G). Consistent with its postulated role as a general base, the NMR-spectroscopically measured pKa* value of 6.6 for His188 agrees well with the apparent whole enzyme pKa of 6.5 determined from the pH-dependence of kcat/Km (Figure 2.2).  2.3.8 Substrate binding mode of CstII STD-NMR spectroscopy is an approach for determining the binding determinants (“epitopes” by analogy to antigen-antibody interactions) of low affinity ligands with proteins. In this technique, the closer in distance or longer in time that a ligand proton is bound to the protein, the more its signal will be attenuated due to saturation transfer from the protein. Importantly, due to weak binding, the effect is observed via the relatively simple NMR spectrum of the free ligand. The STD-NMR spectrum of CMP-3FNeu5Ac in the presence of WT CstII (1:100 enzyme:inhibitor) was obtained (Figure 2.9). The binding epitopes of CMP-3FNeu5Ac are mainly on H5 and H6 in the cytosine ring, H1’ in ribose, and the acetyl methyl group and H9” of Neu5Ac. H6 in the cytosine ring of CMP-3FNeu5Ac shows the strongest STD effect, which indicates it is the proton that is the closest to the protein.  66  Figure 2.9: STD-NMR studies of CMP-3FNeu5Ac binding to CstII-F121D. STD-NMR spectrum (top) and reference spectrum (bottom) show the binding of 1 mM CMP3FNeu5Ac with CstII-F121D (10 μM protein, 20 mM d11-Tris, 150 mM NaCl, 5 mM DTT, pH 7.5 in D2O). The indicated STD signals were standardized using H6 of the cytosine ring, as it has the strongest STD effect (defined as 100%). Based on the STD signals, epitopes in CMP3FNeu5Ac involving binding are H5 and H6 in the cytosine ring, H1’ in ribose ring, and H9a”, H9b”, and Hac in Neu5Ac. The signal of H2O at 4.7 ppm was broadened during spectral processing. The reason for the apparent STD effect for the DTT reducing agent is unknown. Although there are STD effects on the acetyl methyl group and H9” of Neu5Ac, the magnitudes are relatively small. In the crystal structure of CstII binary complex, the Neu5Ac moiety is exposed to the solvent even though the lid motif covers the active site (Figure 2.10). At the same time, the acetyl methyl group interacts with the lid motif. Note that in the crystal structure of the tetrameric CstII, the electron density of lid motif was only observed in 1 of 4 subunits, indicating the lid motif is still dynamic when CMP-3FNeu5Ac is bound. If the lid motif remains in an opened and closed equilibrium upon CMP-3FNeu5Ac binding, the acetyl methyl group of CMP-3FNeu5Ac might not be saturated by transfer from the protein.  67  Figure 2.10: Surface presentation of CstII binary complex. The surface representation of the crystal structure of the CstII binary complex with CMP3FNeu5Ac shows that the 3FNeu5Ac moiety of the sugar donor analog is still exposed to the solvent even though the flexible motif (green) covers the active site. The protons of acetyl methyl group (Hac”) of CMP-3FNeu5Ac also interact with the lid motif when it is “closed”. The proposed lactose binding site is indicated by the blue dashed line.  STD-NMR studies of the CstII binding mode of the sugar acceptor analog lactose in the presence of CMP-3FNeu5Ac were also attempted (Figure 2.11). Note that kinetic studies indicate that CMP-Neu5Ac/CMP-3FNeu5Ac is required for formation of sugar binding site. No STD signal was observed for lactose, likely due to its very high Km value (35 mM) (Chiu, Watts et al. 2004). However, STD signals from CMP-3FNeu5Ac could still be observed. The strongest binding epitope was changed from H6 in the cytosine ring of CMP-3FNeu5Ac in the CstIIF121D•CMP-3FNe5Ac complex to the acetyl methyl group of CMP-3FNeu5Ac in CstIIF121D•CMP-3FNeu5Ac•lactose complex. Two explanations for this observation are proposed.  68  One possibility is that the binding orientation of this inert donor is changed by the presence of lactose. However, in the overlaid crystal structures of CstII binary complexes with CMP3FNeu5Ac and with CMP and trisaccharide, orientations of CMP moieties overlapped perfectly (Figure 1.11). Therefore, a change of the binding mode of CMP-3FNeu5Ac seems an unlikely explanation. A second possibility is that the acetyl methyl group of CMP-3FNeu5Ac forms part of the acceptor binding site (Figure 2.10), and thus its motions or precise contacts with CstII are altered in the presence of lactose, thereby leading to increased saturation transfer.  Figure 2.11: Binding of lactose with CstII-F121D-CMP-3FNeu5Ac complex. STD-NMR spectrum (top) and reference spectrum (bottom) to investigate the binding of 1 mM lactose with CstII-F121D•CMP-3FNeu5Ac complex (10 μM protein, 20 mM d11-Tris, 150 mM NaCl, 1 mM CMP-3FNeu5Ac, 5 mM DTT, pH 7.5 in D2O). No STD signal for lactose was observed (3.5 ppm – 4.5 ppm), possibly due to the low binding affinity. However, the STD spectrum of CMP-3FNeu5Ac was changed in the presence of lactose with more pronounced effects for Neu5Ac. STD signals were standardized with Hac of Neu5Ac (defined as 100%). The signal of H2O at 4.7 ppm was removed during spectral processing.  69  2.4  DISCUSSION AND CONCLUSIONS The crystal structures of both monofunctional CstI and bifunctional CstII have provided  important insights into the mode of donor substrate binding, as well as testable hypotheses regarding the inverting mechanism of sialyl-transfer. To examine the proposal that His188 serves as the catalytic base, we have used kinetic studies and NMR spectroscopy to determine the pKa value of this important active site residue. Dynamic properties of the protein were provided by NMR relaxation dispersion studies. At the same time, STD-NMR spectroscopy was employed to investigate the sugar acceptor lactose binding.  2.4.1 Monomerization of CstII Preliminary attempts to measure the pKa value of His188 in tetrameric wild-type CstII by NMR spectroscopy proved unsuccessful, suggesting that monomerization of the protein would be necessary for this analysis. Fortunately, the active site of CstII is contained entirely within a subunit, thus allowing the generation of active monomers by mutating interfacial hydrophobic residues to polar or charged species (Table 2.1). Note that the interface between the subunits of CstII contains multiple aromatic residues (i.e. Tyr60, Tyr98, His85, Phe121, Tyr125, and Phe126) indicating that self-association results at least in part from stacking and hydrophobic interactions between these side-chains from adjacent monomers. Several of the tested mutants remained tetrameric, while others lost activity upon monomerization. Fortunately, CstII-F121D and CstII-Y125Q were found to be monomeric by gel filtration, static light scattering, and NMR spectroscopic measurements. Both of these variants retained ~30% of the wild-type activity. The CD spectra of wild-type CstII, CstII-F121D, and CstII-Y125Q were nearly indistinguishable (data not shown), indicating that no significant change in secondary structure results from either mutation. Furthermore, the crystal structure of CstII-Y125Q complexed with CMP-3FNeu5Ac superimposes closely with that of the wild-type protein with no measureable difference in conformation at the crystallographic subunit interface or at the active site. Interestingly, although monomeric in solution, the tetrameric state of CstII-Y125Q is reconstituted by the four-fold crystallographic axis. This is suggestive of a possible shift in a monomer/tetramer equilibrium towards the tetramer upon crystallization. Even though the sugar donor analog CMP-3FNeu5Ac was clearly observed in the active site of CstII-Y125Q, electron density from part of the lid domain (residues 157-161 and 180-186) was absent, indicating 70  disorder in this region of the enzyme. In the case of the wild-type protein complexed with CMP, the lid domain is also disordered in crystals. However, when bound to CMP-3FNeu5Ac, wild type CstII crystallizes as a tetramer in the asymmetric unit, with the lid domain becoming observable in 1 of 4 subunits (Chiu, Lairson et al. 2007). The differing mobility of the lid domain in these various species may be due to crystal packing restraints and/or to long range effects of the mutations (either directly or via self-association equilibria) on the internal dynamics of CstII. If monomeric CstII-F121D and CstII-Y125Q mutants are active and retain wild-type tertiary structure, why has CstII evolved to exist as a tetramer? This is unlikely to be due to any requirement for allosteric behavior, since no cooperativity has been observed in substrate binding with wild-type enzyme (L. Lairson, PhD thesis, Univ. of British Columbia, 2007). To analyze for any such effect in the monomerized species, the CstII-Y125Q mutant was assayed with saturating concentrations of donor CMP-Neu5Ac and acceptor lactose as a function of enzyme concentration (0 – 2.5 μM). A linear relationship was observed between the kcat/Km and the amount of enzyme (data not shown), also indicating a lack of cooperativity such as that which could result from concentration effects on a monomer/tetramer equilibrium of this mutant species. One possibility is simply that tetramerization confers greater stability. This is consistent with the higher Tm value of wild-type CstII relative to the two monomeric species (Table 2.1). Alternatively, because its C-terminal membrane association sequences are expected to be aligned along the same face of the tetramer, oligomerization may also be required for effective attachment of CstII to the bacterial inner membrane or interaction with other partner proteins involved in the synthesis or export of LOS (Chiu, Lairson et al. 2007).  2.4.2 NMR spectroscopic analysis of monomeric forms of CstII The 1H/15N-HSQC spectra of CstII-F121D and CstII-Y125Q are obviously improved compared with that of tetrameric wild-type enzyme, confirming that the two mutants are predominantly monomeric and folded properly (Figure 2.5). Of the two variants, the 1H/15NHSQC spectrum of CstII-F121D is slightly better than that of CstII-Y125Q. This is consistent with the greater stability of CstII-F121D relative to CstII-Y125Q, as measured by Tm values (Table 2.1), and by the observation that the former species is less prone to aggregation at the concentrations (~100 μM) used for NMR measurements. Given the remoteness of residue 125  71  from the active site, significant differences in structure and enzymatic mechanism between these two mutants would not be anticipated. Although monomerization of CstII vastly improved its  1  H/15N-HSQC spectra,  approximately half of the expected 1HN-15N cross peaks still remained undetected. The absence of these signals is unlikely to be a simple result of the ~30 kDa mass of CstII-F121D and CstIIY125Q (albeit rather large for NMR analysis) as the relaxation properties of all residues within well-ordered regions of a globular protein should be comparable. Furthermore, 1H/15N-TROSYHSQC spectra of uniformly deuterated samples of CstII-Y125Q were not significantly improved relative to standard 1H/15N-HSQC spectra, indicating that this is not simply a problem of rapid relaxation due to high molecular mass (data not shown). The absence of numerous 1H/15N-HSQC signals likely results from conformational exchange broadening on the chemical shift (msec-sec) time-scale. This could arise from two pathways. The first of these might involve internal backbone motions of regions of monomeric CstII. Of course, it is tempting to speculate that such motions might be important for enzymatic activity. Indeed, in the crystal structure of wild-type CstII, the lid domain lacks detectable electron density in the apo protein, and is observed in only 1 out of 4 monomer units when bound with CMP-3FNeu5Ac (Chiu, Lairson et al. 2007). This indicates that the lid domain is dynamic and that inhibitor binding does alter the structure and mobility of the CstII active site, at least in the crystalline state. However, the lid domain remained disordered in the crystal structure of CstII-Y125Q in complex with CMP-3FNeu5Ac, and the 1H/15N-HSQC spectra of both monomeric mutants in the presence of saturating CMP-3FNeu5Ac were of similar quality to those obtained without any bound ligands (not shown). Thus, motions of the lid domain associated with crystallographic disorder may occur on timescales other than those leading to the postulated 1H/15N-HSQC exchange broadening. An alternative explanation for the missing NMR signals could be an equilibrium with a sufficient population of tetramers (or other oligomeric forms) to alter the relaxation properties of residues whose chemical shifts differ between these species. Although no evidence for tetramer formation was obtained from analytical gel filtration chromatography and static light scattering experiments, these measurements were carried out at concentrations lower than those used for NMR experiments. Distinguishing these two pathways will require additional studies, including the assignment of the ~150 peaks detectable in the 1  H/15N-HSQC spectra of CstII-Y125Q or CstII-F121D. 72  Methyl-TROSY spectra of background-deuterated, selectively [1H/13C]-methyl labeled CstII-Y125Q with and without substrates were also obtained (Figure 2.6). Interestingly, most of the expected methyl signals were detected in the spectra of this monomeric mutant, indicating that the chemical shifts of these sidechain moieties are less sensitive those of the backbone amides to any postulated conformational exchange in CstII. Although unassigned, the ~13 missing peaks in the methyl-TROSY spectrum of the apo enzyme might be associated with the flexible lid domain, as essentially the full complement of expected signals is detected in the presence of saturating amounts of the stable donor substrate analogue CMP-3FNeu5Ac. Methyl relaxation dispersion experiments, which sensitively detect conformational exchange on the msec-sec timescale, were also carried out with selectively [ 2H/13C]-methyl-labeled CstIIY125Q (Figure 2.7). Of the ~70 detectable methyl groups in the apo enzyme, ~18 showed dispersion profiles indicative of conformational exchange. However, upon addition of 5 mM CMP-3FNeu5Ac, only 3 of these methyl groups showed reduced dispersion suggestive of dampened motions. Conversely, ~19 out of ~80 methyl groups in the substrate-bound enzyme still exhibited dispersion profiles, and 4 of these did not show any such behaviour in the apo form. Together, these NMR spectroscopic and X-ray crystallographic data suggest that CstII exhibits a complex repertoire of dynamics in both its unbound and CMP-3FNeu5Ac bound states.  2.4.3 Catalytic His188 In glycosidases and inverting glycosyltransferases, with some exceptions, the role of the catalytic general base is typically played by the side chain carboxylate group of a glutamate or aspartate residue (Zechel and Withers 2001; Lairson, Henrissat et al. 2008). However, in the crystal structure of CstII complexed with CMP-3FNeu5Ac, there are no suitably disposed carboxylate groups in the active site. The closest is that of Glu57, which is 14 Å away from the anomeric reaction center. Of the side chains situated adjacent to the anomeric carbon of CMP3FNeu5Ac (Figure 2.3B; Asn31 (3.9 Å), Asn51 (4.0 Å), Ser132 (6.0 Å), and His188 (4.8 Å)), His188 is the only feasible general base candidate (Chiu, Lairson et al. 2007). This residue is situated appropriately on the alpha face of the donor sugar and adjacent to an open cleft in the enzyme, which would be the most obvious site to bind the acceptor sugar. Thus, the imidazole side chain is suitably positioned to act as the base in catalysis, deprotonating the incoming 73  hydroxyl group of the acceptor during nucleophilic attack at the anomeric carbon of the donor sugar. Consistent with this key proposed catalytic role, mutation of His188 to alanine results in a loss of all detectable CstII transferase activities (Chiu, Lairson et al. 2007). Further precedence for the role of His188 as the base catalyst derives from the fact that side chain imidazoles are found to play this role in other classes of enzymes (Legler, Massiah et al. 2002; Whiteson, Chen et al. 2007). In particular, studies on several human ST's (ST3Gal-I, ST8Sia-II, and ST8Sia-IV) have revealed the identity of an invariant histidine residue that is essential for enzymatic activity (Close, Mendiratta et al. 2003; Jeanneau, Chazalet et al. 2004). The recently reported crystal structures of sialyltransferases CstI (Chiu, Lairson et al. 2007) and PmST1 (Ni, Chokhawala et al. 2007) also revealed active site histidine residues with nearly identical positioning to that of His188 in CstII. Whether the histidine residue functions as the base in the case of PmST1 is unclear. Another possible candidate for the base catalyst is the carboxylate moiety of the donor substrate itself (Figure 2.3). This contention would be consistent with the complete absence of positively-charged residues in the immediate vicinity of the carboxylate, in contrast to what is seen with sialidases, which feature three arginines surrounding the substrate carboxylate (Burmeister, Ruigrok et al. 1992; Luo, Li et al. 1998; Crennell, Takimoto et al. 2000; Amaya, Buschiazzo et al. 2003). Thus, the substrate carboxyl would be expected to have an elevated pKa value when bound to CstII. However, the observed in-plane conformation of the carboxylate is not consistent with a role as base catalyst, as such a role would require that the moiety be oriented perpendicular to the ring plane (Figure 2.3A). While it is still possible that its conformation could change upon formation of a ternary complex with the acceptor sugar, it is interesting to note that this near-planar conformation of the carboxylate is what has been predicted by ab initio calculations and supported by kinetic isotope effect (KIE) experiments to exist during the transition state for the formation of a sialyloxocarbenium ion during the spontaneous hydrolysis of CMP-Neu5Ac (Horenstein 1997). Additionally, more recent KIE experiments have cast serious doubt on the ability of the carboxylate of CMP-Neu5Ac to function as an intramolecular base catalyst during spontaneous hydrolysis (Horenstein and Bruner 1998). Consistent with a mechanism involving a general base catalyst, kcat/Km for wild-type CstII increases with rising pH and displays an apparent pKa value of 6.5 ± 0.1. Note that kcat/Km is the 74  second order rate constant for the reaction of free enzyme and substrate, and thus this value could reflect a deprotonation event in either CstII or CMP-Neu5Ac. Since the CMP-Neu5Ac carboxylic acid moiety has a distinctly lower pKa value of 2.6 (Vimr, Kalivoda et al. 2004), it is reasonable to assign this kinetically relevant pKa of 6.5 to the general base in CstII. Such a value could correspond to a histidine with a relatively “normal” pKa or to an aspartic or glutamic acid with a moderately perturbed pKa. Based on NMR-monitored pH titrations, His188 has a sitespecific pKa* value of 6.6 ± 0.1. This is in close agreement with the pKa value governing the pHdependence of kcat/Km. In parallel with crystallographic studies and the observed chemical rescue of the H188A mutant (Figure 2.1), these data strongly support the hypothesis that CstII indeed employs His188 as the general base in a direct displacement S N2-like inverting mechanism.  2.4.4 Substrate binding mode and dynamics of the lid motif of CstII Crystal structures of the CstII binary complex with CMP (PDB code: 1RO8) or CMP3FNeu5Ac (PDB code: 1RO7) and the ternary complex with CMP and the trisaccharide Neu5Ac-α-2,3-Gal-β-1,3-GalNAc (PDB code: 2X61) provided some insights into the initial and final steps of the sialyltransfer reaction by CstII (Chiu, Watts et al. 2004; Lee, Lairson et al. 2011). However, the related binding modes of both sugar donor and sugar acceptor remain unclear. In an attempt to address this question, STD-NMR spectra were recorded for CMP3FNeu5Ac and lactose in the presence of CstII-F121D (Figure 2.9 and Figure 2.11). Unfortunately, no STD effect was observed for lactose probably due to its low affinity. However, the relatively STD effects observed for the acetyl methyl group and H9” of CMP-3FNeu5Ac implicate the dynamic motion of the lid motif (Figure 2.9). Although the lid motif covers the entire active site upon CMP-3FNeu5Ac binding, the Neu5Ac moiety is exposed to the solvent in the crystal structure of the CstII binary complex with the acetyl methyl group interacting with the lid motif (Figure 2.10). Note that in the crystal structure of the tetrameric CstII, electron density for the lid motif was only observed in 1 of 4 subunits, indicating that the lid motif might still be mobile with CMP-3FNeu5Ac bound. If the lid motif interconvert between “opened” and “closed” states with CMP-3FNeu5Ac bound, the acetyl methyl group of CMP-3FNeu5Ac might not be saturated by cross-relaxation from the irradiated protein.  75  2.4.5 Proposed conformational models of the CstII reaction pathway By combining the information from the known CstII structures, with kinetic and NMR spectroscopic data reported in this thesis, a model linking substrate binding and conformational changes of the lid motif of CstII along its reaction pathway can be proposed (Figure 2.12). In apo CstII, the flexible lid motif exists in an equilibrium of “opened” and “closed” states. The absence of approximately half of the expected signals in the 1H/15N-HSQC spectrum of CstII-F121D could result from this conformational equilibrium occurring on an msec-μsec timescale (Figure 2.5B). When the sugar donor CMP-Neu5Ac is bound, the motion of the lid domain is dampened and the equilibrium shifted towards the “closed” state. This is confirmed by the observation of electron density of the lid motif in one of the four subunits in the crystal structure of CstII binary complex (Chiu, Watts et al. 2004). At the same time, the fact that electron density from the lid motif is absent in the other three subunits indicates that lid is still mobile in the presence of the sugar donor. This conclusion is also supported by the small STD effects of the acetyl methyl group of CMP-3FNeu5Ac (Figure 2.9), as well as the observation that more than half of the signals in the 1H/15N-HSQC spectrum of the CstII-F121D binary complex are still absent.  76  Figure 2.12: Model of substrate binding modes and dynamics of the lid motif of CstII along its reaction pathway.  No crystal structure of the CstII ternary complex with sugar donor and acceptor has been solved, possibly due to the weak binding of lactose (Km = 35 mM) rendering crystallization difficult. At the same time, the association and dissociation rates are fast. Hence, no STD effects of lactose could be observed (Figure 2.11). More importantly, the STD effect of acetyl methyl group of CMP-3FNeu5Ac was dramatically increased in the presence of lactose (Figure 2.12). The effect is probably not due to the change of its proximity within the enzyme active site because the positions of CMP moieties in both the CstII binary complex with CMP (Figure 1.6A) and the ternary complex with CMP and trisaccharide (Figure 1.11A) are exactly the same. Therefore, the STD effect might be caused by lactose altering the motions or precise contacts of the CMP-3FNeu5Ac methyl with CstII. 77  Upon completion of the α-2,3-sialyl-transfer reaction, CMP and the trisaccharide are produced and the lid motif is mobile as evidenced by the lack of observable electron density for the lid domain in the crystal structure of the CstII ternary complex (Lee, Lairson et al. 2011). After dissociation of the trisaccharide, CMP remains in the active site and the lid motif is still dynamic, as suggested by the absence of electron density for the lid motif in the crystal structure of the CstII binary complex with CMP (Chiu, Watts et al. 2004). Together, the effects of substrate binding and the motion of the lid motif of CstII along the α-2,3-sialyl-transfer pathway leads to the model shown in Figure 2.12.  .  78  Chapter 3 NMR spectral assignment of lipooligosaccharide α-1,4galactosyltransferase  (LgtC)  from  Neisseria  meningitidis  LgtC is responsible for the transfer of α-galactose from sugar donor UDP-Gal to the LOS terminal sugar acceptor lactose. Crystal structures of its binary and ternary complexes with only sugar donor analog UDP-2FGal or both sugar donor analog and sugar acceptor analog 4’deoxylactose provided key insights into the galactosyl-transfer mechanism. Combining with the kinetic analyses and MD simulations, its enzymatic mechanism appears to involve a "front-side attack" SNi mechanism or a SNi-like mechanism with a short-lived oxocarbenium-phosphate ion pair intermediate. Furthermore, based upon X-ray crystallographic studies, two flexible loops were proposed to become ordered over the active site of LgtC upon sugar donor binding. Accordingly, NMR spectroscopy was used to investigate to the dynamic properties of the enzyme with an emphasis on the roles of these motions in catalysis. The amide 1H/15N-TROSY-HSQC and methyl-TROSY spectra of LgtC were partially assigned using a variety of NMR spectroscopic approaches, combined with mutagenesis of all the isoleucine residues in the protein. More than expected number of methyl signals was observed, indicating that LgtC adopts multiple conformational states.  79  3.1 INTRODUCTION 3.1.1 Crystal structures of LgtC The structures of LgtC with a UDP-2-deoxy-2-fluoro-galactose (UDP-2FGal) sugar donor analog in the presence and absence of the acceptor sugar analog 4’-deoxylactose have been determined using X-ray crystallography (Figure 1.7) (Persson, Ly et al. 2001). The LgtC binary and ternary structures are essentially the same. The full sugar acceptor binding site is proposed to be only formed upon the binding of a sugar donor in the active site and concomitant ordering of two dynamic loop regions. These loops are likely an opened conformational state when there is no sugar donor (Persson, Ly et al. 2001; Snajdrova, Kulhanek et al. 2004). However, there is no X-ray crystallographic structural information for apo LgtC to support or refute this hypothesis.  3.1.2  Proposed enzymatic mechanism of LgtC Similarly, a detailed description of enzymatic mechanism of LgtC remains to be obtained.  At first, based on the well-defined studied in retaining GHs, a double displacement S N2 mechanism was proposed for LgtC. By analogy with hexosaminidases, the side-chain amide of Gln189 was cautiously proposed as the catalytic nucleophile in this glycosyltransferase (Zechel and Withers 2000) (Figure 1.13). However, 3% residual enzyme activity was observed in a Q189A mutant lacking the glutamine side-chain, thus bringing into question the essential role of this residue. Later, an attempt was carried out to trap the putative covalent glycosyl-enzyme (CGE) intermediate using ESI-MS to investigate the covalent modification of the LgtC-Q189E mutant with UDP-galactose (Lairson, Chiu et al. 2004). Surprisingly, the trapped residue was Asp190 and not Glu189. In the crystal structure of the wild-type enzyme, the distance between the carboxylic side-chain of Asp190 and anomeric carbon is ~ 9 Å in the ternary complex previously solved. Thus formation of the covalent bond between the carboxylic side-chain of Asp190 and galactose in LgtC-Q189E would require a dramatic conformational change. Alternatively, the results of this study could be an artifact as the sample was acidified and unfolded before MS analysis. Therefore, due to no possible or clear catalytic nucleophile, it is unlikely for LgtC to follow double displacement SN2 reaction mechanism. Based on extensive kinetic studies, as well as recent theoretical calculations, LgtC appears to exploit an S Ni-like  80  mechanism (Figure 1.5) (Ly, Lougheed et al. 2002; Lairson, Chiu et al. 2004; Tvaroska 2004; Lairson, Henrissat et al. 2008; Gomez, Polyak et al. 2012).  3.1.3  Proposed conformational dynamics of LgtC The proposed “front-side attack” mechanism with net retention of stereochemistry would  require at least localized conformational and electrostatic changes to stabilize a short-lived oxocarbenium-phosphate ion pair intermediate. This transition state has to be stabilized in a suitable electrostatic environment by the rearrangements of the residues in the active site of LgtC. Hence, dynamic properties of LgtC and its catalytic mechanism could be highly correlated. To provide experimental evidence for this hypothesis, NMR spectroscopy was used to investigate the substrate binding and structural dynamics of LgtC in its apo and substrate-bound forms. The requisite first step towards these studies was to assign the NMR spectra of LgtC.  81  3.2  METHODS  3.2.1  Cloning and site-directed mutagenesis of LgtC The previously described gene encoding "wild-type" LgtC, with the mutation C128/174S  and a deletion of the C-terminal 25 residue membrane association sequence was cloned into the pET22a (Novagen) vector with C-terminal His6-tag (Persson, Ly et al. 2001). Primers with sequences and  5’-CCCGGGCATATGGACATCGTATTTGCGGCAGACGACAACTATGCC-3’  5’-GGGCCCCTCGAGGCCCTGGAAATACAAGTTTTCGTGCGGGACGGCAAGTTT-  GCCGCGCCATTCTTC-3’, that also encode a TEV protease cleavage site, were used to clone the “wild-type” LgtC gene. Additional mutations were introduced sequentially using the QuikChange site-directed mutagenesis kit (Stratagene).  3.2.2  Protein expression and purification The His6-tagged LgtC proteins were expressed in E. coli BL21 (DE3) cells. The cells  were grown in 2xYT broth media at 37 °C to an OD600 of 0.8, and then induced with IPTG at a final concentration of 0.5 mM. After further growth at 16 °C for 16 hours, the cells were harvested by centrifugation and lysed by sonication in the presence of 50 mg/L lysozyme (Sigma). The cell debris was removed by centrifuging at 15,000 rpm in a Sorvall SS32 rotor and LgtC was isolated from the supernatant using a HisTrap HP column (GE Healthcare). The His 6tag of the purified recombinant protein was then cleaved by TEV protease. The construct of TEV protease was provided by the Structural Genomics Consortium and the enzyme was expressed and purified in the laboratory according to the manufacturer’s protocol. The tag cleavage reaction was done by dialysis in the buffer consisting of 50 mM Tris, 0.5 mM EDTA, 1 mM TCEP, pH 8.0 at room temperature with TEV protease (8 nM) for 16 hrs. After TEV protease digestion, HisTrap HP was used to remove the His 6-tag, uncleaved His6-tagged proteins, and the tagged protease. The C-terminal hexapeptide Glu-Asn-Leu-Tyr-Phe-Gln remained after cleavage of the tag. 15  N-labeled proteins were expressed in M9 minimal media containing 1 g/L  Proteins selectively labeled with  15  N-alanine,  15  N-leucine,  15  N-tyrosine,  15  15  NH4Cl.  N-valine,  15  N-  glutamate, or 15N-aspartate were produced using the auxotrophic strain BL21 (DE3) ilvE, tyrB, aspC, avtA, and trpB, grown in a medium containing a mixture of unlabeled amino acids along with one of the following labeled amino acids: 100 mg/L [ 15N]-L-alanine, 100 mg/L [15N]-L82  leucine, 50 mg/L [15N]-L-tyrosine, 50 mg/L [15N]-L-valine, 500 mg/L [15N]-L-glutamate, or 200 mg/L [15N]-L-aspartate (Sigma-Aldrich) (McIntosh, Wand et al. 1990; Waugh 1996; Venter, Ashcroft et al. 2002). Proteins selectively labeled with Ileδ1-[1H/13C], Leu-[13CH3,  12  CD3] and Val-[13CH3,  12  CD3] in an otherwise deuterated background were expressed as follows. The salts, antibiotics,  and IPTG were dissolved in 99% D2O and lyophilized to remove the exchangeable protium isotopes. The cells were first grown in 2xYT broth media at 37 °C to an OD600 of 0.8 and then diluted in standard M9/H2O media to an OD600 of 0.2. The cells continued to grow at 37 °C to an OD600 of 0.7 and then diluted in standard M9/D2O media containing 1 g/L 15ND4Cl and 3 g/L D[D7]-glucose in 99% D2O (Cambridge Isotope Laboratories) to an OD600 of 0.1. The D2O culture continued to grow until OD600 of 0.5 and was then diluted to OD600 of 0.25. After the cells reached an OD600 of 0.5 again, they were further diluted to OD600 of 0.1. After further growth to OD600 0.25, 70 mg/L of 2-keto-3-D2-4-13C-butyrate and 120 mg/L of 2-keto-3-methyl-D3-3-D14-13C-butyrate (that is, α-ketoisovalerate deuterated at the β-position and with one of the two methyl groups  13  CH3 and the other  12  CD3) were added (Cambridge Isotope Laboratories)  (Tugarinov and Kay 2004; Sprangers and Kay 2007; Sprangers and Kay 2007). Precursor 2-keto3-D2-4-13C-butyrate was prepared in advance from the 2-keto-3-H2-4-13C-butyrate (Cambridge Isotope Laboratories) by incubating in a buffer consisting of 50 mM Na 2DPO4 (D2O-exchanged Na2HPO4), pH* 10 in 99% D2O at 45 °C for 16 hrs to exchange the beta protiums to deuteriums. D2O-exchanged IPTG (0.5 mM final) was added when the culture OD 600 reached 0.8. The cells were then grown at 16 °C for 8 hrs post-induction before harvesting. For the sample used in methyl-TROSY spectral assignments, protein selectively labeled with Ileδ1-[1H/13C], Leu-[13CH3,  12  CD3] and Val-[13CH3,  12  CD3] in an otherwise deuterated  13  C  background was expressed using the same method as described above except 3 g/L D-[D7]-13C6glucose and the methyl labeling precursors of 70 mg/L of 2-keto-3-D2-1,2,3,4-13C-butyrate and 120 mg/L of 2-keto-3-methyl-D3-3-D1-1,2,3,4-13C-butyrate were used (Cambridge Isotope Laboratories) (Tugarinov and Kay 2004; Sprangers and Kay 2007; Sprangers and Kay 2007). Proteins selectively labeled with [13Cɛ1]-histidine in an otherwise deuterated background were biosynthesized in the histidine auxotrophic strain BL21 (DE3) hisG, grown in M9 media containing 50 mg/L [13Cɛ1]-L-histidine (Icon Isotopes), 1 g/L ND4Cl and 10 g/L D-glucose in 99% D2O (Waugh 1996; Venter, Ashcroft et al. 2002). 83  Unless stated otherwise, all purified proteins were concentrated to 300 μM and bufferexchanged into 20 mM Tris, 5 mM TCEP, pH 8.5 in H2O or D2O, using an Amicon Ultra-15 Centrifugal Filter Devices. Protein concentrations were determined by UV absorbance using the predicted 280 value (66,350 M-1 cm-1) (Wilkins, Gasteiger et al. 1999).  3.2.3  Optimization of LgtC NMR buffer conditions LgtC (C128/174S) at a final concentration of 500 μM was used to optimize the buffer  condition. The test solutions were set up in PCR tubes (20 μL each) containing 50 mM of MES buffer for pH 5.0 – 6.0 and 20 mM Tris buffer for pH 7.0 – 9.0 along with varying amounts of additives, such as NaCl, MgCl2, UDP, UDP-2FGal, and lactose, were tested (Table 3.1) and incubated at 30 °C. At Day 7 and Day 11, the solutions were centrifuged (10,000 rpm) at room temperature for 10 min, and the remaining soluble protein in supernatant was quantified by UV absorbance at 280 nm.  3.2.4  CD spectroscopy CD spectra of ~10 μM protein in a 0.2 cm path length cuvette were recorded on Jasco-  J810 spectropolarimeter. Four scans at 50 nm/min were averaged, followed by subtraction of a buffer blank spectrum. Thermal denaturation measurements were monitored at 220 nm with heating at a rate of 1 °C/min. Mid-point unfolding temperatures (Tm) were determined by fitting to a standard two-state model (Pace 1990).  3.2.5  Refolding screen of LgtC A HisTrap HP column (GE Healthcare) was first used to test for on-column refolding of  LgtC-25. The protein was unfolded in denaturing buffer consisting of 20 mM Na 2HPO4, 500 mM NaCl, 20 mM imidazole, and 4 M Gdn-HCl, pH 7.4. The unfolded protein was injected onto the column pre-incubated with denaturing buffer, and washed with a 150 mL gradient of 0 - 4 M Gdn-HCl. The refolded protein was eluted according to the manufacturer’s protocol for a HisTrap HP column and the recovery of soluble protein was quantified by SDS-PAGE. LgtC-25 and LgtC (C128/174S) were also tested for refolding under 96 conditions selected from the literature (Table 3.1) (Vincentelli, Canaan et al. 2004). The refolding solutions were buffered using sodium acetate (pH 4), MES (pH 5 – 6), and Tris (pH 7 – 9). 84  Table 3.1: Screening of refolding conditions for LgtC Number  Condition  Number  Condition  1  pH 4, 5 mM MgCl2, 5 mM CaCl2, 500 mM Arg  49  pH 7, 100 mM NDSB-201, 500 mM Arg  2  pH 4, 5 mM MgCl2, 5 mM CaCl2, 0.1 % Triton X100  50  pH 7, 5 mM MgCl2, 5 mM CaCl2, 0.1 % PEG3350  3  pH 4, 5 mM MgCl2, 5 mM CaCl2, 100 mM NaCl  51  pH 7, 500 mM Gly, 500 mM Pro, 1 % sucrose  4  pH 4, 20 % glycerol  52  pH 7, 5 mM MnCl2, 5 mM MgCl2, 5 mM CaCl2, 500 mM Gly, 500 mM Pro  5  pH 4, 0.1 % PEG2000, 1 % sucrose  53  pH 7, 5 mM MgCl2, 5 mM MnCl2, 5 mM CaCl2, 100 mM NaCl, 100 mM KCl, 0.75 M GdnHCl  6  pH 4, 5 mM MnCl2  54  pH 8, 500 mM Arg  7  pH 4, 100 mM NDSB-201  55  pH 8, 5 mM MnCl2, 0.1 % Triton X-100, 1 % sucrose  8  pH 4, 5 mM MgCl2, 5 mM CaCl2, 100 mM NaCl, 100 mM KCl, 500 mM Arg, 500 mM Gly, 500 mM Pro  56  pH 8, 20 % glycerol, 5 mM MgCl2, 5 mM CaCl2  9  pH 4, 5 mM MgCl2, 5 mM CaCl2, 100 mM KCl, 0.1 % Triton X-100  57  pH 8, 100 mM KCl, 5 mM MgCl2, 5 mM CaCl2, 100 mM NDSB-201, 500 mM Arg  10  pH 4, 5 mM MgCl2, 5 mM CaCl2, 100 mM NDSB201, 1 % sucrose  58  pH 8, 1 % sucrose  11  pH 4, 500 mM Gly, 500 mM Pro, 100 mM KCl  59  pH 8  12  pH 4, 0.1 % PEG3350, 1 % sucrose  60  pH 8, 0.1 % PEG3350, 500 mM Gly, 500 mM Pro  13  pH 5, 5 mM MnCl2, 500 mM Arg  61  pH 8, 5 mM MgCl2, 5 mM CaCl2, 1 % sucrose  14  pH 5, 100 mM NaCl, 500 mM Arg  62  pH 8, 100 mM NaCl, 5 mM MgCl2, 5 mM CaCl2, 500 mM Pro, 500 mM Gly  15  pH 5, 0.1 % PEG3350, 5 mM MgCl2, 5 mM CaCl2, 100 mM NDSB-201  63  pH 8, 100 mM NaCl, 100 mM KCl, 5 mM MgCl2, 5 mM CaCl2, 1 % sucrose  16  pH 5, 100 mM NaCl, 100 mM KCl, 500 mM Arg, 500 mM Gly, 500 mM Pro  64  pH 8, 0.1 % PEG2000, 500 mM Pro, 500 mM Gly  17  pH 5, 100 mM NaCl, 100 mM KCl  65  pH 8, 5 mM MgCl2, 5 mM CaCl2  18  pH 5, 0.1 % Triton X-100  66  pH 8, 100 mM NDSB-201  19  pH 5, 5 mM MnCl2, 5 mM CaCl2, 5 mM MgCl2, 100 mM NDSB-201  67  pH 9, 0.1 % PEG3350, 5 mM MgCl2, 5 mM CaCl2, 1 % sucrose  20  pH 5, 5 mM MgCl2, 5 mM CaCl2, 20 % glycerol  68  pH 9, 100 mM KCl, 5 mM MgCl2, 5 mM CaCl2, 100 mM NDSB-201, 1 % sucrose  21  pH 5, 0.1 % PEG2000, 5 mM MgCl2, 5 mM CaCl2, 0.1 % Triton X-100  69  pH 9, 5 mM MgCl2, 5 mM CaCl2  22  pH 5, 5 mM MgCl2, 5 mM CaCl2, 0.1 % Triton X100  70  pH 9, 100 mM NaCl, 0.1 % Triton X-100  85  Table 3.1: Screening of refolding conditions for LgtC Number  Condition  Number  Condition  23  pH 5, 20 % glycerol, 5 mM MgCl2, 5 mM CaCl2, 500 mM Gly, 500 mM Pro  71  pH 9, 5 mM MgCl2, 5 mM CaCl2, 500 mM Gly, 500 mM Pro  24  pH 5, 5 mM MgCl2, 5 mM CaCl2  72  pH 9, 100 mM NaCl, 100 mM KCl, 5 mM MgCl2, 5 mM CaCl2, 0.1 % Triton X-100  25  pH 5, 100 mM KCl, 1 % sucrose  73  pH 9, 20 % glycerol, 100 mM NDSB-201  26  pH 5, 5 mM MgCl2, 5 mM MnCl2, 5 mM CaCl2, 100 mM NaCl, 100 mM KCl, 0.75 M GdnHCl  74  pH 9, 100 mM NaCl, 1 % sucrose  27  pH 6, 0.1 % PEG2000, 5 mM MgCl2, 5 mM CaCl2, 100 mM NDSB-201  75  pH 9, 5 mM MgCl2, 5 mM CaCl2, 500 mM Gly, 500 mM Pro, 500 mM Arg  28  pH 6, 5 mM MgCl2, 5 mM CaCl2, 500 mM Gly, 500 mM Pro, 1 % sucrose  76  pH 9, 0.1 % PEG2000  29  pH 6, 20 % glycerol, 0.1 % Triton X-100, 500 mM Arg  77  pH 9  30  pH 6, 100 mM NaCl, 100 mM KCl, 5 mM MgCl2, 5 mM CaCl2, 1 % sucrose  78  pH 9, 5 mM MnCl2, 5 mM MgCl2, 5 mM CaCl2, 500 mM Arg  31  pH 6, 5 mM MgCl2, 5 mM CaCl2, 100 mM NDSB201, 500 mM Arg  79  pH 9, 500 mM Pro, 500 mM Gly, 500 mM Arg  32  pH 6, 0.1 % PEG2000 0.1 % Triton X-100  80  pH 9, 5 mM MgCl2, 5 mM MnCl2, 5 mM CaCl2, 100 mM NaCl, 100 mM KCl, 0.75 M GdnHCl  33  pH 6, 100 mM KCl  81  pH 4, 5 mM MnCl2, 5 mM CaCl2, 500 mM Arg  34  pH 6, 0.1% Triton X-100, 1% sucrose  82  pH 5, 5 mM MnCl2, 5 mM CaCl2, 500 mM Arg  35  pH 6, 5 mM MgCl2, 5 mM CaCl2, 20 % glycerol  83  pH 6, 5 mM MnCl2, 5 mM CaCl2, 500 mM Arg  36  pH 6, 5 mM MnCl2  84  pH 7, 5 mM MnCl2, 5 mM CaCl2, 500 mM Arg  37  pH 6, 5 mM MgCl2, 5 mM CaCl2, 100 mM NaCl, 500 mM Gly, 500 mM Pro  85  pH 8, 5 mM MnCl2, 5 mM CaCl2, 500 mM Arg  38  pH 6  86  pH 9, 5 mM MnCl2, 5 mM CaCl2, 500 mM Arg  39  pH 6, 0.1 % PEG3350  87  pH 7, 100 mM NaCl, 5 mM MnCl2, 5 mM CaCl2, 500 mM Arg  40  pH 7, 0.1 % PEG2000, 5 mM MgCl2, 5 mM CaCl2, 0.1 % Triton X-100, 1 % sucrose  88  pH 6, 100 mM NaCl, 5 mM MnCl2, 5 mM CaCl2, 500 mM Arg  41  pH 7  89  pH 4, 5 mM MnCl2, 5 mM CaCl2, 500 mM Arg, 0.1 % Triton X-100  42  pH 7, 100 mM KCl  90  pH 5, 5 mM MnCl2, 5 mM CaCl2, 500 mM Arg, 0.1 % Triton X-100  43  pH 7, 100 mM KCl, 100 mM NaCl, 100 mM NDSB-201  91  pH 6, 5 mM MnCl2, 5 mM CaCl2, 500 mM Arg, 0.1 % Triton X-100  44  pH 7, 5 mM MgCl2, 5 mM CaCl2, 20 % glycerol  92  pH 7, 5 mM MnCl2, 5 mM CaCl2, 500 mM Arg, 0.1 % Triton X-100  86  Table 3.1: Screening of refolding conditions for LgtC Number  Condition  Number  Condition  45  pH 7, 0.1 % PEG3350, 0.1 % Triton X-100, 500 mM Arg  93  pH 8, 5 mM MnCl2, 5 mM CaCl2, 500 mM Arg, 0.1 % Triton X-100  46  pH 7, 100 mM NaCl, 100 mM NDSB-201  94  pH 9, 5 mM MnCl2, 5 mM CaCl2, 500 mM Arg, 0.1 % Triton X-100  47  pH 7, 5 mM MgCl2, 5 mM CaCl2, 500 mM Arg  95  pH 8, 100 mM NaCl, 5 mM MnCl2, 5 mM CaCl2, 500 mM Arg, 0.1 % Triton X-100  48  pH 7, 5 mM MgCl2, 5 mM CaCl2  96  ddH2O  For these assays, LgtC (200 μM final) was denatured in a buffer consisting of 6 M GdnHCl, 20 mM Tris, 20 mM MgCl2, and 5 mM DTT. The denatured protein was slowly diluted (1:20) into each of the refolding buffers, followed by the addition of 5 mM final DTT, then left to fold at room temperature in the dark for 3 days. The presence of refolded LgtC was tested using the TLC assay described below and quantified by size exclusion chromatography standardized with native protein.  3.2.6  Activity assays Approximate activities of LgtC and its mutants were determined by a TLC assay with  fluorescent detection using UDP-Gal and bodipy-lactose as substrates (Lairson, Wakarchuk et al. 2007). Steady-state kinetic parameters for LgtC were determined by a continuous coupled enzyme assay (Gosselin, Alhussaini et al. 1994).  3.2.7  NMR spectroscopy NMR spectra were acquired at 25 °C on a Varian Inova and Bruker Avance III 600 MHz  spectrometers equipped with cryogenic  1  H/2D/13C/15N probes, and a Bruker 850 MHz  spectrometer equipped with a regular 1H/2D/31P/13C/15N probe. Spectra were processed using NMRPipe (Delaglio, Grzesiek et al. 1995) and analyzed with SPARKY 3 (Goddard 1999). High concentrations of MgCl2 (10 mM final), UDP (1 mM final), UDP-2FGal (1 mM final), and lactose (300 mM final) relative to their Km values were added into the protein (300 μM final) to form binary and ternary complexes. One-bond sensitivity-enhanced 1H/15N-HSQC spectra (Kay, Keifer et al. 1992) were recorded for  15  N-labeled LgtC and its variants. The sequence-specific  backbone assignments of 2D/13C/15N-labeled LgtC (C128/174S, T273A) (500 μM protein, 20  87  mM Tris, 5 mM TCEP, pH 8.5 in 10 % D2O) were achieved using TROSY-based experiments: 2D 1H/15N-TROSY, 3D TROSY-HNCA, 3D TROSY-HNCACB, 3D TROSY-HN(CO)CA, 3D TROSY-HN(CO)CACB, 3D TROSY-HNCO, and 3D TROSY-HN(CA)CO (Yang and Kay 1999). Methyl-TROSY and methyl relaxation dispersion experiments (Tugarinov and Kay 2003; Sprangers and Kay 2007; Sprangers and Kay 2007; Sprangers, Li et al. 2008) were recorded with [1H/13C]-methyl-labeled deuterated LgtC (C128/174S, T273A). Unless stated otherwise, the samples used in methyl-TROSY experiments were 500 μM protein in D2O with 20 mM d11-Tris, 5 mM TCEP, pH* 8.5. The assignments of methyl protonated [Ile(δ1 only), Leu(13CH3, 12CD3), Val(13CH3,  12  CD3)] U-[15N,13C,2H] LgtC (C128/174S, T273A) were obtained using 3D Ile,Leu-  (HM)CM(CGCBCA)NH, 3D Val-(HM)CM(CBCA)NH, 3D Ile,Leu-HMCM(CGCBCA)CO, 3D Val-HMCM(CBCA)CO, and HMCM[CG]CBCA experiments (Tugarinov and Kay 2003). In addition, a complete set of LgtC mutants with systematic isoleucine to alanine or valine substitutions were generated and selectively 13CH3-methyl labeled (with deuteration) for analysis by methyl-TROSY spectroscopy. One-bond sensitivity-enhanced 1H/13C-HSQC spectra with a CPMG pulse train (Mulder, Spronk et al. 1996) were recorded as a function of pH* with [13Cɛ1]-histidine labeled LgtC(C128/174S, T273A) in D2O sample buffer. The sample pH* values were adjusted by addition of L amounts of 0.05 M DCl and 0.05 M NaOD and measured using an NMR Tube Micro Probe Electrode (IQ Scientific Instruments).  88  3.3 RESULTS 3.3.1 Constructs of LgtC Several different LgtC constructs had been made in the past years. LgtC-19 and LgtC-25 are the truncated forms of the enzyme in which 19 and 25 C-terminal residues that form a potential membrane-associated region are deleted (Wakarchuk, Cunningham et al. 1998). Deletion of a C-terminal section of up to 28 residues does not significantly affect the enzymatic activity. LgtC-19 and LgtC-25 (Figure 3.1A) have been used for published enzyme kinetics and structural studies (Lougheed, Ly et al. 1999; Persson, Ly et al. 2001; Ly, Lougheed et al. 2002; Lairson, Chiu et al. 2004). LgtC-25 was used for structural studies because of its relatively high expression level compared with other truncated versions (Wakarchuk, Cunningham et al. 1998; Persson, Ly et al. 2001). Furthermore, the double cysteine mutant, LgtC (C128/174S) (Figure 3.1E), was used as it appeared to be more stable and produced better crystals (Persson, Ly et al. 2001). Although LgtC (C128/174S) is stable and expressed well, its production is inconvenient because purification requires anion exchange and gel filtration chromatography. To facilitate this, a His6-tag was initially introduced at the N-terminus of LgtC (C128/174S) with an intervening thrombin cleavage site (Figure 3.1B). However, the His6-tag proved resistant to cleavage. Since the efficiency of thrombin is adversely affected by reducing reagents, such as DTT or TCEP, which are required to prevent oxidation of the cysteine residues in LgtC, constructs with a TEV protease site were generated (Figure 3.1C). In contrast to thrombin, TEV protease requires reducing reagents to be active. Unfortunately, the His 6-tag remained uncleavable, possibly due to interactions with and/or steric blockage by the remainder of LgtC. Hence, the His6-tag was moved to the C-terminus of LgtC (C128/174S) and TEV protease cleavage then occurred readily (Figure 3.1D).  89  Figure 3.1: Constructs of LgtC. (A) LgtC-25 with uncleavable His6-tag at its C-terminus. (B) LgtC (C128/174S) with an Nterminal thrombin cleavable His6-tag. (C) LgtC (C128/174S) with an N-terminal TEV protease (TP) cleavable His6-tag. (D) LgtC (C128/174S) with a C-terminal TEV protease cleavable His6tag. (E) LgtC (C128/174S). (F) LgtC (C128/174S, T273A) with a C-terminal TEV protease cleavable His6-tag. Subsequently, by using directed evolution, Dr. Roman Kittl in the Withers laboratory found that the mutation T273A and two other silent mutations helped to increase the expression of LgtC without affecting its enzymatic activity (Figure 3.1F). Although the reason for this behavior is unclear, LgtC (C128/174S, T273A) with a C-terminal TEV protease cleavable His6tag was used for the NMR studies presented herein. Unless specified otherwise, the resulting protein is denoted as “wild-type” LgtC.  3.3.2  Optimization of NMR buffer conditions LgtC (C128/174S) was crystallized in a buffer composed of 50 mM NH4OAc, 3 mM  TCEP, 3 mM MnCl2, pH 7.0 (Persson, Ly et al. 2001). In contrast, for NMR spectroscopy, a protein must remain soluble and monodispersed at high concentrations. Therefore, it was necessary to optimize the buffer conditions for studying LgtC in solution. This was achieved using a sedimentation assay to test the solubility of LgtC under a wide range of conditions (Table 3.2).  90  Table 3.2: Results of NMR buffer condition screen for LgtC Conditions  % Recovery (Day 7)  % Recovery (Day 11)  20 mM Tris, 5 mM DTT, pH 8.3  100  100  1  20 mM Tris, 5 mM DTT, pH 7.0  61  23  2  20 mM Tris, 10 mM NaCl, 5 mM DTT, pH 7.0  54  21  3  20 mM Tris, 25 mM NaCl, 5 mM DTT, pH 7.0  58  24  4  20 mM Tris, 50 mM NaCl, 5 mM DTT, pH 7.0  61  26  5  20 mM Tris, 75 mM NaCl, 5 mM DTT, pH 7.0  52  29  6  20 mM Tris, 100 mM NaCl, 5 mM DTT, pH 7.0  56  30  7  50 mM MES, 5 mM DTT, pH 5.0  9  13  8  50 mM MES, 5 mM DTT, pH 6.0  14  13  Control  9  20 mM Tris, 5 mM DTT, pH 7.0  47  40  10  20 mM Tris, 5 mM DTT, pH 8.0  79  86  11  20 mM Tris, 5 mM DTT, pH 9.0  76  79  12  50 mM MES, 50 mM NaCl, 5 mM DTT, pH 5.0  10  11  13  50 mM MES, 50 mM NaCl, 5 mM DTT, pH 6.0  13  11  14  20 mM Tris, 50 mM NaCl, 5 mM DTT, pH 7.0  52  27  15  20 mM Tris, 50 mM NaCl, 5 mM DTT, pH 8.0  58  15  16  20 mM Tris, 50 mM NaCl, 5 mM DTT, pH 9.0  82  90  17  20 mM Tris, 1 mM EDTA, 5 mM DTT, pH 7.0  54  34  18  20 mM Tris, 5 mM DTT, pH 7.0  46  29  19  20 mM Tris, 0.5 mM MgCl2, 5 mM DTT, pH 7.0  25  30  20  20 mM Tris, 1 mM MgCl2, 5 mM DTT, pH 7.0  52  17  21  20 mM Tris, 2.5 mM MgCl2, 5 mM DTT, pH 7.0  40  15  22  20 mM Tris, 5 mM MgCl2, 5 mM DTT, pH 7.0  16  19  23  20 mM Tris, 10 mM MgCl2, 5 mM DTT, pH 7.0  38  28  24  20 mM Tris, 50 mM MgCl2, 5 mM DTT, pH 7.0  33  17  25  20 mM Tris, 1 μM UDP, 5 mM DTT, pH 7.0  47  24  26  20 mM Tris, 10 μM UDP, 5 mM DTT, pH 7.0  45  26  27  41  27  33  17  28  15  16  11  31  20 mM Tris, 100 μM UDP, 5 mM DTT, pH 7.0 20 mM Tris, 1 μM UDP, 10 mM MgCl2, 5 mM DTT, pH 7.0 20 mM Tris, 10 μM UDP, 10 mM MgCl2, 5 mM DTT, pH 7.0 20 mM Tris, 100 μM UDP, 100 mM MgCl2, 5 mM DTT, pH 7.0 20 mM Tris, 1 μM UDP-2FGal, 5 mM DTT, pH 7.0  47  26  32  20 mM Tris, 10 μM UDP-2FGal, 5 mM DTT, pH 7.0  36  34  33  20 mM Tris, 100 μM UDP-2FGal, 5 mM DTT, pH 7.0 20 mM Tris, 1 μM UDP-2FGal, 10 mM MgCl2, 5 mM DTT, pH 7.0  48  33  25  15  28 29 30  34  91  Table 3.2: Results of NMR buffer condition screen for LgtC Conditions  % Recovery (Day 7)  % Recovery (Day 11)  40  15  44  17  37  20 mM Tris, 10 μM UDP-2FGal, 10 mM MgCl2, 5 mM DTT, pH 7.0 20 mM Tris, 100 μM UDP-2FGal, 100 mM MgCl2 , 5 mM DTT, pH 7.0 20 mM Tris, 10 mM lactose, 5 mM DTT, pH 7.0  52  23  38  20 mM Tris, 50 mM lactose, 5 mM DTT, pH 7.0  51  16  39  20 mM Tris, 100 mM lactose, 5 mM DTT, pH 7.0 20 mM Tris, 10 μM UDP-2FGal, 10 mM MgCl2, 50 mM lactose 5 mM DTT, pH 7.0  67  13  50  18  35 36  40  As summarized in Table 3.2, LgtC (C128/174S) is more soluble and stable under the basic conditions of pH 8.0 and 9.0 (conditions #10 and #11 in Table 3.2). Additives, such as NaCl, UDP, UDP-2FGal, and lactose, did not help to further solubilize LgtC (C128/174S). According to the literature, the divalent metal ion, Mn 2+, stabilizes the UDP leaving group as well as the structure of the enzyme (Persson, Ly et al. 2001). However, Mn2+ cannot be used in general protein NMR studies because it is paramagnetic and will cause severe line-broadening. In the case of LgtC, Mg2+ can replace Mn2+ with only a 50% reduction in activity (B. Lougheed, M.Sc. thesis, 1998). Unfortunately, based on the solubility screen, Mg2+ promotes precipitation of the enzyme. Given that Mn2+ and Mg2+ are required for substrate binding, this leads to difficulties in studying LgtC complexes. As a result of this extensive screen, LgtC (C128/174S) was adjusted to pH 8.5 for subsequent NMR spectroscopic studies. Furthermore, since salt does not affect the solubility of LgtC (C128/174S), a low ionic strength buffer was utilized in order to maximize the sensitivity of the cryo-probe used.  3.3.3  NMR spectroscopy of LgtC  3.3.3.1 1H/15N-HSQC spectrum of 15N-labeled LgtC An early conventional 1H/15N-HSQC spectrum of uniformly  15  N-labeled LgtC-25 is  shown in Figure 3.2. Although the signals from 1HN-15N groups are well dispersed (and indicative of a stable folded structure), only ~120 out of 300 expected peaks are observed. This behavior is both consistent with the size of LgtC (32 kDa) and suggestive of aggregation or 92  extensive conformational exchange broadening. Nevertheless, since the conventional 1H/15NHSQC spectrum looked promising, we expected a marked improvement using TROSY-based experiments with a deuterated sample (Fernandez and Wider 2003; Tugarinov and Kay 2003; Sprangers and Kay 2007; Sprangers and Kay 2007; Sprangers, Li et al. 2008).  Figure 3.2: 1H/15N-HSQC spectrum of 15N-labeled LgtC-25. Shown is the spectrum of uniformly 15N-labeled LgtC-25 (300 μM) in 20 mM Tris, 5 mM DTT, pH 7.5, with 10% D2O. The data were recorded at 25 °C for 16 hrs using a 600 MHz NMR spectrometer. The aliased peak, from an arginine side-chain at 6.6 ppm, is colored in green.  3.3.3.2 Refolding of LgtC Although the sample for TROSY-based NMR experiments was produced in 100% D2O, the backbone amide deuterium atoms of LgtC must be exchanged back to protium atoms for 1Hdetection. The simplest method is to simply unfold the deuterated protein in H 2O buffer, followed by refolding. Therefore, possible refolding conditions of LgtC were explored.  93  3.3.3.2.1  CD spectroscopy of LgtC  The reversible refolding of LgtC had not been reported. Therefore, the thermal and chemical denaturation and refolding of LgtC were studied by CD spectroscopy (Figure 3.3).  Figure 3.3: CD spectra of LgtC-25. (A) CD spectrum of LgtC-25 (10 μM protein, 20 mM Tris, 20 mM MgCl 2, 5 mM DTT, pH 7.5). (B) The thermal denaturation and refolding of the LgtC-25 was monitored at 222 nm. The protein denatured cooperatively with increasing temperature (blue), yet remained irreversibly unfolded upon cooling (red). (C) The Gdn-HCl induced denaturation of LgtC-25 was also monitored at 222 nm. The CD spectrum of LgtC-25, with minima at 207 nm and 222 nm, showed that LgtC indeed contains a high population of α-helices (Figure 3.3A). This, of course, is consistent with the crystal structure of the LgtC ternary complex. Upon heating, LgtC was found to unfold at a mid-point Tm value of 55 °C (Figure 3.3B). Unfortunately, LgtC could not be refolded reversibly with cooling. Similarly, LgtC can tolerate 1 M Gdn-HCl and unfolds at a mid-point concentration of 2.5 M Gdn-HCl (Figure 3.3C). However, refolding did not occur upon removal of the denaturant. Accordingly, an exhaustive screen of conditions was undertaken with the goal of finding conditions under which LgtC might reversibly refold to its active conformation. 94  3.3.3.2.2  Optimization of LgtC refolding conditions  Optimization of refolding conditions can be very complicated because numerous protocols and buffer conditions must be tested, ideally with minimal amounts of protein. Simple rapid dilution, slow dialysis, and on-column refolding are common methods used to refold proteins. Refolding buffer conditions can be dependent upon an enormous number of factors, such as sample pH, overall ionic strength, specific ions, detergents, reducing agents, and other additives. Also, an easy and efficient screening method must be available to determine if the protein is folded. For our first attempt, an on-column protocol with a HisTrap HP column (GE Healthcare) was tested for refolding LgtC-25. The on-column refolding method has the advantage of easy buffer manipulation and is thought to minimize aggregation by immobilizing a protein on a solid matrix. However, Gdn-HCl-denatured LgtC could not be eluted in a soluble form from the HisTrap column. For our second attempt, a simple dilution procedure with a wide range of 96 conditions (modified from (Vincentelli, Canaan et al. 2004)) was used to test for refolding of Gdn-HCl denatured LgtC. Since structure and function are correlated, the folding of the LgtC was monitored by a TLC assay for which the mobility of a fluorescently-labeled substrate changes upon its enzymatic modification. Two different constructs, LgtC-25 and LgtC (C128/174S) were used in these screens.  95  Figure 3.4: Activity assay to screen refolding of LgtC-25. LgtC-25 (1 μM) was examined in each refolding condition of Table 3.1. As shown in the positive control lane (+ve), active LgtC results in slower mobility of a fluorescent substrate. Four refolding buffers (#28: pH 6, 5 mM MgCl2, 5 mM CaCl2, 500 mM L-glycine, 500 mM L-proline, 1 % sucrose, #29: pH 6, 20 % glycerol, 0.1 % Triton X-100, 500 mM L-arginine, #31: pH 6, 5 mM MgCl2, 5 mM CaCl2, 100 mM NDSB-201, 500 mM L-arginine, and #90: pH 5, 5 mM MnCl2, 5 mM CaCl2, 500 mM L-arginine, 0.1 % Triton X-100) were found to yield active LgtC-25 according to a TLC fluorescence reporter assay (Figure 3.4). Across these conditions, the common factors are sample pH and presence of amino acids such as L-glycine, L-arginine, and L-proline. Although the addition of amino acids has been reported to facilitate protein folding (Meng, Park et al. 2001; Ou, Park et al. 2002; Arakawa, Ejima et al. 2007), this pattern was somewhat unexpected because LgtC was found to be more stable and soluble at pH 8.5.  96  Figure 3.5: Activity assay to screen refolding of LgtC (C128/174S). LgtC (C128/174S) (1 μM) in each refolding condition from Table 3.1 was used in the TLC fluorescent assay. As shown in the negative (-ve) and positive control lane (+ve), active LgtC results in slower mobility of a fluorescent substrate.  97  Two refolding buffers (#29: pH 6, 20 % glycerol, 0.1 % Triton X-100, 500 mM Larginine, #31: pH 6, 5 mM MgCl2, 5 mM CaCl2, 100 mM NDSB-201, 500 mM L-arginine) were also found to yield active LgtC (C128/174S) (Figure 3.5). In these conditions, the common factors are pH 6 and the addition of L-arginine. These conditions are similar to those found in refolding tests with LgtC-25. The yields of the refolded protein using these conditions were quantified by using size exclusion chromatography. The maximum yield was 18%, indicating that only a small amount of LgtC could be refolded properly. Therefore, it was not economically or even practically feasible to prepare large quantities of isotopically labeled LgtC for NMR studies by this approach.  3.3.3.2.3  Protium-deuterium exchange of LgtC  As an alternative to an unfolding-refolding strategy, the amides of a deuterated protein can be also reprotonated simply by incubation in H2O buffer for a sufficient time to allow for exchange via native state fluctuations. Fortunately, LgtC was found to be stable and soluble at pH 8.5. Note that amide exchange is base catalyzed and thus occurs more rapidly with increasing sample pH. To test the feasibility of this approach, the exchange of protonated LgtC in D2O buffer was monitored by NMR spectroscopy (Figure 3.6).  98  Figure 3.6: LgtC undergoes rapid amide hydrogen exchange under alkaline conditions. (A) 1H/15N-HSQC spectrum of fully protonated 15N-labeled LgtC in H2O buffer consisting of 20 mM Tris, 5 mM TCEP, pH 8.5 with 10 % D2O. (B-H) The protein was buffer-exchanged into D2O buffer at pH 8.5* by using Amicon Ultra-15 Centrifugal Filter Devices and spectra were acquired as a function of time. (I) The protein was exchanged back to H2O buffer for 4 hrs, confirming that LgtC remains soluble and folded under these conditions and that disappearance of the NMR signals was not due to the unfolding or aggregation of the protein. An aliased peak is colored in green. After incubation for 16 hrs in D2O buffer at pH 8.5*, signals from only 5 amides remained detectable (Figure 3.6C). These signals most likely arise from the most protected amides buried within the core of LgtC. After 56 hours, only 3 peaks were detected (Figure 3.6H). However, LgtC does aggregate after ~ 5 days under these conditions. In order to confirm that the disappearance of the signals was due to H-D exchange and not degradation or aggregation, LgtC was exchanged back into H2O buffer. The reappearance of the full 1H/15N-HSQC spectrum of LgtC confirms that the protein exchanges reversibly under these conditions (Figure 3.6A & I). From the H-D exchange experiment, we concluded that the amides of deuterated LgtC could be re-protonated simply by incubating the protein in H2O buffer at pH* 8.5 for 1-2 days.  99  Thus, TROSY-based NMR experiments could be carried out with deuterated LgtC even though the protein cannot be refolded reversibly with acceptable yields.  3.3.3.3 1H/15N-TROSY-HSQC spectra of uniformly deuterated 15N-labeled LgtC The 1H/15N-TROSY-HSQC spectrum of uniformly deuterated 15N-labeled apo LgtC in 20 mM Tris, 5 mM TCEP, pH 8.5, with 10% D2O, shown in Figure 3.7A, was acquired in 3 hrs at 25 °C using a 600 MHz spectrometer. The quality of this spectrum is clearly superior to the initial conventional 1H/15N-HSQC spectrum acquired in 16 hrs (Figure 3.2).  100  101  Figure 3.7: 1H/15N-TROSY-HSQC spectra of LgtC. 1  H/15N-TROSY-HSQC spectra of (A) the uniformly deuterated 13C/15N-labeled apo LgtC (C128/174S, T723A) and (B) 15N-labeled LgtC binary complex saturated with 1 mM MgCl2 and 1.5 mM UDP-2FGal. Both samples were in 20 mM Tris, 5 mM TCEP, pH 8.5, with 10% D 2O and the spectra were acquired with 600 MHz NMR spectrometer run at 25 °C for 3 hrs and 1 hr, respectively. The deuterated samples were incubated in this buffer for 16 hrs prior to spectral acquisition. (C) The overlaid spectra of apo LgtC (red) and its binary complex (blue) show chemical shift perturbations due to inhibitor binding. Aliased signals are colored in green. Although improved greatly, only ~210 of ~300 expected peaks are observed in the spectrum of apo LgtC. The missing peaks might be due to incomplete D-H exchange, spectral overlap, or unfavorable relaxation or conformational dynamics. Importantly, upon formation of a binary complex with Mg2+ and UDP-2FGal, many amides show chemical shift perturbations, indicative of a change in the structure and dynamics of LgtC (Figure 3.7C). Over the course of a titration experiment, only signals from the free or bound LgtC were detected (not shown). Thus the binding of this substrate analog occurs in slow-exchange on the chemical shift timescale (i.e. kex < , where  is the chemical shift difference between free and bound states). Complete spectral assignments are desirable, but not necessary, because the ternary structure of LgtC had been previously solved by X-ray crystallography (Persson, Ly et al. 2001). If a subset of peaks corresponding to the active site residues could be assigned, then changes in active site dynamics upon substrate binding could be studied. Therefore, TROSY-based 3D 1  H/13C/15N correlation experiments, along with site-directed mutagenesis and selective  amino acid labeling were undertaken to assign as best possible the amide  15  N  15  N and 1HN signals of  LgtC. (Gardner and Kay 1998; Loria, Rance et al. 1999; Yang and Kay 1999; Tossavainen and Permi 2004)  102  3.3.3.3.1  Sequence-specific 15N backbone assignment of LgtC by NMR spectroscopy  2D 1H/15N-TROSY-HSQC, 3D TROSY-HNCA, 3D TROSY-HNCACB, 3D TROSYHN(CO)CA, 3D TROSY-HN(CO)CACB, 3D TROSY-HNCO, and 3D TROSY-HN(CA)CO spectra were recorded to obtain the sequence-specific assignment of the main-chain 1HN, 13  2  13  15  N,  15  and C nuclei of H/ C/ N labeled LgtC (C128/174S, T273A) (Gardner and Kay 1998; Loria, Rance et al. 1999; Yang and Kay 1999; Tossavainen and Permi 2004) (Figure 3.8). The protein sample was left in NMR buffer in H2O for more than 16 hrs before acquisition to ensure reprotonation of the amide groups.  103  Figure 3.8: Partially assigned 1H/15N-TROSY-HSQC spectrum of uniformly deuterated 13  C/15N-labeled LgtC.  (A) The 1H/15N-TROSY-HSQC spectrum of uniformly deuterated 13C/15N labeled LgtC (C128/174S, T273A) (500 μM) in 20 mM Tris, 5 mM TCEP, pH 8.5, with 10% D 2O, was recorded at 25 °C using a 600 MHz NMR spectrometer. (B) The strips from the same 15N planes of (i) 3D TROSY-HNCO and (ii) 3D TROSY-HN(CA)CO spectra show the backbone assignments of I59, R60, and F61 obtained by linking signals from the 13C’ nuclei of residue i-1 (strip i) and both residues i-1 and i (strip ii) with 1HN and 15N of residue i. (C) The strips from the same 15N plane of (i) 3D TROSY-HN(CO)CA and (ii) 3D TROSY-HNCA show the backbone assignments of I59, R60, and F61 by linking signals from the 13Cα of residue i-1 (strip i) and both residues i-1 and i (strip ii) with 1HN and 15N of residue i. (D) The strips from the same 15N plane of (i) 3D TROSY-HN(CO)CACB and (ii) 3D TROSY-HNCACB show the backbone 104  assignments of I59, R60, and F61 by linking signals from the 13Cβ of residue of residue i-1 (strip i) and both residues i-1 and i (strip ii) with 1HN and 15N of residue i.  By using the 3D TROSY-based NMR methods, 146 out of ~ 210 observable peaks in the 1  15  H/ N-TROSY-HSQC spectrum of LgtC were assigned confidently (Figure 3.8A and Table  C.1). Note that this corresponds to only ~50% of the total number of residues in LgtC. Furthermore, when mapped onto the structure of protein (Figure 3.9), it is apparent that the assigned amides are generally distal from the active site. Thus, signals from the active site residues of LgtC were not detected in the NMR spectra of the protein and hence remained unassigned. There are several possible reasons for this result. First of all, severe line-broadening could be caused by msec-μsec timescale conformational dynamics of the protein. Indeed, we speculate that such dynamics might be required within the active site of LgtC for catalysis. Secondly, unassigned amides could be incompletely re-protonated, thus leading to reduced signal-to-noise in 1HN-detected experiments. Although a plausible explanation for buried amides, which are expected to show protection from rapid hydrogen exchange, many unassigned residues are also in exposed (and hence readily exchangeable) surface loop regions. Third, peaks could be overlapping and thus not immediately recognizable. In the end, since the spectral assignment of the active site residues could not be accomplished by global sequence specific methods, sitedirected mutagenesis and selective amino acid labeling were attempted.  105  Figure 3.9: Residues with assigned signals in the 1H/15N-TROSY-HSQC spectrum of LgtC are mapped onto the crystal structure of the protein. The assigned residues are colored in red and the two flexible loops are colored in green. The residues in the active site could not be assigned due to the absence of detectable signals in NMR spectra. Parenthetically, of the assigned amides, the signals of Ile40, Arg53, Leu74, and Leu112 are in the unusual down-field region of the spectrum whereas that of Asp2 is in the unusual upfield region. Based on the X-ray crystal structure of LgtC (PDB code: 1GA8), the amide protons of Ile40 and Arg53 are hydrogen bonded with the carbonyl oxygens of the side-chains of Asp37 and Glu33, respectively, and this could lead to their downfield signals. In contrast, the amide protons of Asp2 and Leu112 are close to imidazole ring of His26 and the aromatic ring of Trp116, respectively, and thus are likely perturbed by aromatic ring current effects. Interestingly, the amide proton of Leu74 does not hydrogen-bond with other atoms and has no nearby aromatic ring. One should keep in mind that the structures of apo LgtC and its ternary complex are not  106  necessarily the same or similar. The conformation of the residue Leu74, in the hinge region of one of the flexible loops, might differ in apo LgtC versus the ternary complex.  3.3.3.3.2  15  N backbone assignment of LgtC by site-directed mutagenesis  Site-directed mutagenesis was attempted to assign the NMR signals for Gln189 and Asp190, as well as surrounding active site residues. Ideally, one peak, which is present in the spectrum of wild-type LgtC, should be absent in the mutant protein, thus providing an immediate, unambiguous assignment. However, as shown in the  1  H/15N-TROSY-HSQC  spectrum of LgtC-Q189E (Figure 3.10A), numerous chemical shift perturbations occurred due to this key mutation, and hence the signal of Gln189 could not be assigned by the site-directed mutagenesis approach. Although less dramatic, the similar situation arose with several additional mutants, including LgtC-I76A, LgtC-I79A, LgtC-I104A, LgtC-V106A, and LgtC-V133A that were generated to assign the methyl signals from several isoleucine and valine residues (Figure 3.10B-F & Figure 3.20). Furthermore, the constructs of LgtC-D190A, LgtC-D190E, LgtCD190N, and LgtC-D190S were made but their proteins could not be expressed in a soluble form despite efforts to optimize bacterial growth temperature, concentration of IPTG, and expression time.  107  Figure 3.10: 1H/15N-TROSY-HSQC spectra of uniformly 15N-labeled LgtC mutants. Spectra of uniformly 15N-labeled (A) LgtC-Q189E, (B) LgtC-I76A, (C) LgtC-I79A, (D) LgtCI104A, (E) LgtC-V106A, and (F) LgtC-V133A in 20 mM Tris, 5 mM TCEP, pH 8.5 with 10% D2O, were acquired at 25 °C at 600 MHz console. The spectra of mutants (blue) were overlaid with that of wild-type LgtC (red). Due to the different 15N spectral widths, the aliased peaks (green) in (B)-(F) correspond to the peaks in the up-field region of the wild-type spectrum.  3.3.3.3.3  Enzymatic activities of LgtC IA mutants  In order to determine whether the single mutation altered the protein structure, the enzymatic activities of the mutants were characterized using TLC fluorescent (Figure 3.11) and coupled-enzyme assays (Table 3.3). Based on the qualitative TLC assay, LgtC-I79A is less active than the wild-type enzyme. Quantitatively, the effects are relatively modest with the largest change being a 12-fold drop in the kcat/Km value for UDP-2FGal and 20-fold drop in that for lactose of LgtC-I79A. Although the affinities of UDP-Gal were not significantly affected for LgtC-I76A, LgtC-I79A, and LgtC-I104A, their turnover rates (kcat) were decreased, suggesting  108  these residues might involve in catalysis. Ile76 and Ile79 are on one of the flexible loops flanking the active site and thus mutations of these residues might perturb substrate binding or change the dynamic properties of the active site (Figure 3.12). In conclusion, since these mutants remained active, their structures should resemble that of wild-type enzyme. However, chemical shifts are exquisitely sensitive to even small structural perturbations, and also depend upon many factors, such as electric field effects. Thus it is difficult to interpret the observed spectral changes.  Figure 3.11: TLC activity assay of the LgtC IA and VA mutants. LgtC mutants (1 μM) were used in the TLC assay. As shown in the negative (-ve control) and positive control lane (WT), LgtC activity is detected by the reduced mobility of a fluorescent substrate.  109  Table 3.3: Steady state kinetic parameters for LgtC IA and VA mutants b  c  UDP-Gal  Lactose  Km (μM)  kcat (s-1)  kcat / Km (μM-1 s-1)  Km (mM)  kcat (s-1)  kcat / Km (mM-1 s-1)  WT LgtC  29  16  0.6  101  23  0.2  a  LgtC-I76A  35  5  0.1  59  6  0.1  a  LgtC-I79A  19  1  0.05  275  2  0.01  a  LgtC-I104A  18  3  0.2  19  3  0.2  a  LgtC-V106A  32  26  0.8  24  26  1.1  a  a  Data were obtained from 15N-labeled proteins by Sophie Weissbach.  b  Data for sugar donor UDP-Gal in the presence of 160 mM acceptor lactose at pH 7.5 and 25 °C.  c  Data for sugar acceptor lactose in the presence of 10 mM donor UDP-Gal at pH 7.5 and 25 °C.  110  Figure 3.12: Crystal structure of LgtC ternary complex showing the isoleucine and valine residues nearest the active site. The isoleucine and valine residues in the active site are colored in red and the flexible loops are in green. Also shown are Mn2+•UDP-2FGal and lactose (carbon, grey; oxygen, red; nitrogen, blue; phosphorus, orange; Mn2+, magenta).  3.3.3.3.4  Selective 15N amino acid labeled LgtC  In a final attempt to assign signals from backbone amides of LgtC, we utilized selective 15  N amino acid labeling. By using this approach, the assignments presented in section 3.3.3.3.1  could be confirmed and the amino acid types of the unassigned peaks could be identified. Samples of LgtC selectively labeled with 15N-alanine, 15N-leucine, 15N-tyrosine, 15N-valine, 15Nglutamate, and  15  N-aspartate were prepared with auxotrophic strain BL21 (DE3) ilvE, tyrB,  aspC, avtA, and trpB. (Waugh 1996) This auxotrophic strain is incapable of metabolically synthesizing alanine, leucine, tyrosine, and valine and thus must use the exogenously provided amino acids for growth. The 1H/15N-TROSY-HSQC spectra of the resulting proteins are shown in Figure 3.13. 111  112  Figure 3.13: 1H/15N-TROSY-HSQC spectra of selectively 15N-labeled LgtC. Spectra of apo LgtC, selectively labeled with (A) 15N-alanine, (B) 15N-aspartate, (C) 15Nglutamate, (D) 15N-leucine, (E) 15N-valine, and (F) 15N-tyrosine in 20 mM Tris, 5 mM TCEP, pH 8.5 with 10% D2O, were acquired at 25 °C with a 600 MHz NMR spectrometer. The indicated peak assignments were transferred from the previously assigned 1H/15N-TROSY-HSQC spectrum of uniformly deuterated 13C/15N-labeled LgtC. In the spectra of 15N-leucine, 15N-valine, and 15N-tyrosine labeled LgtC, there was no labeling of other amino acids (D-F). However, due to metabolic interconversion, additional amino acid were also labeled in LgtC samples prepared from bacteria grown in media containing 15N-alanine, 15N-aspartate, and 15N-glutamate (A-C). Based on the spectra of selectively  15  N-labeled LgtC, the previous main-chain spectral  assignments are confirmed by amino acid type (Figure 3.8 & Figure 3.13). Unfortunately, the missing peaks in the spectrum of uniformly 15N-labeled protein are also absent in the spectra of selectively 15N-labeled protein. This indicates that the lower than expected number of signals is not simply due to peak overlap, but rather most likely result from severe exchange broadening. In the cases of labeling with 15N-alanine, 15N-aspartate, and 15N-glutamate, additional amino acid types were also isotopically enriched. This is due to the essential roles of these amino acids in 113  bacterial metabolism and hence their biosynthetic interconversion (Waugh 1996). The auxotrophic E. coli strain used in these labeling does not block this interconversion. In the cases of labeling with 15N-tyrosine, 15N-leucine, and 15N-valine, only a single type of amino acid was observed in each 1H/15N-TROSY-HSQC spectrum (Figure 3.13D-F). Interestingly, although LgtC contains 24 leucine and 19 valine residues, 29 and 25 peaks were observed in the spectra of the protein labeled with  15  N-leucine and  15  N-valine, respectively (Figure 3.13D & F). The  observation of more than the expected number of peaks implies multiple conformations of LgtC that interconvert slowly on a chemical shift time-scale. This will be addressed further in Chapter 4.  3.3.3.3.5  15  N-glutamate selectively labeled LgtC-Q189A and LgtC-Q189E  In order to identify the signal from Gln189, the 1H/15N-TROSY-HSQC spectrum of uniformly  15  N-labeled LgtC-Q189E was obtained (Figure 3.10A). Unfortunately, this strategy  was hampered by extensive chemical shift perturbations in the spectrum of the mutant relative to that of wild-type LgtC. In an effort to circumvent this problem, both LgtC-Q189A and LgtCQ189E were selectively labeled with  15  N-glutamate as described in section 3.3.3.3.4. The  biosynthesis of glutamine (as well as lysine and alanine) from glutamate is not affected by the auxotrophic strain and thus both residues are labeled. We hoped to identify the signal from Gln189 based on the disappearance of the Gln189 peak in the 1H/15N-TROSY-HSQC spectrum of each selectively labeled protein, along with the appearance of new signals from Ala189 and Glu189 in the spectra of LgtC-Q189A and LgtC-Q189E, respectively. The spectra of the selectively labeled mutants were acquired and overlaid with that of wild-type LgtC (Figure 3.14).  114  115  Figure 3.14: 1H/15N-TROSY-HSQC spectra of selectively  15  N-glutamate labeled LgtC  mutants. Spectra of selectively 15N-glutamate labeled apo (A) LgtC-Q189A and (B) LgtC-Q189E in 20 mM Tris, 5 mM TCEP, pH 8.5 with 10% D2O were acquired at 25 °C with a 600 MHz NMR spectrometer. In addition to glutamate, the glutamine, alanine, and lysine residues were also labeled due to the metabolic pathways of E. coli. The spectra of the mutants (blue and green) were overlaid with that of the wild-type protein (red) and peak assignments taken from the previously assigned 1H/15N-TROSY-HSQC spectrum of uniformly deuterated 13C/15N-labeled LgtC. (C) Overlaid spectra of 15N-glutamate labeled apo LgtC-Q189A (blue) and LgtC-Q189E (green). (D) The crystal structure of LgtC ternary conformation with substrate removed shows the chemical shift perturbations of the assigned residues (red) around the mutated Gln189 (blue). The flexible loops are colored in green. As expected, residues with assigned, perturbed chemical shifts cluster broadly around position 189 in the structure of LgtC (Figure 3.14D). Furthermore, by comparing the spectra of the mutants, chemical shift perturbations of most of the signals due to mutations were similar (Figure 3.14C). Note that there was one missing peak (δ H: 9 ppm, δN: 117 ppm) and two new peaks (δH: 7.6 ppm, δN: 123 ppm) in the overlaid spectra of LgtC-Q189A and LgtC-Q189E (Figure 3.14C). The missing peak might be potentially Gln189 and one of the two new peaks might be Ala189. Chemical shift perturbation of this potential Gln189 peak was also observed when sugar donor analog UDP-2FGal was added (Figure 3.7C). Unfortunately, no signal for that potential Gln189 peak was observed in the 3D TROSY-based spectra used in spectral assignment, thus the identity of that peak (δH: 9 ppm, δN: 117 ppm) could not be confidently determined. On the other hand, there is a possibility that the signal of Gln189 is exchangebroadened and could not be observed in any 1H/15N-TROSY-HSQC spectra.  3.3.3.3.6 Summary of 15N amide backbone assignment of LgtC Approximately one-half of the signals in the 1H/15N-TROSY-HSQC spectrum of apo LgtC were assigned by combination of sequence-specific NMR correlation experiments, sitedirected mutagenesis, and selective  15  N amino acid labeling methods. Although a significant  achievement, it is unfortunate that half of the signals remained unassigned, and more so, that many of these correspond to active site residues. The sequence-specific NMR correlation and selective labeling methods were hampered by line-broadening, and site-directed mutagenesis caused additional chemical shift perturbations of the residues around the mutated site. 116  Furthermore, since UDP-Gal and lactose binding are both slow-exchange events, these assignments could not be transferred from the spectrum of apo LgtC to its complexes through titration experiments. Thus, independent spectral assignments of LgtC binary and ternary complexes would have to be carried out. However, the complexes are not stable and soluble in the presence of Mg2+ for the long time periods (days-weeks) required for the acquisition of the required NMR spectra. Thus, we turned our attention to the methyl groups of LgtC.  3.3.4  Methyl-TROSY of selectively 13C-methyl-labeled LgtC The methyl-TROSY spectrum of deuterated LgtC selectively labeled with Ileδ1-[1H/13C],  Leu-[13CH3,12CD3], and Val-[13CH3,12CD3] is shown in Figure 3.15. The peaks are generally well dispersed, especially in the upfield  13  C region corresponding to the isoleucine residues. We  expected to observe 15 peaks from the 15 isoleucine δ1-methyls and 86 peaks from the 24 leucine and 19 valine methyls. Somewhat surprisingly, approximately 21 and 95 peaks could be counted in the isoleucine and leucine/valine regions, respectively, of the methyl-TROSY spectrum of the apo enzyme (Figure 3.15). Although it is difficult to reliably analyze the crowded spectral regions, the observation of more signals than expected suggested two hypotheses. First of all, the sample used in the NMR experiments could be impure, perhaps containing chemically modified forms of LgtC (i.e. oxidized or proteolyzed) or other unrelated E. coli contaminants. The sample appeared to be >90% pure, giving a single band in SDS-PAGE. However, two peaks differing mass by ~132 Da were observed in the MALDI-TOF mass spectrum (Figure 3.16). The larger species is about 10% in the total population according to the intensities of the peaks in the MS spectra. Interestingly, this phenomenon was also observed in the MALDI-TOF mass spectrum of LgtC-A249X where X = CF3Phe (Figure 3.16) and ESI-MS spectrum of LgtC-Q189E (Figure 4.26). It was thought that galactose or glucose was covalently linked with Asp190 in the active site during protein expression in the bacteria because this phenomenon had previously been observed with LgtC-Q189E (Lairson, Chiu et al. 2004). However, despite many efforts to cleave the proposed galactose or glucose from the active site of LgtC by the addition of UDP, Mn2+, and lactose to complete the glycosyl transfer reaction, the higher molecular weight species remained in the sample according to MALDI-TOF mass spectra. The larger species thus is assumed to be an unknown modification of the enzyme.  117  Importantly, the extra NMR signals were consistently observed in the methyl-TROSY spectra of various mutants (discussed below) and different batches of protein, suggesting that they reflect a property of LgtC and not simply contaminants. An alternative hypothesis is therefore that apo LgtC exists in equilibrium between conformational states, which interconvert slowly enough to yield distinct NMR signals. Support for this hypothesis will be provided in the following sections.  118  Figure 3.15: Methyl-TROSY spectrum of apo LgtC. The methyl-TROSY spectrum of deuterated apo LgtC selectively labeled with Ile δ1-[1H/13C], Leu-[13CH3,12CD3], and Val-[13CH3,12CD3] in 20 mM D11-Tris, 5 mM TCEP, pH* 8.5 in D2O was acquired at 25 °C on a Varian Inova 600 MHz spectrometer equipped with a cryogenic 1 2 13 15 H/ D/ C/ N probe.  119  Figure 3.16: MALDI-TOF-MS spectra of wild-type LgtC and LgtC-A249X. MALDI-TOF-MS spectra of apo (A) wild-type LgtC and (B) LgtC-A249X show small population of species with 132 Da and 210 Da higher masses, respectively.  3.3.4.1 Effect of substrate binding on the methyl-TROSY spectrum of LgtC To determine whether methyl-TROSY experiments could be used to investigate changes in the structural dynamics of LgtC upon substrate binding, we recorded spectra as a function of added Mg2+, UDP, UDP-2FGal, and lactose. In the spectrum of the LgtC product complex formed with 20 mM Mg2+ and 1 mM UDP, 23 isoleucine peaks and approximately 95 leucine and valine peaks were observed (Figure 3.17B). The spectrum is similar to that of the apo LgtC, yet it does show some perturbations indicating that Mg2+ and UDP are bound (Figure 3.17E). In the case of the LgtC binary complex, formed with 10 mM Mg2+ and 1 mM UDP-2FGal, 25 isoleucine peaks and approximately 97 leucine and valine peaks were observed (Figure 3.17C). Two important conclusions can be drawn from this result. First, many of these peaks do not overlap with signals in the methyl-TROSY spectrum of the apo enzyme, indicating that metal and sugar donor analog binding perturbs the structure of LgtC more substantially than observed with the UDP product (Figure 3.17F). However, upon close inspection, some weak peaks from the complex appear to correspond to strong signals in the spectrum of apo enzyme (Figure 3.17F), suggesting that LgtC might not be totally saturated under these conditions. To address  120  this issue, the fraction of Mg2+ and UDP-2FGal were calculated based on their Km values (assumed to reflect the thermodynamic Kd values) according to equation 3-1:  3-1  The Km values of Mg2+ and UDP-2FGal are 370 μM and 2 μM, respectively (Lougheed 1998; Persson, Ly et al. 2001). Hence, the fractions of Mg2+ bound and UDP-2FGal bound were about 0.96 and 0.99, respectively, given the amounts added in the binary complex. Assuming the stock of UDP-2FGal was pure and its measured concentration was correct, these values suggest the enzyme was saturated. Alternatively, there could be an inactive population of LgtC that still gives a similar methyl-TROSY spectrum to that of the active apo enzyme. This population of LgtC might correspond to the modified species observed in the mass spectra mentioned in the previous section (Figure 3.16). When the low intensity peaks are ignored, the number of isoleucine peaks is 20, which is closer to the number expected, whereas the number of leucine and valine peaks remains the same due to the uncertainty of counting signals from crowded spectral regions. This suggests that substrate analog binding reduces the conformational heterogeneity of LgtC. Second, over the course of titration experiment, signals from apo enzyme disappeared as a function of added UDP-2FGal (in the presence of 10 mM Mg2+), whereas distinct signals from the resulting complex appeared. Thus, as also seen with 1H/15N-TROSY-HSQC spectra, the substrate analog binds in the slow exchange limit on the chemical shift timescale (k ex < ). The methyl-TROSY spectrum of the LgtC ternary complex, formed with 10 mM Mg2+, 1 mM UDP-2FGal, and 300 mM lactose, was also acquired (Figure 3.17D). In the spectrum, 19 isoleucine and approximately 95 leucine and valine peaks were observed. A comparison of the spectra of LgtC in its apo, binary, and ternary complexes shows that lactose causes smaller spectral changes than does Mg2+•UDP-2FGal. Although suggestive of smaller structural perturbations, lactose lacks the charge and aromatic ring of the donor analog (both of which can cause chemical shift changes for neighbouring methyl groups). Interestingly, 2 isoleucine peaks disappeared when lactose was bound, indicative of conformational exchange broadening (k ex ~ ) (Figure 3.17G). 121  122  Figure 3.17: Substrate binding monitored by methyl-TROSY spectra of uniformly deuterated and [1H/13C] selectively methyl labeled LgtC. (A) apo LgtC. (B) LgtC product complex saturated with 20 mM Mg2+ and 1 mM UDP. (C) LgtC binary complex saturated with 10 mM Mg2+ and 1 mM UDP-2FGal. (D) LgtC ternary complex saturated with 10 mM Mg2+, 1 mM UDP-2FGal, and 300 mM lactose. (E) Overlaid spectra of apo (red) and binary complex (purple) show a few chemical shifts upon binding of Mg2+ and UDP. (F) Overlaid spectra of apo (red) and binary complex (blue) show more extensive chemical shift changes upon binding of the substrate analog UDP-2FGal. (G) Overlaid spectra of ternary complex (blue) and quaternary complex (green) show chemical shift changes and disappearance of several peaks upon binding of lactose. In order to observe all the peaks in the Ile region, the contour level of spectrum (B) was lowered because only 100 μM protein (versus 500 μM for other samples) was used to acquire the spectrum. From these data, we hypothesize that the enzyme shifts from a conformational equilibrium involving two or more states to a single conformational state upon substrate binding. In order to understand the effects of substrate binding and changes in the conformation and structural dynamics of LgtC, spectral assignments were obviously mandatory. Since the isoleucine peaks in the spectra are well dispersed and resolved from those of leucine and valine groups, and since there are four isoleucine residues (Ile76, Ile79, Ile81, and Ile104) in the active site of LgtC, our efforts focused primarily on these residues (Figure 3.12).  3.3.4.2 Methyl-TROSY spectral assignments of LgtC 3.3.4.2.1  Assignment by NMR spectroscopy  The assignment of the signals from the isoleucine, leucine, and valine methyl groups in apo LgtC was first attempted using multi-dimensional scalar coupling experiments (Tugarinov and Kay 2003). The method, illustrated schematically in Figure 3.18, involves correlating methyl resonances with the previously assigned amide 1HN, correlations along a linearized  13  C',  Cα, and  13  Cβ signals via scalar  13  13  CD spin system. This required the preparation of a methyl  protonated [Ile(δ1 only), Leu(13CH3, 12CD3), Val(13CH3, 12CD3)] sample of otherwise uniformly [15N,13C,2D]-labeled LgtC that had been incubated in H2O buffer to re-protonate the amides via hydrogen exchange.  123  Figure 3.18: Schematic diagrams of the intraresidue magnetization transfer steps from the 13  CH3-methyl groups of Ile, Leu, and Val to the amide (A) 1HN and (B) 13C’.  (A) The sequential intra-residue magnetization transfer steps from the δ1-methyl of Ile and one of the methyls of Leu to 1HN are illustrated for the Ile,Leu-(HM)CM(CGCBCA)NH experiment. The intra-residue magnetization transfer steps from one of the methyls of Val to 1HN are also shown for the Val-(HM)CM(CBCA)NH experiment. (B) The out-and-back intra-residue magnetization transfer steps from δ1-methyl of Ile and one of the methyls of Leu to 13C’ are shown for the Ile,Leu-HMCM(CGCBCA)CO experiment. The intra-residue magnetization transfer steps from one of the methyls of Val to 13C’ are also shown for the ValHMCM(CBCA)CO experiment. The diagram is adapted from (Tugarinov and Kay 2003).  124  The  Ile,Leu-(HM)CM(CGCBCA)NH  and  Val-(HM)CM(CBCA)NH  methyl-HN  correlation experiments proved useful for partially assigning the methyl-TROSY spectra of LgtC (Figure 3.18A) (Tugarinov and Kay 2003). A similar pair of “out-and-back” experiments, Ile,Leu-HMCM(CGCBCA)CO and Val-HMCM(CBCA)CO, were also valuable (Figure 3.18B). However, the sensitivity of these experiments is somewhat low in large proteins. Therefore, a complementary magnetization transfer experiment, HMCM(CG)CBCA, was used to correlate the methyl proton signals with those of the  Cα and  13  Cβ of Ile, Leu, and Val residues derived  13  from TROSY-HNCA and TROSY-HNCACB. The strategies of using these methods in the assignment of the methyl-TROSY spectrum are illustrated in Figure 3.19.  125  126  Figure 3.19: Partial assignment of the methyl-TROSY spectra of uniformly deuterated and [1H/13C] selectively methyl labeled apo LgtC obtained using magnetization transfer experiments. (A) Methyl-TROSY spectrum of uniformly deuterated and [1H/13C] selectively methyl labeled apo LgtC. (B) Assignment of the δ1-methyl of Ile169 in the (i) methyl-TROSY spectrum obtained by correlating the (ii) previously assigned amide 1H/15N-TROSY-HSQC signal of this residue with its 13C’ via (iii) HMCM(CGCBCA)CO and (iv) TROSY-HN(CA)CO spectra and with its 13Cα and 13Cβ via (v) HMCM(CG)CBCA, (vi) TROSY-HNCACB, and (vii) TROSYHNCA spectra. (C) Assignment of δ-methyl groups of Leu195 in the (i) methyl-TROSY spectrum obtained by correlating the (ii) previously assigned amide 1H/15N-TROSY-HSQC spectrum signal of this residue with its 13C’ via (iii) HMCM(CGCBCA)CO and (iv) TROSYHN(CA)CO spectra (13CO plane (180.3 ppm) and with its 13Cα and 13Cβ via the HMCM(CG)CBCA of (v) peak i and (vi) peak ii, (vii) along with TROSY-HNCACB, and (viii) TROSY-HNCA spectra. Red/green peaks are positive/negative signs and the peaks marked with asterisks belong to the i – 1 residues. In the spectra of HMCM(CG)CBCA for Ile and Leu methyl assignments, 3 signals with alternating phases, corresponding to 13Cα, 13Cβ, and 13Cγ, were observed (Figure 3.19B (v) and C (v & vi)). In the spectra of HMCM(CG)CBCA for Val methyl assignments, 2 signals with opposite phases, corresponding to  13 α  C and  Cβ, were observed (spectra not shown). These  13  signals were correlated to the TROSY-HNCA and TROSY-HNCACB spectra in order to obtain the peak assignments in methyl-TROSY (Figure 3.19B (vi & vii) and C (vii & viii)). In parallel, HMCM(CGCBCA)CO spectra were acquired (Figure 3.19B (iv) and C spectra (iv)). The signals were correlated to the previously assigned peaks in TROSY-HN(CA)CO (Figure 3.19B (iii) and C  (iii)).  One  should  notice  that  the  TROSY-HN(CA)CO  correlated  plane  of  HMCM(CGCBCA)CO for Leu and Val contains two signals (Figure 3.19C (iv)). These two carbonyl signals are from the same Leu/Val residues. By using the scalar coupling NMR method, signals from 9 isoleucine, 5 leucine, and 7 valine residues were assigned (Figure 3.19A). Most of the assigned residues are located away from the active site of LgtC (Figure 3.19B). This of course reflects the fact that mainchain 1NH and  13  C' assignments are unavailable for many amides near the active site of the enzyme. One  exception is Val133, which is near the active site and interacts with lactose. This will be discussed in the next chapter. Unfortunately, this approach could not be applied to substrate-bound forms of LgtC as the protein aggregated within the time period required to record the necessary 3D spectra (~ 2 127  days for each experiment). Furthermore, since Mg2+•UDP-2FGal binds in the slow exchange regime, it was not possible to use titration experiments to track chemical shift changes upon complex formation. In order to investigate the substrate binding and structural dynamics of LgtC, the spectral assignments of the remaining methyl groups, especially the ones near the active site, had to be obtained in apo and bound states using a site-directed mutagenesis approach.  3.3.4.2.2  Assignment of methyl-TROSY spectra by site-directed alanine mutagenesis  In an initial attempt to assign the methyl-TROSY signals of active site residues in LgtC, Ile76, Ile79, Ile81, Ile104, Val106, and Val133, were mutated to alanine (Figure 3.12). With the help of a visiting German exchange student, Sophie Weissbach, the genes encoding the LgtCI76A, LgtC-I79A, LgtC-I81A, LgtC-I104A, LgtC-V106A, and LgtC-V133A mutants (IA and VA mutants) were made and deuterated proteins selectively labeled with Ileδ1-[1H/13C], Leu[13CH3,12CD3], and Val-[13CH3,12CD3] were prepared. These mutants were active as confirmed by kinetic measurements (Table 3.3) and hence their structures should be similar to wild-type enzyme. The methyl-TROSY spectra of the mutants were acquired and are shown in Figure 3.20B-G. Ideally, when compared to the spectrum of the wild-type protein, 1 peak should be absent due to an isoleucine mutation and 2 due to a valine mutation. However, in the cases of I76A, I79A, I81A, I104A, and V106A mutants, multiple spectral changes were observed, making it difficult to confidently identify which signal was indeed absent and which was shifted, presumably due to a conformational perturbation resulting from the mutation. In the end, only the two signals from Val133, identified previously by the NMR magnetization transfer method, could be confirmed (Figure 3.20G). Although somewhat disappointing, it is noteworthy that three of the isoleucine mutants (I76A, I79A, and I81A) appear to perturb a common set of poorly dispersed signals (circled in Figure 3.20A). These residues are within or flanking one of the active site loops, and likely conformationally linked (i.e. mutation of any one perturbs the NMR signals of the others). As will be shown in section 3.3.4.2.3, this set of methyl-TROSY peaks indeed arise from these residues.  128  Figure 3.20: Methyl-TROSY spectra of uniformly deuterated and [1H/13C] selectively methyl labeled apo (A) wild-type LgtC, (B) LgtC-I76A, (C) LgtC-I79A, (D) LgtC-I81A, (E) LgtC-I104A, (F) LgtC-V106A, and (G) LgtC-V133A. In panels B-G, the spectra of the mutants (green) are superimposed over those of the wild-type protein (red). The potential peaks corresponding to active site isoleucine residues are circled in blue.  129  As a step towards assigning NMR signals from the binary LgtC complex, methylTROSY spectra were also acquired for the mutants saturated with 10 mM Mg 2+ and 500 μM UDP-2FGal (Figure 3.21). Due to the better spectral dispersion of the inhibitor-bound enzyme, it was easier to distinguish absent versus shifted signals. Thus, the peaks from Ile76, Ile81, Ile104, and Val133 could be assigned with reasonable confidence (Figure 3.21B, D, E & G). However, in the spectra of LgtC-I79A and LgtC-V106A (Figure 3.21C & F), the changes were more extensive, precluding assignment of the peaks corresponding to these two residues.  130  Figure 3.21: Methyl-TROSY spectra of uniformly deuterated and [1H/13C] selectively methyl labeled (A) wild-type LgtC, (B) LgtC-I76A, (C) LgtC-I79A, (D) LgtC-I81A, (E) LgtC-I104A, (F) LgtC-V106A, and (G) LgtC-V133A binary complex. The mutant binary complexes were saturated with 10 mM Mg 2+ and 500 μM UDP-2FGal. The spectra of wild-type LgtC are colored in blue and those of mutants are colored in green. All of the peaks in the spectrum of the LgtC-I104A binary complex (E) look like doublets due to an unknown NMR acquisition error. However, this error does not affect the spectral assignment of Ile104 in the binary complex.  131  3.3.4.2.3  Assignment of methyl-TROSY spectra by site-directed isoleucine to valine mutagenesis  Site-directed mutagenesis is a common, albeit tedious, approach for assigning methyl groups of some large protein complexes. For example, it was successfully used to assign the Alaβ and Ile-δ1 methyl groups of the 468-kDa multimeric aminopeptidase, PhTET2 (Amero, Asuncion Dura et al. 2011). Key to this accomplishment was the generation of a complete set of single mutants with every isoleucine and alanine in the protein changed to leucine and valine, respectively. By considering the patterns of methyl-TROSY spectral perturbations for the entire set, the researchers were able to confidently distinguish absent versus shifted peaks and obtain desired assignments. Another example is the spectral assignments of Ile-δ1, Val-γ, and Leu-δ methyl groups of the 230-kDa nucleosome complex (Kato, van Ingen et al. 2011). The authors assigned the methyl groups by mutating a set of isoleucine residues to valine and a set of leucine and valine residues to isoleucine in histones H2A, H2B, H3, and H4 in the nuclesome complex. This was also facilitated by reconstituting the nuclesome with only one labeled histone at a time. Then a series of NOESY spectra were acquired to assign the rest of the residues, based on knowledge of the crystal structure of the nucleosome. We therefore decided to follow this strategy, mutating each isoleucine in LgtC to a valine residue. We chose valine, rather than alanine as we used initially or leucine as used by (Amero, Asuncion Dura et al. 2011), because this is also a hydrophobic β-branched residue that differs from isoleucine by the loss of only one -CH2- group. In principle, this will not introduce any new unfavorable contacts as might occur with a leucine substitution, while retaining most van der Waals contacts that would otherwise be lost with an alanine substitution. Thus, the isoleucine to valine mutations should have minimal effects on the structure and NMR spectra of the protein. Each of the 15 isoleucine residues of LgtC was mutated to valine individually (IV mutants). Two of these mutants, LgtC-I3V and LgtC-I153V, could not be expressed. Fortunately, these isoleucine residues had been already assigned by scalar correlations, as described previously. The activities of the remainder of IV 13 mutants were determined by TLC assays in order to make sure each mutation did not alter the structure significantly (Figure 3.22).  132  Figure 3.22: TLC activity assay of the LgtC IV mutants. LgtC mutants (1 μM) were used in the TLC fluorescent assay. As shown in the negative (-ve control) and positive control lane (WT), LgtC activity is detected by the reduced mobility of a fluorescent substrate. As expected, LgtC-I76V, LgtC-I79V, and LgtC-I104V, in which the mutation sites are in the active site and flexible loop, are less active than the wild-type enzyme (Figure 3.22). Interestingly, mutations of Ile31 and Ile40, which are not nearby the active site, also impaired catalysis. Among the IV mutants, mutation of Ile31 has the largest reduction of the enzymatic activity. As it will be shown later, Ile31 yields two methyl-TROSY signals (Figure 3.23) and exhibited detectable dynamic behavior in apo LgtC (Figure 4.23). This implicates that the activity of LgtC is dependent upon its global structural and dynamic properties. Overall, the IV mutants are active and thus their structures should resemble the one of wild-type enzyme. Therefore, deuterated samples of the 13 IV mutants selectively labeled with Ile δ1-[1H/13C], Leu[13CH3,12CD3], and Val-[13CH3,12CD3] were prepared, and their methyl-TROSY spectra were acquired (Figure 3.23A-M).  133  134  Figure 3.23: Methyl-TROSY spectra of uniformly deuterated and [1H/13C] selectively methyl labeled LgtC IV mutants in their apo-form. Spectra of (A) LgtC-I31V, (B) LgtC-I40V, (C) LgtC-I59V, (D) LgtC-I62V, (E) LgtC-I76V, (F) LgtC-I79V, (G) LgtC-I81V, (H) LgtC-I93V, (I) LgtC-I104V, (J) LgtC-I129V, (K) LgtC-I143V, (L) LgtC-I169V, and (M) LgtC-I191V overlaid with that of wild-type LgtC. The spectra of mutants and the wild-type are colored in green and red, respectively. Assigned peaks from the mutated residues are labeled in black. The peaks that are most significantly shifted in the mutant spectra are labeled in purple. The “a” and “b” peaks, labeled in cyan, arise from two states of the same isoleucine residue, and will be discussed in section 3.3.4.2.4. The relative intensities of the “a” and “b” change upon mutation. (N) Overall assigned methyl-TROSY spectrum of the isoleucine residues of apo wild-type LgtC. (O) The crystal structure of LgtC shows Ile81 is in the hinge of the flexible loop (green), which also contains Ile76 and Ile79. Mutations of Ile81 might alter the structure and/or motions of the loop and perturb the chemical shifts of Ile76 and Ile79. (P) The crystal structure of LgtC shows the residues nearby Ile62. The mutation of Ile62 causes chemical shift perturbation of Ile93 in the spectrum of LgtC-I62V, but not vice versa. The side-chains of isoleucine residues are colored in cyan and their C δ1 methyl are labeled in red. The crystal structure of the LgtC ternary complex (PDB code: 1GA8) with substrates deleted is used for these figures. By comparing individual methyl-TROSY spectra of the apo IV mutants with that of apo wild-type LgtC, the signals from all isoleucine δ1-methyl groups were assigned (Figure 3.23N). The assignments of Ile59, Ile62, Ile93, Ile129, Ile143, Ile169, and Ile191 previously obtained via scalar correlation experiments were also confirmed. In a few cases, such as LgtC-I169V, the only significant spectral change was the loss of one peak, thus providing an immediate assignment (Figure 3.23L). However, in most cases including the active site isoleucines, additional spectral perturbations were observed and thus the comparison of the spectra from several mutants to deduce a consistent set of assignments was required. Some of these perturbations can be rationalized by the proximity of interacting residues in the crystal structure of the enzyme. For example, the δ1-methyls of Ile62 and Ile93 are in van der Waals contact and indeed the chemical shift of one changes upon the mutation of the other (Figure 3.23D, H & P). However, Ile159 is also near Ile93, and Ile159 is within the same β-strand as Ile62, yet neither pair appears coupled (Figure 3.23D, H & P). More perplexing are changes in the signals of Ile31 and Ile104 due to the mutation of the distal Ile129 (Figure 3.23J). This reflects the difficulty in rationalizing chemical shift perturbations based on simple expectations from a static crystal structure. Furthermore the available structure is that of the LgtC binary/ternary complex, which may differ from the structure of apo LgtC.  135  Efforts to assign the active site isoleucine residues in apo LgtC were originally unsuccessful because we only considered 4 IA mutants, and each caused spectral perturbations. Fortunately, effects of the valine mutations in LgtC-I76V, LgtC-I79V, and LgtC-I104V (3.23E, F & I) were relatively smaller than the corresponding alanine mutations (Figure 3.20B, C & E). Combined with knowledge of the effects of all 13 isoleucine mutations, signals from the active site residues, Ile76, Ile79, and Ile104, could be assigned easily by this approach. In the case of LgtC-I81V (Figure3.23G) and the similar LgtC-I81A (Figure 3.20D), the signals from Ile76 and Ile79 are perturbed by the mutation. Ile81 is located at the hinge of a flexible loop, which contains Ile76 and Ile79 (Figure 3.23O). The mutation of Ile81 might influence the motion of the loop and hence affect the chemical environments of Ile76 and Ile79. Note also that the peaks of Ile40 and Ile81 overlap each other. However, by comparing the spectra of the apo-form and binary complex of LgtC-I40V and LgtC-I81V (Figure 3.23B & G, Figure 3.25B & G), the signals from these two residues can be distinguished. This is because Ile40 is distal from the active site in the crystal structure of LgtC, whereas Ile81 is close to the active site. Accordingly, one would expect the chemical shift of Ile40 not to be changed greatly whereas that of Ile81 should be changed upon substrate binding. On the other hand, previously assigned residue Val133 was also confirmed from the spectrum of LgtC-I143V. These two residues are close to each other in the structure of LgtC ternary complex and thus mutation of Ile143 caused chemical shift perturbations of the signal of Val133 (Figure 3.24). Although two signals from Val143 of LgtC-I143V were expected in the spectrum, only one of them was observed (Figure 3.24A). The other one would be expected to be in the crowded regions of the spectrum. These phenomena were also observed for other spectra of the mutants.  136  Figure 3.24: Chemical shift perturbations of Val133 caused by the mutation of I143V. (A) Overlaid methyl-TROSY spectra of apo wild-type LgtC (red) and LgtC-I143V (green) show the region of the Leu/Val signals. Peak assignments of wild-type LgtC and LgtC-I143V are labeled in black and cyan, respectively. (B) The crystal structure of LgtC shows the residues Ile143 and Val133. The side-chains of isoleucine and valine residues are colored in red and yellow, respectively. The crystal structure of the LgtC ternary complex (PDB code: 1GA8) with substrates deleted is used for this figure. From the completely assigned methyl-TROSY spectrum of apo LgtC, Ile3, Ile31, Ile76, Ile104, Ile129, and Ile143 each show two peaks (Figure 3.23N). Furthermore, the relative intensities of the two peaks from each isoleucine changed upon mutation of another. These observations imply that apo LgtC adopts at least two conformational states that are in equilibrium. This will be discussed further in Chapter 4. Since substrate binding to LgtC is a slow-exchange event, the assignments deduced for the apo-form cannot be transferred to the holo-form by a titration experiment. Therefore, each of the IV mutants was saturated with 20 mM Mg2+ and 1 mM UDP-2FGal and their spectra were also acquired (Figure 3.25).  137  138  Figure 3.25: Methyl-TROSY spectra of uniformly deuterated and [1H/13C] selectively methyl labeled LgtC IV mutants saturated with Mg2+ and UDP-2FGal. Spectra of (A) LgtC-I31V, (B) LgtC-I40V, (C) LgtC-I59V, (D) LgtC-I62V, (E) LgtC-I76V, (F) LgtC-I79V, (G) LgtC-I81V, (H) LgtC-I93V, (I) LgtC-I104V, (J) LgtC-I129V, (K) LgtC-I143V, (L) LgtC-I169V, and (M) LgtC-I191V overlaid with that of wild-type LgtC saturated with 10 mM Mg2+ and 1 mM UDP-2FGal. The spectra of mutants and wild-type are colored in green and blue, respectively. Assigned peaks from the mutated residues are labeled in black. The peaks that are shifted in the mutant spectra are labeled in purple. The “a” and “b” peaks, labeled in cyan, represent two states of the same isoleucine residue and will be discussed in section 3.3.4.2.4. (The “b” peaks are most intense in the wild-type binary complex spectrum, whereas the reverse holds for the mutants.) The “c” peak of Ile79 represent the third state of LgtC. (N) Overall assigned methyl-TROSY spectrum of the isoleucine residues of wild-type LgtC binary complex is shown. (O) The crystal structure (without 4’-deoxylactose) of LgtC showing the active site isoleucine residues. Ile40 and Ile104 are closely pointing to each other thus it is reasonable that the chemical shift of Ile104 changes in the spectrum of LgtC-I40V. Ile79 and Ile104 are well separated from one other and on opposite sides of the bound UDP-2FGal. However, mutation of I79V might influence UDP-2FGal binding and hence perturb the chemical shift of Ile104 in the spectrum of LgtC-I79V. The side-chains of isoleucine residues are colored in cyan and their C δ1 methyls are labeled in red. UDP-2FGal is also shown (carbon, grey; oxygen, red; nitrogen, blue; phosphorus, orange; Mn2+, magenta). Similar to the IA mutants, peak assignments were easier to deduce for the IV mutants in their binary complexes than in their apo-form. This is due to the better dispersion of the peaks, especially those of the active site residues, in the spectra of binary complexes. By comparing the full set of methyl-TROSY spectra of binary complexes of the IV mutants with that of the wildtype species, signals from the δ1-methyl groups of all isoleucines were assigned (Figure 3.25N). The results are consistent with the assignments of Ile76, Ile81, and Ile104 obtained using IA mutants. The chemical shift perturbations observed in some spectra, such as that of LgtC-I62V, can be easily rationalized in terms of the LgtC crystal structure (Figure 3.25D). For example, the signal of Ile93 was shifted in the spectrum of LgtC-I62V because of the van der Waals interaction of the δ1-methyl group of Ile62 (Figure 3.23P, Figure 3.25D). Conversely, although one might expect a chemical shift perturbation of Ile59 in the spectrum of the LgtC-I62V binary complex or of Ile62 in the spectrum of the LgtC-I59V binary complex because they are close to each other in amino acid sequence (Figure 3.23P), no such effect was observed in either the apo or substrate-bound forms of LgtC. It is noteworthy that in the spectra of LgtC-I40V and LgtC-I79V, the signal from Ile104 is both shifted and visibly broadened (Figure 3.25B & F). The side-chains of Ile104 and Ile79 are 139  approximately 10 Å apart and are located on the opposite sides of the bound UDP-2FGal (Figure 3.25O). Mutation of Ile79 to a valine may alter the interaction of the donor analog with Ile104, both structurally and dynamically, thereby yielding the observed spectral changes. In contrast, Ile40 is approximately 8 Å from Ile104, yet more distant from the active site of LgtC (Figure 3.25O). Mutation of this residue might alter packing within the hydrophobic core of the enzyme and thus indirectly perturb the interaction of Ile104 with the bound UDP-2FGal. In the spectrum of LgtC-I104V, the signals of Ile40 and Ile79 are not significantly changed (Figure 3.25I). However chemical shifts depend upon the environment of each residue within the protein, and there is no reason to expect reciprocal spectral perturbations due to the mutations of these three isoleucine residues. The spectral assignment of Ile79 in the LgtC binary complex was initially not obtained using IA mutants due to relatively large chemical shift perturbations in the spectrum of the LgtCI79A binary complex (Figure 3.21C). Fortunately, the chemical shift perturbations of the peaks in the spectrum of LgtC-I79V binary complex were relatively smaller (Figure 3.25F). Still, it was difficult to assign Ile79 in the spectrum of the LgtC-I79V binary complex because the intensities of the peaks (b and c) are small and one of them is overlapping with Ile40 (Figure 3.25N). Peak Ile79b refers to the peak that disappeared when lactose was added. Peak Ile79c refers to the peak overlapping with the peak of Ile40 that remained unperturbed when lactose was added. Since the intensities of the Ile79 peaks are small, they were deduced after the spectra of binary complexes of LgtC-I31V, LgtC-I40V, LgtC-I59V, LgtC-I62V, LgtC-I81V, LgtC-I93V, LgtC-I129V, LgtCI143V, LgtC-I169V, and LgtC-I191V were obtained (Figure 3.25A-D, G, H, & J-M). In the spectrum of the LgtC-I40V binary complex, the weak peak overlapping with Ile40 can be deduced as Ile79c (Figure 3.25B). Another peak of I79b (δH: 0.82 ppm, δC: 15.39 ppm) was identified in the spectra of binary complexes of LgtC-I31V, LgtC-I59V, LgtC-I62V, LgtC-I81V, LgtC-I93V, LgtC-I129V, LgtC-I143V, LgtC-I169V, and LgtC-I191V (Figure 3.25A, C, D, G, H, & J-M). The intensities of this Ile79b peak are relatively stronger compared with that in the wildtype binary complex. The difference in intensities of the Ile79 peaks in the wild-type and mutants might arise because the conformation of Ile79b is more favorable in the mutants. The other active site residues, Ile76 and Ile81, were assigned from the spectra of the binary complexes of IV mutants and are consistent with the assignments obtained using IA mutants. From the completely assigned methyl-TROSY spectrum for the binary complex, most 140  of the residues, except for Ile31, Ile40, Ile59, Ile62, Ile81, and Ile104, have two peaks (Figure 3.25N). More importantly, the intensities of these double peaks shifted from population “a” to population “b” upon substrate binding. This phenomenon, which suggests that LgtC exists in multiple conformations in both the apo and substrate bound forms, will be further discussed in Chapter 4.  3.3.4.2.4  PRE NMR experiment of [1H/13C] selectively methyl labeled LgtC binary complex  In order to confirm the assignment of the methyl-TROSY spectra of LgtC, paramagnetic relaxation enhancement (PRE) NMR measurements using Mn2+ were attempted. Mn2+ is the native metal ion bound to the active site of LgtC (Persson, Ly et al. 2001). Since it is paramagnetic, the relaxation rates (or signal line-width) of methyl in LgtC should increase with the inverse 6th power of their separation from the bound metal. Therefore, PRE NMR experiment can be potentially used to confirm the assignments of methyl-TROSY spectra. Since LgtC cannot bind Mg2+/Mn2+ in the absence of UDP-2FGal (section 4.3.1.1), this method cannot be used to confirm the assignments of the methyl-TROSY spectrum of apo LgtC. In the PRE NMR experiment, the LgtC binary complex saturated with Mg2+ and UDP2FGal was titrated with Mn2+. Since the Kd value of Mn2+ (27 μM) is 13-fold lower than that of Mg2+ (370 μM) according to previous kinetic data (Lougheed 1998), we assumed that Mn2+ would be able to compete with Mg2+ for the binding site and eradicate the NMR signals of the residues nearby without changing the structure/spectrum of the LgtC binary complex. The spectra acquired are shown in Figure 3.26.  141  142  Figure 3.26: PRE NMR experiment of LgtC binary complex (A) Methyl-TROSY spectrum of [1H/13C] selectively methyl labeled LgtC binary complex (200 μM protein) saturated with 10 mM MgCl2 and 1 mM UDP-2FGal. The complex was titrated with (B) 25 μM, (C) 50 μM, (D) 100 μM, (E) 200 μM, and (F) 400 μM MnCl2. The methyl-TROSY spectra were acquired at 25 °C using an 850 MHz spectrometer. (G) The plot shows the change of intensity of each isoleucine δ1-methyl group with respect to increasing concentrations of MnCl2. First of all, the methyl-TROSY spectrum of LgtC binary complex acquired using an 850 MHz spectrometer showed better dispersion of several peaks which overlapped in the previous spectrum acquired using a 600 MHz spectrometer (Figure 3.25N & Figure 3.26A). For example, the double peaks of Ile59, Ile93, Ile143, Ile159, and Ile169 were better dispersed when acquired with the higher field spectrometer. Previously, the “a” and “b” peaks were proposed to represent two states of the same isoleucine residue. From the PRE NMR experiment, it was found that the signals of the “a” peaks were still present at a 1:2 ratio of protein and Mn2+ (Figure 3.26F), indicating that LgtC in state “a” does not bind Mn 2+. This observation suggests two possible scenarios. One is that there is an inactive species of LgtC that has a similar structure to the active enzyme but is unable to bind substrates, as all the “a” peaks overlapped with the peaks in the spectrum of apo LgtC (Figure 3.27F). Another scenario is that there are two conformational states of LgtC in equilibrium observed in the methyl-TROSY spectrum of the LgtC binary complex. In other words, there is still a small population of apo LgtC in the sample used to acquire the spectrum of LgtC binary complex, even though the concentrations of Mg 2+ and UDP2FGal in the sample should be enough to saturate the enzyme. Based on these observations, the “a” peaks would be considered as the signals from the isoleucine residues in apo state or “a” state, and “b” peaks would be considerd as the signals from the isoleucine residues in Mg2+•UDP-2FGal bound state or “b” state. More importantly, when the concentrations of Mn2+ and LgtC were equi-molar (200 μM), most of the peaks of the residues nearby the active site (Ile76, Ile79, and Ile104) were eradicated by the paramagnetic effect (Figure 3.26E & G, Table 3.4). Further, the order of disappearance of signals at increasing [Mn2+] is consistent with the distance derived from the crystal structure of the LgtC ternary complex. This result helps to confirm the assignments of the methyl-TROSY spectra of the LgtC binary complex.  143  Table 3.4: Correlation of the distance between Mn2+ and isoleucine δ1-methyl groups with PRE data. [MnCl2] (μM)  a  Residuea  Distance between Mn2+ and Cδ1 of Ile (Å)b  I79a  7.0  I79b  7.0  I104a  8.8  I104b  0  25  50  100  200  400  *  *  *  *  *  *  *  *  8.8  *  *  I76b  9.9  *  *  I159a  13.9  I159b  13.9  I3a  15.0  *  *  *  I3b  15.0  I191a  15.6  *  *  I191b  15.6  *  *  I40  15.8  *  I81  16.9  *  I143a  17.0  I143b  17.0  I129a  17.6  I129b  17.6  I93a  17.8  I93b  17.8  I62  19.1  I31b  19.4  I169a  20.9  I169b  20.9  I59a  22.2  I59b  22.2  The column is sorted in increasing order of the distance between Mn  *  * *  * * * * * *  2+  δ1  and C of isoleucine  residues. b  The distances between Mn2+ and Cδ1 of isoleucine residues were measured based on the crystal  structure of the LgtC ternary complex. The absence of the resonance at specific concentrations of Mn2+ is noted with an asterisk.  144  3.3.4.2.5  Summary of the methyl-TROSY assignments of LgtC  Signals from the isoleucine δ1-methyl groups in the methyl-TROSY spectra of LgtC in its apo-form and the binary complex with Mg2+•UDP-2FGal were completely assigned by the combination of NMR correlation experiments and site-directed mutagenesis methods (Figure 3.27A & C). Since the methyl-TROSY spectra of apo LgtC and LgtC•Mg2+•UDP binary complex (Figure 3.27E) and those of the LgtC•Mg2+•UDP-2FGal binary and LgtC•Mg2+•UDP2FGal•lactose ternary complexes superimpose closely (Figure 3.27G), we can also directly transfer these assignments to the latter UDP-bound and lactose-bound species (Figure 3.27B & D). Although the peaks are shifted in the spectra of the binary complex with UDP (Ile104 and Ile191) and the ternary complex (Ile81 and Ile191) slightly, they could be assigned with high confidence since the chemical shift perturbations are small (Figure 3.27E & G). With the partial spectral assignments of valine and leucine methyl groups in the methyl-TROSY spectrum of apo LgtC and complete spectral assignments of isoleucine δ1-methyl groups in the methyl-TROSY spectra of apo LgtC and its binary and ternary complexes, the structural dynamics of LgtC could be studied, as discussed in Chapter 4.  145  Figure 3.27: Assigned methyl-TROSY spectra of isoleucine residues of uniformly deuterated and [1H/13C] selectively methyl labeled LgtC along its reaction pathway (A) Apo form, (B) binary complex saturated with 10 mM Mg2+ and 1 mM UDP-2FGal, (C) ternary complex saturated with 10 mM Mg2+, 1 mM UDP-2FGal, and 300 mM lactose, and (D) product complex saturated with 20 mM Mg2+ and 1 mM UDP. Also shown are the spectra of the (E) binary UDP-2FGal•Mg2+ complex overlaid on that of the apo enzyme, (F) the ternary UDP2FGal•Mg2+•lactose complex overlaid on that of binary UDP-2FGal•Mg2+ complex, and (G) product UDP•Mg2+ complex overlaid on that of the apo enzyme.  146  3.3.5  H/13C-HSQC spectra of [13Cε1]-histidine-labeled LgtC  1  There are 8 histidine residues in LgtC, of which 7 are defined in the crystal structure of the enzyme (Figure 3.28). The missing His286, which is the C-terminal residue of the construct used in solving the crystal structure, is in the flexible C-terminal region where no electron density was observed. Three histidine residues are in the active site. They are His78, which is in the flexible loop of residues 75-80, His244, which coordinates Mn2+/Mg2+, and His253, which is in the hinge of the flexible loop of residues 246-251. These histidine residues can be monitored via 1H/13C-HSQC spectra of the [13Cε1]-histidine-labeled LgtC (Figure 3.28). In the spectrum of apo LgtC, the histidine peaks are crowded and only 6 signals could be counted (Figure 3.28B). In the spectrum of the binary complex, 1 additional weak, dispersed peak was observed (Figure 3.28C) and the relative intensities of Peaks 1 and 2 are interchanged (Figure 3.28D). Although in principle one could assign these peaks by site-directed mutagenesis, one major concern is that the signals in 1H/13C-HSQC spectra of the [13Cε1]-histidine-labeled LgtC are highly overlapped and thus it might be difficult to distinguish absent and shifted peaks. On the other hand, pH titrations, 13  13  C relaxation dispersion experiment, and PRE experiment had been carried out with  ε1  [ C ]-histidine-labeled LgtC. Due to overlapping of the peaks and aggregation of the proteins, no clear and useful result was obtained. In conclusion, the histidines were found not to be a useful probe for the study of the structure and dynamics of LgtC.  147  Figure 3.28: 1H/13C-HSQC spectra of deuterated [13Cε1]-histidine-labeled LgtC (A) The crystal structure of LgtC ternary complex shows the location of the seven histidine residues (red), of which His78, His244, and His253 are in the active site. UDP-2FGal and 4’deoxylactose are also shown (carbon, grey; oxygen, red; nitrogen, blue; phosphorus, orange; Mn2+, magenta). 1H/13C-HSQC spectra of deuterated [13Cε1]-histidine-labeled (B) apo LgtC (200 μM protein, 20 mM D11-Tris, 5 mM TCEP, pH* 8.5 in D2O) and (C) its binary complex saturated with 10 mM Mg2+ and 500 μM UDP-2FGal were acquired at 25 °C using a 600 MHz spectrometer. (D) Overlaid spectra of deuterated [13Cε1]-histidine-labeled apo LgtC (red) and its ternary complex (blue).  148  3.4 DISCUSSION AND CONCLUSION 3.4.1 General insights concerning LgtC NMR spectra In order to study the structural dynamics of LgtC by NMR spectroscopy, several different types of NMR spectra were acquired. These included the amide 1H/15N-HSQC-TROSY of either uniformly labeled or amino acid selectively  15  N-labeled LgtC, methyl-TROSY of uniformly  deuterated and [1H/13C] selectively methyl labeled protein, and 1H/13C-HSQC of selectively [13Cε1]-histidine-labeled protein (Figure 3.8, Figure 3.17, & Figure 3.28). Despite the use of complementary methods, including scalar correlation experiments, site-directed mutagenesis and selective isotopic labeling, the 1H/15N-HSQC-TROSY and methyl-TROSY spectra of LgtC could not be completely assigned. The spectral assignments were hampered by signal overlapping and line-broadening due to the size and dynamics of LgtC. Fortunately, the isoleucine signals in the methyl-TROSY spectra are well dispersed and could be completely assigned by a combination of NMR correlation experiments and site-directed mutagenesis strategies (Figure 3.27). More importantly, several isoleucine residues are located in the active site of the enzyme and thus serve as a probe for the structural dynamics and substrate binding studies.  3.4.2  Insights into of the mode of substrate binding to LgtC Based on the NMR experiments, we discovered that substrate binding to LgtC is a slow-  exchange event, occurring on the msec to μsec time-scale. However, the affinities of Mg2+ (Kd = 370 μM), Mn2+ (Kd = 27 μM), UDP-Gal (Kd = 30 μM), and UDP-2FGal (Kd = 2 μM) with LgtC are relatively weak (Lougheed 1998; Persson, Ly et al. 2001; Ly, Lougheed et al. 2002). These substrate bindings are weak and thus one might expect their binding to be in the intermediate to fast time-scale. However, Kd is the ratio of the rate constants for dissociation (k off) and association (kon) and thus the relatively weak binding (or ratio of koff and kon) might reflect both slow dissociation and association of the substrates, possibly due to a significant conformational change in LgtC upon complex formation. This will be addressed further in Chapter 4.  3.4.3  Implication of the multi-conformational states of LgtC One interesting observation from the methyl-TROSY spectra of LgtC (as well as the  amide 1H/15N-HSQC-TROSY spectra of selectively  15  N-labeled protein) is that there are more  than the expected number of peaks for some residues (Figure 3.27). In particular, several 149  isoleucines yield two NMR signals. Although this could be due to impurities or chemically modified forms of LgtC, we disfavor this possibility as the protein is sufficiently pure and the multiple signals appear consistently in all samples of wild-type and mutant LgtC. Furthermore, as will be shown in Chapter 4, the relative ratios of these signals change with sample temperature. Thus, these data suggest that LgtC adopts at least two conformational states that interconvert slowly on the chemical shift time-scale. The conformational equilibrium of LgtC is also dependent upon mutation and whether the substrates/products are present (Figure 3.29).  150  Figure 3.29: Proposed different conformational states of LgtC. Four proposed conformational states of LgtC are shown, based on either the methyl-TROSY spectra of either uniformly deuterated and [1H/13C] selectively methyl-labeled or 1H/15NTROSY-HSQC spectra of 15N-tyrosine selectively-labeled apo LgtC and its binary and ternary complexes. (A) State “a” represents the “opened” state of apo LgtC. (B) State “b” represents the “closed” state of LgtC with sugar donor UDP-Gal bound. (C) State “c” represents the “closed” state of LgtC with sugar donor UDP-Gal and acceptor lactose bound. (D) State “d” represents the “opened” state of LgtC with UDP bound. The flexible loops are colored in green and the scaffold of the protein is colored in yellow. Isoleucine residues are colored in blue and tyrosine residues are colored in red.  151  Four conformational states of LgtC are proposed here (Figure 3.29). State “a” represents the “opened” state of apo LgtC. State “b” represents the “closed” state of LgtC upon sugar donor UDP-Gal binding. State “c” represents an additional “closed” state of LgtC upon sugar donor UDP-Gal and acceptor lactose binding. State “d” represents the more “opened-like” state of LgtC upon UDP binding. The proposed multi-conformational states and their equilibria will be addressed further in the next chapter.  152  Chapter 4 Investigation of structural dynamics and substrate binding  of  lipooligosaccharide  α-1,4-  galactosyltransferase (LgtC) from Neisseria meningitdis by NMR spectroscopy  With the spectral assignments obtained from the previous chapter, substrate binding and structural dynamics of LgtC are studied by NMR spectroscopy in this chapter. Using an NMR titration, metal ions (Mg2+) and sugar donor analog (UDP-2FGal) are shown to bind simultaneously to LgtC. More importantly, the observation of a greater number of methyl resonances than that expected indicates that LgtC adopts multiple conformational states. These states, termed "opened" and "closed", are in equilibrium on a time-scale of seconds, and their relative populations change upon mutation and substrate binding. These multiple conformational states were observed in several NMR experiments, including methyl-TROSY and 1H/13C-HMQC exchange spectroscopy. At the same time, methyl relaxation dispersion studies indicated substantial µsec-msec time-scale motions of methyl groups both within and distal to the active site in apo and substrate-bound LgtC. Thus LgtC exhibits a range of dynamic behaviours potentially linked to its catalytic function.  153  4.1 INTRODUCTION 4.1.1  Substrate binding of LgtC From the kinetic analysis, LgtC follows an ordered bi-bi kinetic mechanism with  2+  Mn •UDP-Gal binding before the sugar acceptor (Ly, Lougheed et al. 2002). The full sugar acceptor binding site is proposed to be only formed upon the binding of a sugar donor in the active site, with concomitant ordering of two dynamic loop regions. MD simulations suggested that manganese ion is tightly bound in the active site in the absence of sugar donor (Snajdrova, Kulhanek et al. 2004). However, there are no experimental data showing the order of metal ion and sugar donor binding to the active site. In this chapter, substrate binding to LgtC is monitored by NMR spectroscopy.  4.1.2  Multiple conformational equilibria of LgtC As presented in chapter 3, multiple signals (“a” and “b” peaks) for several isoleucine  residues were observed in the methyl-TROSY spectra of LgtC (Figure 3.27). This is suggestive of a conformational equilibrium between at least 2 states for the apo enzyme. Furthermore, the relative intensities of these signals changed upon Mg2+•UDP-2FGal binding, indicating that the conformational equilibrium is linked to substrate binding, and hence catalysis. At the same time, some unusual long range chemical shift perturbations were observed in the mutants used for methyl-TROSY assignments of apo LgtC (Figure 3.23). This suggests that the structure and dynamics of apo LgtC might be functionally different from the known static crystal structures of the LgtC binary and ternary complexes (Persson, Ly et al. 2001). In this chapter, these features of apo LgtC will be investigated by NMR spectroscopy in order to address the key question regarding possible changes in the conformational states of LgtC along its reaction pathway.  4.1.3  Structural dynamics of LgtC As described previously, LgtC appears to exploit an SNi-like mechanism (Ly, Lougheed  et al. 2002; Lairson, Chiu et al. 2004; Lairson, Henrissat et al. 2008; Gomez, Polyak et al. 2012). This “front-side attack” mechanism with net retention of stereochemistry would require at least localized conformational and electrostatic changes to stabilize a short-lived oxocarbeniumphosphate ion pair intermediate. Furthermore, based on the crystal structure of LgtC complexed with UDP-2FGal and 4’-deoxylactose, two flexible loops (residues 75-80 and residues 246-251) 154  are proposed to become ordered and thereby close the active site. Accordingly, as presented in this chapter, we turned to NMR spectroscopy to investigate the structural dynamics of LgtC along its reaction pathway. The NMR spectroscopic studies of LgtC presented in this chapter help to address this conundrum by providing evidence for significant conformational flexibility in the active site of the enzyme.  155  4.2 METHODS 4.2.1 Protein expression and purification Isotopically labeled LgtC proteins were expressed and purified as described in the Methods section in Chapter 3. LgtC-F132X and LgtC-Y186X with the unnatural amino acids, X = p-trifluoromethyl-phenylalanine (CF3Phe) and p-13C-methoxy-phenylalanine (O13CH3Phe), respectively, were prepared by Adrienne Cheung according to published protocols (Jackson, Hammill et al. 2007; Cellitti, Jones et al. 2008; Graziano, Liu et al. 2008). The O13CH3Phe was synthesized by Dr. Hongming Chen in the Withers' laboratory. Saturating concentrations of 10 mM MgCl2, 1 mM UDP-Gal/UDP-2FGal, and 300 mM lactose were added to form binary and ternary substrate analog complexes.  4.2.2  NMR spectroscopy NMR spectra were acquired on a Varian Inova 500 MHz spectrometer equipped with a  cryogenic 1H/2D/13C/15N probe and a conventional broad band probe for 1  2  13  19  F and  31  P, a Varian  15  Inova 600 MHz Advance spectrometer equipped with a cryogenic H/ D/ C/ N probe, a Bruker 600 MHz spectrometer equipped with cryogenic 1H/2D/13C/15N and 1H/2H/13C/15N/19F probes, and a Bruker 850 MHz Avance spectrometer with a conventional 1H/2H/13C/15N/19F probe. 1D NMR spectra were processed and analyzed using Topspin 3.0. 2D NMR spectra were processed using NMRPipe (Delaglio, Grzesiek et al. 1995) and analyzed with SPARKY 3 (Goddard 1999).  4.2.2.1 31P NMR spectroscopy Either 1 mM UDP or UDP-2FGal was used for 31P NMR-montitored LgtC titrations in a buffer consisting of 20 mM D11-Tris, 5 mM TCEP, 20 mM MgCl2, pH* 8.5 in 99% D2O. Substrate titration experiments with enzyme were performed at 25 °C.  4.2.2.2 19F NMR spectroscopy 1 mM UDP-2FGal was used for  19  F NMR-montitored LgtC titrations in a buffer  consisting of 20 mM D11-Tris, 5 mM TCEP, 20 mM MgCl2, pH* 8.5 in 99% D2O. Substrate titration experiments with enzyme were performed at 25 °C.  156  4.2.2.3 STD-NMR spectroscopy LgtC (10 μM, in 20 mM D11-Tris, 10 mM MgCl2, 5 mM TCEP, pH* 8.5 in 99% D2O) was used for STD-NMR spectroscopy (Angulo, Rademacher et al. 2006). Either 1 mM UDP or 1 mM UDP-2FGal was added to the NMR tube containing the protein. STD-NMR spectra were acquired at 25 °C using 500 MHz and 600 MHz spectrometers. On-resonance irradiation was performed at 0 ppm and off-resonance at 20 ppm. Irradiation was performed using WALTZ16 pulses and 70 ms duration to give a total satuaration time of 1.5 s. STD-NMR spectra were acquired with a total of 2048 scans. Reference spectra were acquired with a total of 512 scans. The STD effect was calculated as the ratio of peak intensities in the presence (I STD) and absence (Iref) of protein saturation (irradiation), using the following equation: (4-1) The STD effects were normalized by assigning a value of 100% to the ligand epitope showing the largest ASTD value.  4.2.2.4 1H/15N-TROSY-HSQC NMR spectroscopy Unless stated otherwise, uniformly  15  N-labeled LgtC or  15  N amino acid selectively  labeled LgtC (500 μM protein, 20 mM Tris, 5 mM TCEP, pH 8.5 in 10 % D2O) were used in acquiring 1H/15N-TROSY-HSQC spectra (Yang and Kay 1999). All spectra were acquired at 25 °C using 500 MHz and 600 MHz spectrometers.  4.2.2.5 Methyl-TROSY NMR spectroscopy Methyl-TROSY and methyl relaxation dispersion experiments (Tugarinov and Kay 2003; Sprangers and Kay 2007; Sprangers and Kay 2007; Sprangers, Li et al. 2008) were recorded with [1H/13C]-methyl labeled deuterated LgtC (C128/174S, T273A). Unless stated otherwise, the samples used in methyl-TROSY experiments were in D2O with 20 mM D11-Tris, 5 mM TCEP, pH* 8.5 (Tugarinov and Kay 2003).  157  4.2.2.5.1  NMR chemical shift perturbation in methyl-TROSY spectra  LgtC selectively labeled with Ileδ1-[1H/13C], Leu-[13CH3, 12CD3] and Val-[13CH3,  12  CD3]  (300 μM protein, 20 mM D11-Tris, 5 mM TCEP, pH* 8.5, 99% D2O), was prepared as described previously and used in substrate binding NMR experiments. All spectra were acquired at 25 °C unless stated otherwise. LgtC•Mg2+•UDP product complex was saturated with 20 mM MgCl 2 and 1 mM UDP, LgtC•Mg2+•UDP-2FGal binary substrate analog complex was saturated with 10 mM MgCl2 and 1 mM UDP-2FGal, and LgtC•Mg2+•UDP-2FGal•lactose ternary complex was saturated with 20 mM MgCl2 and 1 mM UDP-2FGal and 300 mM lactose. To monitor the substrate binding of LgtC, the average, weighted chemical shift perturbations (CSP values) of LgtC isoleucine δ1-methyl groups were calculated by using the following equation: (4-2)  ΔδH and ΔδC are the chemical shift changes of the methyl 1Hδ1 and 13Cδ1, respectively, weighed by the relative carbon (γC) and proton (γH) magnetogyric ratios (i.e. 0.25).  4.2.2.5.2  Conformational exchange measurements  Slow conformational exchange in LgtC selectively labeled with Ileδ1-[1H/13C], Leu[13CH3,  12  CD3] and Val-[13CH3,  12  CD3] was measured using a modification of a published  experiment (Farrow, Zhang et al. 1994). The experiment, which detects exchange of  13  C  longitudinal magnetization in the form of a 1H/13C-HMQC spectrum, was recorded with delay mixing time of 400 ms at 25 °C using a 600 MHz spectrometer (Velyvis, Schachman et al. 2009). In the case of CF3Phe-labeled LgtC-F132X, a simple 2D  19  F-NOESY experiment was  exploited with a transfer time of 250 ms. All measurements were carried out using a 600 MHz spectrometer with cryogenic probes.  4.2.2.5.3  Methyl relaxation dispersion experiment  LgtC selectively labeled with Ileδ1-[1H/13C], Leu-[13CH3, 12CD3] and Val-[13CH3,  12  CD3]  (300 μM protein, 20 mM D11-Tris, 5 mM TCEP, pH* 8.5, 99% D2O), was prepared as described 158  previously and used in  13  C relaxation dispersion experiments. Methyl relaxation dispersion  experiments were done with constant delay times (T) of 20 ms and 40 ms with νCPMG values ranging from 50 Hz to 1000 Hz at 25 °C using 500 MHz, 600 MHz, and 850 MHz spectrometers (Korzhnev, Kloiber et al. 2004). Values of R2,eff were calculated according to the equation:  (4-3)  I(vCMPG) is the intensity of a cross-peak recorded with CPMG pulses over a constant time delay T, and I(0) is that of a control spectrum recorded without the delay. Rex, kex, Δω, and population values were calculated by fitting the dispersion curves using GUARDD (Kleckner and Foster 2012).  4.2.3 LgtC-Q189E glycosyl-enzyme intermediate trapping by NMR spectroscopy 15  N-labeled LgtC-Q189E was expressed and purified using standard procedures.  However, the expression level of the mutant was about 5-fold lower than that of the wild-type enzyme. The mutant (300 μM in 20 mM Tris, 5 mM TCEP, pH 8.5, 10% D2O) was used in the glycosyl-enzyme intermediate trapping experiment. In the NMR tube, 15 Units of PK/LDH (Sigma-Alrich), 20 mM UDP-galactose, and 50 mM phosphoenolpyruvate were added into the protein. A 1H/15N-TROSY-HSQC spectrum of the sample was acquired for an hour before initiating the reaction. Then, 10 mM MgCl 2 was added to initiate the glycosyl transfer reaction and 10× 1-hr 1H/15N-TROSY-HSQC spectra were acquired to monitor the progress of the reaction. Due to aggregation and resultant loss of signal, after the ten initial 1-hr spectra had been acquired, six 2-hrs spectra followed by four 4-hrs spectra were acquired to increase the signal-tonoise.  4.2.4 LgtC-Q189E glycosyl-enzyme intermediate trapping observed by electrospray ionization mass spectrometry After completion of the NMR experiments of section 4.2.3, 15N-labeled LgtC-Q189E and 15  N-labeled LgtC-Q189E-Gal were analyzed by ESI-MS, as described previously (Lairson, Chiu  et al. 2004).  159  4.3 RESULTS  4.3.1 Substrate binding by LgtC 4.3.1.1 Simultaneous binding of the metal ion and sugar donor to LgtC Based on molecular dynamics simulations, it was suggested that LgtC binds Mn 2+ prior to its sugar donor UDP-Gal (Snajdrova, Kulhanek et al. 2004). However, there was no clear experimental test of this binding order. Therefore, NMR spectroscopy was used to monitor the titration of LgtC with Mg2+ and the donor analog UDP-2FGal (Figure 4.1). The addition of 10 mM Mg2+ or 1 mM UDP-2FGal independently did not significantly perturb the methyl-TROSY or 1H/15N-TROSY-HSQC spectra of LgtC (Figure 4.1A, C & D). Thus neither binds LgtC alone. The methyl signals of the solvent-exposed Ile76 and Ile79 were slightly shifted in the presence of the metal ion, possibly because of a non-specific electrostatic effect. In contrast, when both Mg2+ and UDP-2FGal were present, several peaks in the methyl-TROSY and 1H/15N-TROSY-HSQC spectra were significantly shifted (Figure 4.1B & E).  160  Figure 4.1: Simultaneous binding of Mg2+ and UDP-2FGal by LgtC. (A, C, D) Overlaid methyl-TROSY and 1H/15N-TROSY-HSQC spectra of LgtC in the absence (red) and the presence of 10 mM Mg2+ (green) or 1 mM UDP-2FGal (yellow) show no significant differences. Thus, Mg2+ and UDP-2FGal alone do not bind LgtC. (B, E) Overlaid methyl-TROSY and 1H/15N-TROSY-HSQC spectra of LgtC in the absence (red) and presence (blue) of 10 mM Mg2+ and 1 mM UDP-2FGal confirmed that Mg2+•UDP-2FGal binds simultaneously as evidenced by clear chemical shift perturbations. Samples of uniformly deuterated and [1H/13C] selectively methyl labeled LgtC used for methyl-TROSY experiments were in 20 mM d11-Tris, 5 mM TCEP, pH* 8.5 in D2O. Samples of selectively 15N-tyrosine labeled LgtC used for 1H/15N-TROSY-HSQC experiments were in 20 mM Tris, 5 mM TCEP, pH 8.5 in 10% D2O. The spectra were acquired at 25 °C on a 600 MHz spectrometer. Based on these titration experiments, it is clear that Mg2+ and UDP-2FGal (and very likely, Mn2+ and UDP-Gal) bind LgtC simultaneously, rather than independently. This indicates that, in addition to stabilizing the UDP leaving group during galactosyl-transfer reaction, the two positive charges of Mg2+/Mn2+ can also stabilize the two negative charges of UDP-2FGal/UDPGal and thereby facilitate their mutual binding to LgtC.  161  4.3.1.2 Direct observation of UDP and UDP-2FGal Thus far in this thesis, NMR spectroscopy has been used to study substrate and product binding by monitoring changes in the amide and methyl groups of LgtC. However, 31P NMR and 19  F NMR methods can be also used to investigate binding of UDP and UDP-2FGal directly.  Substrate binding monitored by 31P NMR spectroscopy  4.3.1.2.1 31  P NMR spectra of product UDP titrated with LgtC are shown in Figure 4.2. With  increasing concentrations of added LgtC, the two sharp peaks from the 31Pα and 31Pβ of free UDP progressively disappeared and one single broad peak from bound UDP appeared. The relative ratios of free and bound UDP over this titration are consistent with the reported K d value of 80 μM for LgtC binding (Lougheed, Ly et al. 1999). As seen previously, when observing signals from the protein, the UDP and LgtC bind in the slow exchange limit on the chemical shift timescale. Thus both dissociation (koff) and association (kon) rate constants are relatively low and their ratio is low, possibly due to a significant conformational change in LgtC upon complex formation.  162  Figure 4.2: UDP binding of LgtC monitored by 31P NMR spectroscopy. 31  P NMR spectra of UDP (500 μM) with increasing molar ratios of LgtC. The signals from three forms of TCEP are indicated (Krezel, Latajka et al. 2003). The resonances from 31Pα and 31Pβ of free UDP were assigned based on published data for ADP (Geraldes and Castro 1989). The spectra were acquired at 25 °C using a 500 MHz spectrometer. 31  P NMR spectra of UDP-2FGal titrated with LgtC were also acquired to determine the  LgtC-bound state of the sugar donor analog (Figure 4.3). With increasing concentrations of added LgtC, the two sharp peaks from  31 α  P and  31 β  P of free UDP-2FGal progressively  disappeared and two broad peaks from the bound compound appeared, again in the slow exchange limit. In comparison with the titration of UDP in Figure 4.2, UDP-2FGal showed better binding than UDP to LgtC. This is consistent with the reported dissociation constants for UDP2FGal (Km = 2 μM) versus UDP (Kd = 80 μM) and make sense in that the stronger binding of UDP-Gal (Km = 30 μM) is required to replace product UDP in the active site of LgtC for the next round of enzymatic reaction (Lougheed, Ly et al. 1999; Persson, Ly et al. 2001; Ly, Lougheed et al. 2002). Also, in contrast to UDP which yielded one bound 31P signal, the two signals observed  163  with UDP-2FGal indicate that the 31Pα and 31Pβ are in distinctly different chemical environments within the active site of the enzyme.  Figure 4.3: UDP-2FGal binding of LgtC monitored by 31P NMR spectroscopy. 31  P NMR spectra of UDP-2FGal (500 μM) in the presence of increasing concentrations of LgtC. A residual amount of free phosphate from the purification was observed. The resonances from 31 α P and 31Pβ of free UDP-2FGal were assigned based on the literature (Hayashi, Murray et al. 1997). The signals from three forms of TCEP are indicated (Krezel, Latajka et al. 2003). The spectra were acquired at 25 °C using a 500 MHz spectrometer. UDP-Gal is the natural sugar donor for LgtC, but it is hydrolyzed slowly by this enzyme (kcat/Km = 0.0039 μM-1 s-1) (Ly, Lougheed et al. 2002). Therefore, for structural studies, a stable inhibitor, UDP-2FGal, was used as its glycosidic bond cannot be readily cleaved by the enzyme (Hayashi, Murray et al. 1997; Persson, Ly et al. 2001). This is because the strongly electronegative fluorine atom at the C2 position of the galactose moiety withdraws electron density from the glycosidic bond, thereby strengthening the bond to the oxygen of the UDP moiety of UDP-2FGal and disfavoring charge development during the transition state for hydrolysis or galactosyl-transfer. However, as shown in Figure 4.4, LgtC-bound UDP-2FGal is also slowly hydrolyzed to yield bound UDP. Fortunately, complete hydrolysis required over a  164  day and thus UDP-2FGal is still a useful ligand for structural studies of LgtC. Nevertheless, this provides a time restraint for NMR experiments involving the complex of UDP-2FGal with LgtC.  Figure 4.4: Hydrolysis of UDP-2FGal by LgtC monitored by 31P NMR spectroscopy. 31  P NMR spectra show the slow hydrolysis of UDP-2FGal by LgtC (1:2 substrate to protein molar ratio). The signals from three forms of TCEP and also phosphate are indicated (Krezel, Latajka et al. 2003). The spectra were acquired at 25 °C using a 500 MHz spectrometer.  4.3.1.2.2 19  Sugar donor binding monitored by 19F NMR spectroscopy F NMR spectra of UDP-2FGal titrated with LgtC were also acquired (Figure 4.5). In the  spectrum of free UDP-2FGal (with or without excess Mg2+), a double doublet from the fluorine atom at the C2 position of the galactose moiety is observed (Figure 4.5A). Upon addition of aliquots of LgtC, the sharp signals of UDP-2FGal progressively disappeared and a very broad signal from the bound analog appeared. Again, binding occurred in the slow exchange limit on the  19  F chemical shift timescale. After prolonged incubation, the UDP-2FGal was hydrolyzed 165  and sharp signals from the released 2FGal (in a mixture of anomeric forms) appeared. The observation of this α/β-2FGal product further confirmed UDP-2FGal could be hydrolyzed by LgtC, and along with the spectra of Figure 4.4, indicates that only the UDP moiety of the hydrolyzed product remained bound.  Figure 4.5: UDP-2FGal binding of LgtC monitored by 19F NMR spectroscopy. 19  F NMR spectra of UDP-2FGal (500 μM) titrated with Mg2+ and LgtC. The spectra were acquired at 25 °C using a 500 MHz spectrometer.  4.3.1.3 STD-NMR spectroscopy of LgtC STD-NMR is one approach for determining the binding determinants (“epitopes” by analogy to antigen-antibody interactions) of a low affinity ligand with protein. In Chapter 2, this approach was used to characterize the interactions of CMP-3FNeu5Ac and lactose with CstIIF121D (section 2.3.8). Although the mode of UDP-2FGal binding to LgtC had been solved by X-ray crystallography (Persson, Ly et al. 2001), there is no direct information about the product  166  complex formed with UDP. Accordingly, STD-NMR studies were carried out to characterize further the binding modes of UDP and UDP-2FGal with LgtC.  4.3.1.3.1  STD-NMR spectroscopy of UDP binding with LgtC  STD-NMR spectra of UDP in the presence of LgtC were acquired and are shown in Figure 4.6A. As noted above, UDP is a product inhibitor of LgtC (Kd = 80 μM) (Lougheed, Ly et al. 1999). Based on the data, binding epitopes in UDP are H5 and H6 in the uridine ring and to a lesser extent, H1’ in the ribose ring (Figure 4.6A & B). This suggests that there is close interaction between uridine and the enzyme. However, in the available crystal structures, the entire sugar analog is covered by the flexible loops and thus all protons in UDP-2FGal should show similar STD effects (Figure 4.6C). Likewise, all protons in UDP would also be expected to exhibit similar STD effects. One possible explanation for this discrepancy is that the flexible loops are more “opened” in the UDP product complex, thus reducing the effective contacts between the ligand and protein that lead to saturation transfer. However, understanding STD quantitative effects really requires a detailed analysis of the relaxation pathways for protons in the bound and free states and thus caution must be exercised not to overly interpret the data of Figure 4.6.  167  Figure 4.6: STD-NMR studies of UDP binding to LgtC. (A) Reference (top) and STD spectra (bottom) show the binding of UDP to LgtC. STD signals were standardized using H6 of the uridine ring, as it has the strongest STD effect (defined as 100%). (B) Based on the STD signals, epitopes in UDP include H5 and H6 in the uridine ring and H1’ in the ribose ring. The asterisk denotes the STD effect of overlapped signals of H2’ and H4’. Removal of the water signal at 4.7 ppm by post-acquisition processing caused distortion of the adjacent signals. Components in the buffer also show STD effects, possibly due to nonspecific interactions with LgtC (note that their concentrations are at least 5-fold higher than that of UDP). (C) The surface representation of the crystal structure of the LgtC binary complex with UDP-2FGal covered by the flexible loops (green).  168  4.3.1.3.2  STD-NMR spectroscopy of UDP-2FGal binding with LgtC  STD-NMR spectra of UDP-2FGal binding with LgtC were acquired and shown in Figure 4.7. In contrast to UDP, no STD signals were detected. This can be attributed to the tighter binding of UDP-2FGal (Km = 2 μM) which leads to efficient relaxation in the bound state and hence little “transfer of information” to the detected free state as the outcome of slow association (kon) and dissociation (koff), possibly due to the conformational changes of LgtC.  Figure 4.7: STD-NMR studies of UDP-2FGal binding to LgtC. Reference (top) and STD spectra (bottom) show the binding of UDP-2FGal with LgtC. No STD signals were observed due to strong binding of UDP-2FGal with LgtC. The water signal at 4.7 ppm was removed by post-acquisition processing.  4.3.1.4 Substrate binding of LgtC monitored by methyl-TROSY spectroscopy As summarized in Chapter 3, methyl-TROSY spectra of LgtC in its apo-form, binary complexes with Mg2+•UDP-2FGal, ternary complex with Mg2+•UDP-2FGal•lactose and product complex with Mg2+•UDP have been obtained and the δ1-methyl signals of all isoleucine residues assigned (Figure 3.17 & Figure 3.27). Importantly, substrate binding is a slow-exchange event, and some isoleucine residues show two signals, “a” and “b”, corresponding to two different states of LgtC that are in equilibrium (Figure 3.27). 169  4.3.1.4.1  Spectral changes upon UDP-2FGal binding  When Mg2+•UDP-2FGal was added, the signals of Ile3, Ile76, Ile79, Ile81, Ile93, Ile104, Ile159, Ile169 and Ile191 were perturbed (Figure 3.27A & C, & Figure 4.1B). Also, the intensities of some “a” peaks (i.e. Ile31 and Ile129) were decreased and those of the “b” peaks were increased, indicating the LgtC conformational equilibrium is shifting from predominantly state “a” to state “b” upon substrate binding (Figure 3.27F). As expected, when mapped onto the structure of LgtC, isoleucines showing the largest chemical shift perturbations generally clustered near the Mg2+•UDP-2FGal binding site (Figure 4.8A). However, Ile31 and Ile129 are distal from the active site and thus the associated conformational change is relatively global. In the case of Ile191, which is also distal from the active site, the chemical shift perturbations upon substrate/product binding might be due to the ring current effects from three adjacent aromatic residues (Figure 4.8B & C). Additionally, the double signals of Ile104 became one upon UDP2FGal binding, whereas the single signal of Ile79 split into two small signals, indicating that the conformational changes might be more complex than simply 2 states.  170  Figure 4.8: CSP of isoleucine δ1-methyl of uniformly deuterated and [1H/13C] selectively methyl labeled LgtC upon UDP-2FGal binding. (A) Isoleucine δ1-methyl chemical shift perturbations for LgtC with 1 mM UDP-2FGal (10 mM Mg2+) versus apo LgtC. Isoleucines with a CSP > 0.08 ppm are highlighted in red and the remainder with smaller perturbations in cyan on the shown structure. Blank values correspond to residues not detected under one or more conditions. The two flexible loops are colored in green. UDP-2FGal and 4’-deoxylactose are also shown (carbon, grey; oxygen, red; nitrogen, blue; phosphorus, orange; Mn2+, magenta). (B) Ile191 is within the same α-helix as Gln189 and Asp190. The N-terminal end of the helix contacts the substrate binding site of LgtC. (C) Ile191 is surrounded by aromatic residues and thus its chemical shift may change readily upon structural changes propagated from the substrate binding site via this helix.  171  4.3.1.4.2  Spectral changes upon lactose binding  Relatively small spectral changes occurred when lactose was added into the binary complex of LgtC•Mg2+•UDP-2FGal, indicating that acceptor binding does not substantially perturb the structure of the enzyme (Figure 4.9). This is consistent with a comparison of the crystal structures of its binary and ternary substrate complexes (Figure 1.7), and the lack of any isoleucine near the acceptor binding site (Figure 4.9). Surprisingly the number of peaks was decreased from 25 to 20 (Figure 3.27C & D). In particular, the signals of Ile76, Ile79, and Ile104 disappeared. These three residues contact the bound UDP-2FGal, indicating that the presence of a sugar acceptor does perturb the donor and/or residues forming the donor binding site. However, exchange broadening due to lactose binding is unexpected given its very low affinity for the LgtC binary complex (Km = 13 mM) (Persson, Ly et al. 2001), and suggests that its association/dissociation rates are slowed to the msec-μsec time-scale due to a requisite structural or dynamic change.  Figure 4.9: CSP of isoleucine δ1-methyl of uniformly deuterated and [1H/13C] selectively methyl labeled LgtC binary complex upon lactose binding. Isoleucine δ1-methyl chemical shift perturbations for LgtC ternary complex with 1 mM UDP2FGal and 300 mM lactose (10 mM Mg2+) versus LgtC binary complex with 1 mM UDP-2FGal (10 mM Mg2+). Isoleucines with a CSP > 0.08 ppm are highlighted in red and the remainder with smaller perturbations in cyan on the shown structure. Blank values correspond to residues not detected under one or more conditions. The two flexible loops are colored in green. UDP-2FGal and 4’-deoxylactose are also shown (carbon, grey; oxygen, red; nitrogen, blue; phosphorus, orange; Mn2+, magenta).  172  4.3.1.4.3  Spectral changes upon UDP binding  When Mg2+•UDP was added, modest isoleucine chemical shift perturbations resulted (Figure 4.10). Specifically, the signals of Ile93, Ile104 and Ile191 were shifted and the relative intensities of two “a” peaks (i.e. Ile31 and Ile129) decreased and those of the “b” peaks increased, indicating the LgtC was stabilized similarly in state “b” (i.e. state “d”). It is noteworthy that the chemical shift perturbations due to UDP binding are substantially smaller than those due to UDP-2FGal binding, despite the fact that several isoleucines are expected to contact the UDP moiety in both cases. This phenomenon is consistent with the weaker binding affinity of UDP (Kd = 80 μM) than that of UDP-2FGal (Km = 2 μM). Although a crystal structure of the LgtC product complex with Mg2+•UDP has not been solved, this suggests that conformational differences may exist relative to the substrate complex.  Figure 4.10: CSP of isoleucine δ1-methyl of uniformly deuterated and [1H/13C] selectively methyl labeled LgtC upon UDP binding. Isoleucine δ1-methyl chemical shift perturbations for LgtC with 1 mM UDP (10 mM Mg 2+) versus apo LgtC. Isoleucines with a CSP > 0.08 ppm are highlighted in red and the remainder with smaller perturbations in cyan on the shown structure. Blank values correspond to residues not detected under one or more conditions. The two flexible loops are colored in green. UDP2FGal and 4’-deoxylactose are also shown (carbon, grey; oxygen, red; nitrogen, blue; phosphorus, orange; Mn2+, magenta). Interestingly, the signals of Ile76 and Ile79, which are within one of the flexible loops, disappeared upon Mg2+•UDP binding. This is suggestive of conformational exchange broadening (i.e. interconversion between states with different chemical shifts at a rate kex ~ Δω). To exclude  173  alternative possibilities, such as aggregation, UDP was depleted “in situ” using pyruvate kinase (PK), which transfers the phosphate from phosphoenolpyruvate (PEP) to UDP and forms UTP and pyruvate (Figure 4.11). Indeed, the relative intensities and chemical shifts of Ile31a, Ile76a, Ile79, Ile129a, and Ile191 returned to that observed with apo LgtC (Figure 4.11B) as UTP no longer binds the enzyme. The signal of Ile76b could not be recovered because the original signal was weak and some aggregation was observed physically after the spectral acquisition.  Figure 4.11: Depletion of UDP from LgtC•Mg2+•UDP binary complex by PK. Overlaid methyl-TROSY spectra of apo LgtC (100 μM) (red) and its binary complex saturated with 20 mM Mg2+ and 1 mM UDP (purple) before (A) and after (B) the addition of 5 mM PEP and 10 units of pyruvate kinase (PK) (blue). As judged for example by the signals of Ile31a, Ile129a, and Ile191, LgtC returned to its apo state after UDP was converted to non-binding UTP. The disappearance of Ile76b and reduced intensities of some peaks might be due to incomplete depletion of UDP.  174  4.3.1.5 Substrate binding to 15N-tyrosine-labeled LgtC Substrate binding was also investigated using 1H/15N-TROSY-HSQC spectra of  15  N-  labeled LgtC (Figure 3.7). However, interpretation of the many spectral changes was not possible due to the complexity of the spectrum of the uniformly labeled protein, combined with the absence of ~ 1/3 of the expected signals and incomplete assignment of observed signals. Therefore, we will focus our attention on the spectra of LgtC selectively labeled with  15  N-  tyrosine (Figure 4.12).  175  Figure 4.12: Substrate binding monitored by  1  H/15N-TROSY-HSQC spectra of  15  N-  tyrosine-labeled LgtC. 1  H/15N-TROSY-HSQC spectra of LgtC selectively labeled with 15N-tyrosine in its (A) apo state, (B) binary state with 1 mM UDP-2FGal and 1 mM MgCl2, and (C) ternary state also with 100 mM 4’-deoxylactose. Also shown are superimposed spectra of (D) the binary complex and apo enzyme, (E) the ternary complex and the apo enzyme, and (F) all three species. The pairs that visibly changed in their relative intensities upon substrate binding are boxed in (A). All samples were in the buffer consisting of 20 mM Tris, 5 mM TCEP, pH 8.5 with 10% D2O. The spectra were acquired for 16 hrs each at 25 °C with a 600 MHz NMR spectrometer. Aliased signals in (B) and (C) are colored in cyan and yellow, respectively. LgtC contains 15 tyrosine residues and indeed 15 signals are observed in the spectrum of the labeled protein. However, this might be coincidental because signals from the active site amides are generally absent in 1H/15N-TROSY-HSQC spectra of this protein, thus some peaks could arise from multiple conformations of LgtC. Indeed, although the 1H/15N-TROSY-HSQC spectrum of  15  N-tyrosine labeled LgtC was not completely assigned, Tyr101, Tyr150, and 176  Tyr202 appear to show a pattern of “a” and “b” peaks similar to that observed in methyl-TROSY spectra (Figure 4.12 & Figure 3.27). Furthermore, when the substrates were present, the relative intensities of “a” peaks decreased and the “b” peaks increased. Residues Tyr150 and Tyr202 are distal to the active site of LgtC, and thus substrate binding appears to perturb the global conformation of the enzyme (Figure 4.13). In the spectra of the binary (Mg2+•UDP-2FGal) and ternary (Mg2+•UDP-2FGal•4’deoxylactose) complexes, 23 and 19 peaks were observed, respectively (Figure 4.12B & C). This increase in signals may reflect ordering of the enzyme to reduce otherwise exchange broadened tyrosines and chemical shift perturbations due to the bound substrate analogs, combined with possible residual signals from incomplete saturation of LgtC. The intensity of the proposed “b” peak of Tyr101 also shifted slightly (Figure 4.12), presumably because it interacts with the UDP moiety of UDP-2FGal (Figure 4.13A). Finally, it is noteworthy that the signal of Tyr186 changed upon addition of UDP-2FGal and then changed further with lactose binding. This is consistent with its location within the acceptor binding site of LgtC.  Figure 4.13: Locations of tyrosine residues in the crystal structure of the LgtC ternary complex. (A) “Top” and (B) “side” views of LgtC highlighting the tyrosine residues Tyr101, Tyr150, Tyr186, and Tyr202, which are colored in yellow. The remaining tyrosine residues are colored in cyan, along with Gln189 in blue and Asp190 in red. The two flexible loops are colored in green. UDP-2FGal and 4’-deoxylactose are also shown (carbon, grey; oxygen, red; nitrogen, blue; phosphorus, orange; Mn2+, magenta).  177  4.3.1.6 Substrate binding investigated using UAA Unnatural amino acids (UAAs) can be site-specifically incorporated into proteins using engineered “suppressor” tRNA/aminoacyl-tRNA synthetases. These non-canonical amino acids serve as unique spectroscopic and chemical probes for studying protein structure, dynamics, and function (Jackson, Hammill et al. 2007; Graziano, Liu et al. 2008; Jones, Cellitti et al. 2010; Liu and Schultz 2010). Using published methods, Adrienne Cheung in the McIntosh laboratory substituted the NMR active UAAs, CF3Phe and O13CH3Phe, for Phe132 and Tyr186 in LgtC (Figure 4.14). Kinetic analyses of the UAA-labeled LgtC were done and the incorporations of UAA at positions 132 and 186 did not affect the enzymatic activity (Table 4.1).  Figure 4.14: Isotopically labeled UAA incorporated at positions 132 and 186 of LgtC. (A) Phe132 and Tyr186 are highlighted in red in the LgtC ternary crystal structure. The two flexible loops are in green, and UDP-2FGal and 4’-deoxylactose are also shown (carbon, grey; oxygen, red; nitrogen, blue; phosphorus, orange; Mn2+, magenta). (B) The UAAs CF3Phe and O13CH3Phe.  178  Table 4.1: Steady state kinetic parameters for UAA-labeled LgtC. b  Km (μM)  kcat (s-1)  WT LgtC  19  46  CF3Phe-labeled LgtC-F132X  14  O13CH3Phe-labeled LgtC-Y186X  10  a  a  a  c  UDP-Gal kcat / Km  Lactose kcat / Km  Km (mM)  kcat (s-1)  2.5  -  -  -  33  2.4  29  35  1.2  29  3.1  14  40  2.8  -1 -1  (μM s )  (mM-1 s-1)  a  Data were provided by Adrienne Cheung.  b  Data for sugar donor UDP-Gal in the presence of 160 mM acceptor lactose at pH 7.5 and 37 °C.  c  Data for sugar acceptor lactose in the presence of 10 mM donor UDP-Gal at pH 7.5 and 37 °C.  4.3.1.6.1  Substrate binding of CF3Phe-labeled LgtC-F132X  The  19  F-NMR spectra of CF3Phe-labeled LgtC-F132X are shown in Figure 4.15. In the  spectrum of the apo protein, a major peak “a” and a minor peak “b” were observed (Figure 4.15A). This phenomenon was also observed in the spectra of CF3Phe-labeled LgtC-Y186X and LgtC-A249X (not shown). This supports further the conclusion that LgtC exists in a conformational equilibrium between at least 2 states. When Mg 2+•UDP-2FGal was added, the relative intensities of the peaks switched, suggesting that peak “b” arises from a conformation resembling that of the donor-bound state. Note that position 132 is not immediately adjacent to the bound Mg2+•UDP-2FGal and thus CF3Phe132 may be reporting a conformational change rather than a direct perturbation from the bound donor. Addition of 100 mM lactose caused a small shift of peak "b", possibly due to a direct interaction of residue 132 and the bound acceptor (Figure 4.15). Similar patterns of the chemical shift perturbations of peaks upon addition of substrates were also observed in the spectra of CF3Phe-labeled LgtC-Y186X and LgtC-A249X (not shown). After ~ 15 hrs, the UDP-2FGal was hydrolyzed and a broad signal with the chemical shift of peak "a" was observed. This presumably arises from the product complex with bound UDP, and CF3Phe132 is in the same environment as with apo LgtC. Thus the  19  F-NMR  spectra of CF3Phe-labeled LgtC-F132X support the conclusions drawn above from methylTROSY studies.  179  Figure 4.15: Substrate binding of CF3Phe-labeled LgtC-F132X monitored by  19  F NMR  spectroscopy. 19  F NMR spectra of CF3Phe-labeled (A) apo LgtC-F132X, (B) its binary complex with 10 mM Mg2+ and 1 mM UDP-2FGal, (C) its ternary complex with 10 mM Mg2+, 1 mM UDP-2FGal, and 100 mM lactose, and (D) the ternary complex after 15 hours. The samples were in buffer consisting of 20 mM Tris, 5 mM TCEP, pH 8.5 with 10% D 2O. The spectra were acquired at 25 °C using a 500 MHz spectrometer.  4.3.1.6.2  Substrate binding of O13CH3Phe-labeled LgtC-Y186X  In the crystal structure of the LgtC ternary complex, Tyr186 is at the N-terminus of a helix that helps to form the binding site for the galactosyl moiety of UDP-2FGal, and also interacts with 4’-deoxylactose (Figure 4.14A). Therefore, it was chosen to be replaced by UAA O13CH3Phe to monitor the changes due to substrate binding via 1H/13C-HSQC spectra (Figure 4.16).  180  Figure 4.16: Substrate binding of O13CH3Phe-labeled LgtC-Y186X monitored by 1H/13CHSQC NMR spectroscopy. 1  H/13C-HSQC NMR spectra of O13CH3Phe-labeled (A) apo LgtC-Y186X (red), (B) its binary complex with Mg2+•UDP-galactose (blue), (C) its ternary complex with Mg2+•UDPgalactose•lactose (green), and (D) its ternary complex with addition of 5 mM PEP and 10 units of PK (green). The superimposed spectrum of apo LgtC is included in B-D to illustrate the chemical shift perturbations. (E) Expanded overlaid regions of the 1H/15N-TROSY-HSQC spectra of apo LgtC selectively labeled with 15N-tyrosine (red), its binary complex with Mg2+•UDP-2FGal (blue), and its ternary complex with Mg2+•UDP-2FGal•4’-deoxylactose (green) (taken from Figure 4.12). The spectrum of the ternary complex acquired when UDP2FGal had been hydrolyzed completely after 60 hrs is colored in magenta. The samples were at 25 °C in 20 mM Tris, 5 mM TCEP, pH 8.5 with 10% D2O. The 1H/13C-HSQC and 1H/15NTROSY-HSQC spectra were acquired using 850 MHz and 600 MHz spectrometers, respectively. In the 1H/13C-HSQC spectrum of apo LgtC-Y186X, a single peak from O13CH3Phe186 with a tiny shoulder was observed (Figure 4.16A). When UDP-Gal was added, the main peak was shifted to the position of the shoulder and a residual signal from apo enzyme remained. (Note that UDP-Gal was used due to the lack of UDP-2FGal and the realization that the unmodified substrate is also only slowly hydrolyzed and thus suitable for NMR studies.) A similar pattern change for 2 peaks from Tyr186 in the 1H/15N-TROSY-HSQC of selectively 15N181  tyrosine labeled LgtC was also observed (Figure 4.12 & Figure 4.16E). Again, this is indicative of “a” and “b” conformational states for the apo enzyme, with “b” resembling the substrate bound state. When lactose was added into the binary complex, the signals remained unchanged in the 1H/13C-HSQC spectrum, even after 24 hours (Figure 4.16C). However, when PK and PEP were added to deplete the UDP in the sample, the peak shifted back to the original position of the apo form (Figure 4.16D). Thus, it appears that O13CH3Phe186 has the same chemical shift when either UDP-Gal and lactose or UDP are bound (or hydrolysis occurred before a signal from the ternary complex could be detected). In contrast, the 1H/15N-TROSY-HSQC of Tyr186 does report the different substrate/product bound states of LgtC.  4.3.2 Multiple-conformational states of LgtC A consistent observation from the NMR spectra of the amides and methyls of LgtC, as well as from incorporated UAAs, is that many residues in the apo enzyme give two or more signals, denoted as "a" and "b". Although, as discussed in section 3.3.4, this could be due to sample heterogeneity, such as degraded or chemically modified forms of LgtC, we disfavor this possibility as the protein appears sufficiently pure and the multiple signals appear consistently in all samples of the wild-type and mutant enzymes. Furthermore, substrate binding shifts the relative ratios of the “a” and “b” peaks. Thus, we hypothesize that apo LgtC adopts at least 2 conformations, with “a” peaks reflecting a substrate-free (or “opened”) state and “b” peaks reflecting a conformation resembling that adopted upon substrate binding (“closed”). The presence of "a" peaks in spectra of the substrate bound enzyme could arise from incomplete enzyme saturation or from an equilibrium between "opened" and "closed" states even with substrate bound. Since the “a” peaks did not disappear, even with 2 mM UDP-2FGal, the latter hypothesis is possible. To further investigate these hypotheses, we investigated the temperature dependence of the LgtC spectra, studied by the observation of magnetization exchange between the "a", and "b" peaks, and used NMR relaxation dispersion measurements for more quantitative insights into LgtC dynamics.  182  4.3.2.1 Temperature effect on the methyl-TROSY of apo LgtC The methyl-TROSY spectra of apo LgtC were recorded as a function of temperature for two reasons (Figure 4.17). First, we wanted to find the optimal temperature for subsequent studies. This requires finding a balance between increased signal lifetimes and hence sharper peaks due to faster global tumbling with elevated temperature versus any detrimental effects such as aggregation, unfolding, or chemical modification. However, more importantly, we wanted to determine whether the additional signals changed as a function of temperature, perhaps due to a conformational equilibrium of LgtC.  183  184  Figure 4.17: Methyl-TROSY spectra of uniformly deuterated and [1H/13C] selectively methyl labeled LgtC as a function of temperature. Methyl-TROSY spectra of apo LgtC were acquired at (A) 16 °C, (B) 20 °C, (C) 25 °C, (D) 30 °C, and (E) 37 °C in random order. For comparison, the contour levels of the spectra were adjusted by eye to compensate for temperature-dependent peak intensities and aggregation of the protein. (F) The overlaid spectra shows some peaks changed progressively in chemical shift and relative intensity when the temperature was increased. The reference for each spectrum was adjusted to correct for the temperature-dependent lock signal by referencing the chemical shift of H2O signals at different temperatures. Overall, the methyl-TROSY spectra of LgtC were similar from 16 °C to 37 °C, indicating that the protein remains well folded over this temperature range. As expected, signals were broader at the lower temperature. However, LgtC aggregated during spectral acquisition at high temperatures (i.e. 30 °C and 37 °C), thus the peaks in the spectrum acquired at 30 °C look sharper and more intense compared with that acquired at 37 °C (Figure 4.17D & E). Based on these data, we chose 25 °C as the optimal temperature for subsequent studies. Most interestingly, the Ia/Ib ratio (i.e. relative intensities of the “a” versus “b” peaks) of the assigned residues, such as Ile31, Leu74, Ile76, Ile104, Leu125, Ile129, Val133, Ile143, and Val200, increased slightly with increasing temperature (Figure 4.17). This is consistent with the hypothesized conformation equilibrium of apo LgtC and indicates that the “opened” state is favored at higher temperatures. Notably, some of these residues, such as Ile31, Ile125, Ile129, Ile143, and Val200, are distal from the active site of LgtC and help to form the “scaffold” of the enzyme.  4.3.2.2 Magnetization transfer between conformational states 13  C and  19  F magnetization transfer experiments were used to detect slow exchange  between the two conformational states of LgtC (Velyvis, Schachman et al. 2009). As shown in Figure 4.18A & B, weak exchange cross peaks were observed for Ile31, Ile129, and an unassigned leucine/valine in 1H/13C-HMQC-detected exchange spectra. Other residues exhibiting “a” and “b” signals did not show clear exchange peaks either due to poor spectral dispersion or poor signal intensities after a long transfer delay. Similar weak magnetization transfer was observed in the 19F/19F-NOESY spectrum of CF3Phe-labeled LgtC-F132X (Figure 4.18C). These observations confirm that apo LgtC exists as an equilibrium mixture of at least two 185  interconverting conformational states that give rise to “a” and “b” peaks of methyl groups and an unnatural amino acid. Unfortunately, due to the long times (48 hrs) required to obtain these spectra, it was not practical to repeat the measurements as a function of transfer time and thereby measure the exchange rate constants. Accordingly, we can only conclude that the detected conformational change of LgtC occurs on the seconds time-scale.  186  187  Figure 4.18: Magnetization exchange experiments reveal the slow interconversion between the two conformational states of apo LgtC. Exchange cross peaks were detected for (A) Ile31 and Ile129, and (B) a pair of unassigned leucines/valines. Panels (i) show control methyl-TROSY spectra and panels (ii) show 1H/13CHMQC-detected exchange spectra with a 400 ms transfer delay. (C) Exchange was also detected for CF3Phe-labeled LgtC-F132X using a 19F/19F-NOESY experiment with (i) 1 msec or (ii) 250 msec transfer times. The additional diagonal signal was not present in the initial spectra recorded with fresh protein and is attributed to sample degradation. The exchange rates (kex) of (D) Ile31 and (E) Ile129 were calculated by fitting into the curves using Prism (Hwang, Choy et al. 2002). Based on the numbers, LgtC consists of more than two conformational states. Exchange measurements were also carried out for the binary complex of LgtC with Mg2+•UDP-2FGal. In contrast to the apo enzyme, no exchange cross peaks were detected for isoleucines with “a” and “b” peaks in the binary complex (400 msec transfer time; not shown). Note that the signals from these residues are generally better dispersed than with apo LgtC, but the "a" signals are weak relative to the "b" signals. Thus the lack of observable transfer might be due to the low signal intensities of "a" peaks or to exchange occurring on a time-scale slower than the 13C transverse relaxation times of the labeled methyl groups (i.e. seconds).  4.3.2.3 Summary of conformational exchange of LgtC LgtC exists in an equilibrium between at least two conformations that interconvert with an exchange rate constant of ~ 1 sec -1, as estimated from magnetization transfer experiments. Increasing temperature favors the state giving rise to "a" signals, whereas substrate binding favors the state with "b" signals. Accordingly, we attribute "a" signals to an "opened" state and "b" signals to a "closed" state of the apo enzyme that resembles its substrate bound state.  4.3.3 Studies of LgtC structural dynamics 4.3.3.1 Methyl relaxation dispersion of 13C selectively methyl labeled LgtC Relaxation dispersion experiments were used to further investigate the dynamics of LgtC. These experiments are a robust approach for detecting the contributions of msec-μsec time-scale conformational exchange to the effective decay of the transverse  13  C signal. TROSY-based  188  methyl relaxation dispersion measurements of deuterated LgtC selectively labeled with Ileδ1[1H/13C], Leu-[13CH3,12CD3], and Val-[13CH3,12CD3] recorded with 500 MHz, 600 MHz, and 850 MHz NMR spectrometers are shown in Figure 4.19. Conformational exchange is manifest as a decrease in the effective relaxation rate, R 2,eff, with increasing frequency of refocusing pulses (νCPMG). The effect also increases with increasing magnetic field strength. Since LgtC is a relatively large protein with relatively fast transverse relaxation decay rates, dispersion experiments were also carried out using a shorter 20-ms constant time delay (Figure 4.20). This led to improved data for several residues including Ile76 and Ile104, although precluded detection of slower motions accessible with a longer delay.  Figure 4.19: Methyl relaxation dispersion curves of apo LgtC. Selected relaxation dispersion curves of (A) Ile79, (B) Ile3a, (C) Ile191, and (D) Ile59 of apo LgtC. The relaxation dispersions were done with T2 constant time delay of 40 ms at 25 °C using 500 MHz (green), 600 MHz (blue) and 850 MHz (red) NMR spectrometers. (E) Assigned methyl-TROSY spectrum of apo LgtC showing the isoleucine δ1-methyl signals Note that the signal-to-noise ratio of data recorded with a conventional probe on the 850 MHz instrument was lower than that with cryo-probes on the lower field spectrometers.  189  Figure 4.20: Methyl relaxation dispersion curves of apo LgtC with different delay times. Selected relaxation dispersion curves of (A) Ile104, (B) Ile3a, (C) Ile191, and (D) Ile59 of apo LgtC recorded with constant time delays of 20 ms (green) and 40 ms (blue) at 25 °C using a 600 MHz spectrometer. (E) Assigned methyl-TROSY spectrum of apo LgtC showing the isoleucine δ1-methyl signals. As a preliminary analysis step, Rex values were calculated as R2,eff(25 Hz) – R2,eff(1000 Hz) and plotted as histogram in Figure 4.21 for constant time delay of 40 ms. Rex values with constant time delay of 20 ms were calculated from fitting the dispersion curves using (Kleckner and Foster 2012) and plotted as histogram in Figure 4.22. The data were also mapped onto the structure of the enzyme. The observation of ΔRex terms for many methyls in both the “a” and “b” states clearly reveals that LgtC also undergoes msec-μsec time-scale motions. In particular, the active site residues Ile104 and Val133 showed large dispersion in the apo LgtC. Two residues in the same helix as Gln189 and Asp190, Ile191 and Leu192 showed intermediate effects as well. However, Ile3, Ile31, Val22, and Leu125 also showed R ex terms, indicating that these motions extend throughout the protein scaffold and are not limited to the active site. Interestingly, the loop residues, Ile76, Ile79, and Ile81, did not show dispersion behaviour and thus did not undergo significant motions on the time-scale probed by this approach.  190  191  Figure 4.21: Methyl relaxation dispersion studies of apo LgtC (40 ms delay). Plots of the exchange contributions (Rex) towards the 13C relaxation of assigned (A) isoleucine δ1-methyls and (B) leucine δ-methyls and valine γ-methyls of LgtC are shown. The data were obtained using the 500 MHz (green), 600 MHz (blue) and 850 MHz (red) NMR spectrometers with constant time delay of 40 ms. The light green, cyan, and pink dashed lines indicate values of ΔRex deemed to be significant visually in the 500 MHz, 600 MHz and 800 MHz data sets, respectively. The few negative values for the latter data set reflect errors due to low signal-tonoise ratios. (C) ΔRex values in the 500 MHz data are mapped on the ribbon diagram of LgtC (PDB code: 1GA8 with substrate analogs not shown; flexible loops colored in green). Only the active site residues and the residues with R ex >10 s-1 are labeled. The color code for leucine and valine residues is based on the largest Rex value of either methyl group.  192  Figure 4.22: Methyl relaxation dispersion studies of apo LgtC (20 ms delay). (A) Exchange contributions (Rex) of isoleucine δ1-methyl groups of LgtC. The data were obtained from using the 600 MHz (blue) and 850 MHz (red) NMR spectrometers with a constant time delay of 20 ms. Data was calculated by fitting the dispersion curves using GUARDD (Kleckner and Foster 2012). The pink and cyan dashed-lines indicate values of ΔRex deemed to be significant in the two data sets, respectively. (B) R ex value from the 600 MHz data set are mapped on the ribbon diagram of LgtC (PDB code: 1GA8 with substrate analogs not shown; flexible loops colored in green).  193  Figure 4.23: Comparison of methyl relaxation data for apo LgtC and its binary complex. Exchange contributions (Rex) for the isoleucine δ1-methyls of apo LgtC (red) and its binary complex with Mg2+•UDP-2FGal (blue), obtained using a 600 MHz spectrometer with a constant time delay of 20 ms. Blanks indicate the absence of a signal in the spectra of apo LgtC or its binary complex. The dashed-line indicates a cut-off value for results discussed in the text. To investigate possible dynamic changes of LgtC upon substrate binding, methyl relaxation dispersion experiment was also carried out for its binary complex with Mg 2+•UDP2FGal (20 ms delay at 25 °C using a 600 MHz spectrometer). As summarized in Figure 4.23, the Rex values of most residues, except Ile76a, generally decreased upon complex formation. The substrate-bound LgtC generally retains similar motions as in the apo form on the msec-μsec time-scale. The apparent increase of the Rex value of Ile79 and Ile104 suggests that these residues are still dynamic in the presence of sugar donor and might be involved in the stabilization of the oxocarbenium transition state during catalysis in the S Ni mechanism. The importance of these two residues in catalysis is evidenced by the reduced k cat and kcat/Km values of LgtC-I79A and LgtC-I104A mutants relative to the wild-type (Table 3.3). 194  4.3.4 Intermediate trapping of LgtC-Q189E 4.3.4.1 NMR spectroscopy In the presence of UDP-Gal, LgtC-Q189E formed a covalent glycosyl-enzyme (CGE) intermediate (LgtC-Q189E-Gal) in which Asp190, instead of the proposed nucleophile Glu189, was found to be covalently linked to galactose (Lairson, Chiu et al. 2004). This very surprising result, obtained from mass spectrometry, implies that the mutant LgtC must undergo a significant conformational change during catalysis in order to position the side-chain of Asp190 near the activated donor sugar. However, such a conformational change seems unlikely since crystal structures of wild-type LgtC and LgtC-Q189E with Mn2+•UDP-2FGal bound are essentially identical (Lairson, Chiu et al. 2004). To investigate this perplexing result further, a similar CGE intermediate trapping experiment was undertaken using NMR spectroscopy.  Figure 4.24: Overlaid 1H/15N-TROSY HSQC spectra of  15  N-labeled wild-type LgtC and  LgtC-Q189E. Overlaid spectra of 15N-labeled wild-type LgtC (red, green) and LgtC-Q189E (blue, yellow). The aliased peaks (green, yellow) result from the reduced 15N spectral-width. The 1H/15N-TROSY-HSQC spectrum of 15N-labeled LgtC-Q189E is quite different from that of the wild-type LgtC (Figure 4.24). Many peaks were shifted as a consequence of the single mutation, thus precluding confident spectral assignments (Chapter 3). This also indicates that the introduction of a presumably charged carboxylate side-chain in the active site of LgtC directly 195  (via electric field effects) or indirectly (via conformational change) perturbs the environments of amides throughout the protein. The putative CGE intermediate LgtC-Q189E-Gal was then formed by incubation with Mg2+•UDP-Gal in the presence of coupling enzymes to deplete any product UDP, and its NMR spectrum was acquired (Figure 4.25A). Unfortunately, the LgtCQ189E-Gal aggregated relatively quickly in the presence of Mg2+, thus degrading the quality of the spectra. Nevertheless, as seen in Figure 4.25A, significant spectral changes were observed relative to LgtC-Q189E. However, this might be simply due to UDP-galactose binding and not necessarily to the formation of a CGE intermediate. In support of this hypothesis, the 1H/15NTROSY-HSQC spectrum of  15  N-labeled LgtC-Q189E complexed with Mg2+•UDP-2FGal was  recorded and shown to resemble that of LgtC-Q189E-Gal though differences remain (Figure 4.25B). Both spectra differ significantly from the spectrum of LgtC-Q189E. This indicates that much of the same general conformational change happens when covalently linked galactose or UDP-2FGal occupies the active site. Indeed, the differences between the spectra of LgtC-Q189EGal and LgtC-Q189E•Mg2+•UDP-2FGal may simply be due to the presence or absence of the UDP moiety (Figure 4.25C).  196  Figure 4.25: Overlaid 1H/15N-TROSY HSQC spectra of LgtC-Q189E. Overlaid 1H/15N-TROSY HSQC spectra of LgtC-Q189E (red, aliased peaks in green), LgtCQ189E-Gal (blue, aliased in yellow), and LgtC-Q189E complexed with Mg2+•UDP-2FGal (green). The bottom 3 aliased peaks (yellow) in the spectrum of LgtC-Q189E-Gal correspond to the top 3 green peaks in the spectrum of LgtC-Q189E complexed with Mg2+•UDP-2FGal. All samples were 15N-labeled. The LgtC-Q189E-Gal sample was prepared by incubation with Mg2+•UDP-Gal in the presence of coupling enzymes to deplete UDP and recorded after 15 minutes of incubation.  197  4.3.4.2 ESI-MS ESI-MS was used to re-examine the hypothesis that galactose was covalently linked to LgtC-Q189E in this sample. Interestingly, two species of similar populations were observed in the MS spectrum of  15  N-labeled apo LgtC-Q189E. The theoretical mass of this protein is  33567.4 Da, which is close to that of the observed peak at 33561 Da (Figure 4.26A). Hence, the second peak of 33723 Da is potentially  15  N-labeled LgtC-Q189E-Gal or  15  N-labeled LgtC-  Q189E-Glc (Δm = 162 Da versus an expected increase of 157 Da). Indeed, upon incubation with UDP-Gal, only the higher mass species was observed. Thus, it is possible that the apo LgtCQ189E was partially modified due to UDP-Glc or UDP-Gal present in the expression host E. coli. However, wild-type LgtC (Δm ~ 132 Da) and LgtC-A249X (Δm ~ 210 Da) also showed small populations of higher mass forms with the lower accuracy MALDI-TOF MS approach (Figure 3.16), and these species would not be expected to form any covalent intermediate. Furthermore, the putative covalently-bound galactose or glucose in LgtC-Q189E could not be removed by addition of UDP, Mn2+, and lactose to complete the glycosyl transfer reaction. Therefore, the origin of the higher mass forms of LgtC could not be unambiguously determined.  198  Figure 4.26: ESI-MS spectra of  15  N-labeled LgtC-Q189E and its covalent glycosyl-enzyme  intermediate. ESI-MS spectra of (A) 15N-labeled LgtC-Q189E and (B) 15N-labeled LgtC-Q189E-Gal. The mass difference is 162 Da between the two populations of apo 15N-labeled LgtC-Q189E is suggestive of covalently-linked galactose.  199  4.4 DISCUSSION AND CONCLUSION 4.4.1 Substrate and product binding of LgtC Binding of UDP-Gal and lactose by LgtC occurs in slow-exchange on the chemical shift time-scale (i.e. msec), as demonstrated through titration studies presented in Chapter 3. In this chapter, the substrate binding modes of LgtC were characterized further by NMR spectroscopy. Using both methyl-TROSY and 1H/15N-TROSY-HSQC measurements, Mg2+ and UDP-2FGal were found to bind LgtC simultaneously (Figure 4.1). Although metal ions are known to be mandatory for the enzymatic activities of many glycosyltransferases and the interaction of Mg 2+ and nucleotides had been studied (Powell and Brew 1976; Tran-Dinh and Neumann 1977; Vogel and Bridger 1982), this is the first direct demonstration of the requisite simultaneous binding of a metal ion and sugar donor analog with LgtC. The positive charge of the divalent metal ion might function to neutralize the negative charges of UDP-Gal in order to facilitate its binding to the active site of the enzyme, as well as enhance glycosyl-transfer. Since a metal ion and sugar donor must bind LgtC simultaneously, a single Km value for formation of a given LgtC•metal•sugar-donor complex would be expected. Previously reported data for LgtC (metal ions: Mg2+ (Km = 370 μM with UDP-Gal) and Mn2+ (Km = 27 μM with UDP-Gal); sugar donor: UDP-Gal (Km = 30 μM with Mn2+) and UDP-2FGal (Ki = 2 μM with Mn2+)) were determined by enzymatic assays in which the Km value of one species was measured in the presence of saturating levels of all other substrates (Lougheed 1998; Persson, Ly et al. 2001). As expected, the Km values of Mn2+ and UDP-Gal are indeed similar, thus the Km of the Mn2+•UDP-Gal complex is 30 μM. Although the Km of Mn2+ was not determined with saturating UDP-2FGal, the Km of the Mn2+•UDP-2FGal complex should be 2 μM. The Km value for Mg2+ is ~10-fold higher than for Mn2+, and the corresponding Km value for UDP-Gal in the presence of saturating Mg2+ was not measured. The substitution of Mg2+ for Mn2+ also reduces the activity of LgtC by ~50% (Lougheed 1998). 31  P and 19F NMR spectroscopies were also used to investigate the binding of the substrate  analog UDP-2FGal and the product UDP (in the presence of Mg2+). In each case, binding occurred in the slow exchange limit. Although both species contain two non-equivalent phosphates, only one single broad  31  P-NMR signal was detected for bound UDP (Figure 4.2),  whereas two very dispersed broad signals were observed with UDP-2FGal (Figure 4.3). This indicates that the phosphates of Mg2+•UDP-2FGal are in distinct environments when bound to 200  LgtC. Indeed, X-ray crystallographic analyses revealed that UDP-2FGal is deeply buried within the LgtC active site and adopts an unusual conformation with the galactose ring folded back against the phosphates (Persson, Ly et al. 2001). Each phosphate also participates in distinct hydrogen bonding and electrostatic interactions. In contrast, the  31  P NMR signals of the bound  product Mg2+•UDP are less perturbed relative to those of free UDP and hence may be less structurally constrained within the active site of LgtC due to the lack of the galactosyl moiety. This is also consistent with their relative Kd values.  31  P and  19  F NMR spectroscopies also  demonstrated that UDP-2FGal is slowly hydrolyzed by LgtC (Figure 4.4 & Figure 4.5). STD-NMR was used to probe the binding epitopes of UDP with LgtC (Figure 4.6). From the STD factors, the binding of UDP mainly involves the uridine ring, suggesting that the enzyme is in a more opened state than observed crystallographically with UDP-2FGal. That is, if the enzyme were in a “closed” state with UDP completely buried within the active site, then all protons in UDP should show the same STD effect (Figure 4.6C). However, the STD effects of the ribose protons of UDP are only ~50% of those for the uridine moiety. On the other hand, CSPs of the flexible loop residues (i.e. Ile76, Ile79, and Ile81) from the methyl-TROSY spectra of LgtC•Mg2+•UDP complex are relatively low (Figure 4.10), indicating that the motion of the loop is similar to that in the apo-form. These observations are reasonable because UDP is one of the products of galactosyl-transfer reaction of LgtC and it has to be released at the end of the reaction. Unfortunately, no STD effect could be observed from UDP-2FGal because of its tight binding (Ki = 2 μM).  4.4.2 Multi-conformational states of LgtC More than the number of expected peaks was observed in the methyl-TROSY spectra of apo LgtC and its binary and ternary complexes, implicating multiple conformational states. Furthermore, distinct methyl chemical shift perturbations were observed upon binding of substrates UDP-2FGal and lactose and product UDP, especially for the active site isoleucines, Ile76, Ile79, Ile81, and Ile104 (Figure 3.27). These spectral changes are indicative of conformational changes associated with substrate and product binding. A simple model to explain these results is that apo LgtC exists in at least two states yielding the distinct “a” and “b” peaks in its methyl-TROSY spectrum (Figure 3.27A). We hypothesize that state “a” represents an “opened” state with respect to the flexible loops (Figure 201  3.29A). State “b” represents a “closed” state of LgtC formed upon sugar donor analog UDP2FGal binding (Figure 3.29B). These two states are in equilibrium when no substrate is present and exist at comparable populations. When the enzyme is saturated with Mg 2+•UDP-2FGal, the equilibrium shifts from state “a” to state “b” (Figure 3.27 & Figure 4.12). When a saturating amount of lactose was added to the binary complex, state “c” resulted (Figure 3.29C). States “b” and “c” are similar spectroscopically. The disappearance of the signals of Ile76 and Ile104 might be due to the rapid chemical exchange arising upon lactose binding. Finally, after the galactosyltransfer reaction, a state “d” results from bound Mg2+•UDP (Figure 3.29D). The flexible loops are opened in this state according to data obtained from the STD-NMR spectroscopy (Figure 4.6) and the relatively small CSPs of the loop residues of apo LgtC and LgtC•Mg2+•UDP in methylTROSY (Figure 4.10). However, the scaffold of the protein in state “d” is still similar to that of LgtC•UDP-2FGal (state “b”).  4.4.3 Multiple conformational equilibria of LgtC Based on the data, we proposed that apo LgtC exists in at least two conformational states that interconvert slowly on the chemical shift time-scale. This hypothesis is supported by several lines of evidence. First, within the series of methyl-TROSY spectra of apo LgtC (Figure 4.17), the relative intensities of the “a” peaks decreased with increased temperature and those of the “b” peaks increased, suggesting that the “a” to “b” states interconvert in a temperature-dependent equilibrium. Second, within the methyl-TROSY-based 1H/13C-HMQC-detected exchange experiment was done (Figure 4.18). Cross-peaks correlating “a” and “b” forms were clearly seen for the two sets of isoleucine signals, Ile31 and Ile129, indicating a conformational equilibrium. This conformational exchange of “a” and “b” states of LgtC was also observed through a  19  F/19F-  NOESY experiment of CF3Phe-labeled LgtC-F132X (Figure 4.18). Binding of substrate analog UDP-2FGal displaces the “a”/”b” equilibrium towards the “b” state, suggesting this is the bound or “closed” conformation. However, multiple peaks are still seen from single residues in the methyl-TROSY spectrum (Figure 3.27C), although crosspeaks are not observed in a 1H/13C-HMQC-detected exchange experiment. This suggests that the conformational exchange is occurring more slowly than the detection limit (~ 1 s -1) for this 202  experimental approach. Thus, once the sugar donor is bound into the active site, the flexible loops cover the active site and are not “opened” easily until hydrolysis or galactosyl-transfer reactions have been done. It is noteworthy that weak residual “a” peaks were also detected in the spectra of the LgtC binary, ternary, and product complexes. These residual “a” peaks might result from a small population of inactive form of LgtC that cannot bind substrate. Indeed, mass spectroscopy (ESIMS with apo LgtC-Q189E (Figure 4.26A) and MALDI-TOF-MS with apo wild-type LgtC and LgtC-A249X (Figure 3.16) consistently revealed the presence of higher mass forms of LgtC. Based on the CGE intermediate LgtC-Q189E-Gal previously reported (Lairson, Chiu et al. 2004), these higher mass forms were originally thought to be LgtC covalently linked with glucose or galactose formed during expression in E. coli. However, the hypothesized intermediate could not be “turned over” by adding UDP, Mn 2+, and lactose to complete the galactosyl-transfer reaction, and thus appears to result from an unknown modification of LgtC.  4.4.4 Structural dynamics of LgtC LgtC dynamics on the msec-μsec time-scale were probed using methyl relaxation dispersion measurements. Conformational exchange on this time-scale was clearly observed for Ile104, which is in the active site and interacts with UDP-2FGal, as well as for Ile191 near the active site. Ile3 and Ile31 also showed similar relaxation dispersion behavior, indicating that the effects of conformational exchange extend throughout the enzyme. R ex terms were not observed for isoleucines in the postulated flexible loops, indicating that their motions (if any) likely occur on a faster time-scale that does not lead to exchange broadening. The dynamic properties of LgtC were generally retained upon binding UDP-2FGal (Figure 4.23). However, the “b” peaks of Ile79 and Ile104 in the bound state showed slightly increase in Rex terms, indicative of msec-μsec motions. The importance of Ile79 in enzymatic catalysis was illustrated by the 16-fold reduction in kcat upon mutation to alanine (Table 3.3). The side-chain of Ile79 might be involved in van der Waals interaction with the galactose moiety of UDP-Gal. Since the conformation of UDP-Gal is bent to facilitate the nucleophile attack by the hydroxyl group of lactose (Persson, Ly et al. 2001), the side-chain of Ile79 might be required to stabilize the orientation of galactose moiety. Ile104 is in the active site and is the residue in the  203  center of the DXD (residues 103-105) motif involved in Mn2+ binding. These NMR data suggest that Ile104 becomes more dynamic during catalysis.  4.4.5  Conformational movements of the catalytic loop and α-helix J of LgtC Gln189 was initially suggested to be the catalytic nucleophile of LgtC due to the  proximity of its amide side-chain to the anomeric carbon of UDP-2FGal (~ 4.2 Å) (Persson, Ly et al. 2001). However, residual enzymatic activity remained in LgtC-Q189A mutant, making the catalytic role of Gln189 as nucleophile suspicious (Persson, Ly et al. 2001). Later, Asp190 was found to be covalently modified in a competent CGE intermediate of LgtC-Q189E (Lairson, Chiu et al. 2004). Given that the distance between the carboxylic side-chain of Asp190 and the anomeric carbon is ~ 8.3 Å, formation of this intermediate must require a dramatic conformational change. Hence, SNi-like or ion pair mechanism was proposed for the nature of the enzyme at that time. Subsequent experimental and theoretical studies also support a single displacement SNi or ion-pair mechanism (Figure 1.5) (Tvaroska 2004; Lairson, Henrissat et al. 2008; Gomez, Polyak et al. 2012). And this has been backed up by density functional theory (DFT) and quantum mechanics/molecular mechanics (QM/MM) calculations (Gomez, Polyak et al. 2012). Based on the calculated shorter distance (0.5 Ǻ) between the anomeric carbon of the computed oxocarbenium ion and the side-chain amide oxygen (Oε1) of the Gln189, a role in stabilizing the increasing positive charge at the anomeric carbon was suggested for Gln189 (Gomez, Polyak et al. 2012). Lys250, which is in one of the flexible loops, was proposed to stabilize the UDP leaving group during the catalysis while the side-chain of Asp188 was proposed to stabilize the α-galactose moiety of the acceptor by hydrogen bonds. In order for these interactions to occur, a rearrangement of the residues in the active site upon substrate binding is expected. Unfortunately, the structure of apo LgtC has not yet been solved. However, based on the data obtained from the NMR experiments, the conformation of apo LgtC is expected to be different from that of the substrate bound states. Movements of two potential active site regions upon substrate binding are proposed (Figure 4.27), namely the loop (cyan) that contains the DXD motif (Asp103 and Asp105) and the α-helix J (pink) that contains Gln189 and Asp190.  204  Figure 4.27: Active site α-helix J and loop of LgtC. (A) Ribbon diagram of LgtC ternary complex (PDB code: 1GA8) indicates the active site α-helix J (pink) and loop (cyan), with aspartate and glutamine residues colored in red and blue, respectively, hydrophobic residues in yellow, and the flexible loops in green. UDP-2FGal and 4’deoxylactose are also shown (carbon, grey; oxygen, red; nitrogen, blue; phosphorus, orange; Mn2+, magenta). (B) Ribbon diagram of LgtC ternary complex (substrates deleted), with the proposed rearrangements of catalytic loop, flexible loops, and α-helix in the absence of substrates indicated by cyan, green, and pink arrows, respectively, in the absence of substrates. (C) Ribbon diagram of LgtC ternary complex (4’-deoxylactose deleted) represents the LgtC binary complex. (D) Ribbon diagram of the LgtC ternary complex. Some useful insights into interactions and movements can be obtained from careful inspection of chemical shift data for some of the mutants made during spectral assignments. Thus, among the mutants used for methyl-TROSY spectral assignments of apo LgtC, LgtC205  I104A (Figure 3.20E) and LgtC-I104V(Figure 3.23I) caused chemical shift perturbations of Ile31 and Ile129 (Figure 4.28). Similarly, spectra of LgtC-I31V and LgtC-I129V also showed perturbations to the chemical shift of Ile104 and altered intensities of its “a” and “b” peaks. The interaction between Ile31 and Ile104 is somewhat expected because the δ1-methyl groups of two residues are separated by only ~15 Å in the crystal structure of the binary complex (Figure 4.28). However, the chemical shift perturbation of Ile129 in LgtC-I104V, or vice versa, is surprising since they are well separated from one another (24 Å). One possible explanation is that in apo LgtC, the loop containing Ile104 and the DXD motif shift inward towards the vacant active site and hence closer to Ile129 (Figure 4.27B).  Figure 4.28: Locations of Ile104 and Ile129 relative to the LgtC active site. Ribbon diagram of LgtC ternary complex (PDB code: 1GA8, with substrates deleted) shows the side-chains of Ile31, Ile104, and Ile129 in cyan and their 13C-labeled δ1-methyl groups in red. The flexible loops are colored in green. From the 1H/15N-TROSY-HSQC and methyl-TROSY spectra of LgtC, distinct chemical shifts of Tyr186 (Figure 4.12 & Figure 4.16) and Ile191 (Figure 3.27) were observed upon binding different substrates. These two residues are both located in or near the α-helix J that also 206  contains Gln189 and Asp190 (Figure 4.27). The side-chain of residue Ile191 is pointing away from the active site and towards several aromatic residues. Accordingly, we speculate that perturbations in α-helix J alter the juxtaposition of Ile191 relative to these residues and hence change its methyl chemical shifts. Thus changes in chemical shift (Δδ) of Ile191 might be used as a measure of the position of this helix.  Table 4.2: Chemical shift perturbation of residue Ile191 in methyl-TROSY spectra. *Δδ (ppm) of δ1methyl of Ile191 apo LgtC 0.000 LgtC binary complex with Mg2+•UDP-2FGal 0.275 2+ LgtC ternary complex with Mg •UDP-2FGal and lactose 0.172 LgtC product complex with Mg2+•UDP 0.120 1 13 * Δδ value is the weighed CSP of both H and C based on Equation 4-2 using the chemical shift of Ile191 in the methyl-TROSY spectrum of apo LgtC. By assuming the α-helix J moves inward in the apo enzyme (Figure 4.27A), we set the reference shift of Ile191 as Δδ = 0 ppm. When Mg2+•UDP-2FGal is bound in the active site, the α-helix J shifts outward, causing Δδ = 0.275 ppm (Table 4.2). When lactose is bound, the α-helix J moves slightly inwards (Δδ = 0.172 ppm). However, the solved crystal structures of both binary (PDB code: 1G9R) and ternary (PDB code: 1GA8) complexes overlay perfectly suggesting that any such conformational changes are subtle (Figure 1.7). The inward movement of the α-helix might help to change the conformation of UDP-Gal to facilitate the nucleophilic attack by the 4’-OH of lactose. After α-1,4-galactosyllactose is formed and has dissociated, UDP remains in the active site and the α-helix moves further inward to occupy the space that was originally occupied by the galactose moiety of UDP-Gal (Δδ = 0.120 ppm). In summary, several structural differences between apo LgtC and substrate bound form can be expected. The flexible loops are opened in the apo state according to the STD-NMR data of UDP binding (Figure 4.6). The catalytic loop containing the DXD motif and α-helix J may also shift upon substrate binding.  207  4.4.6 Proposed conformational models of LgtC reaction pathway Integrating several lines of NMR spectroscopic data, we propose a model regarding the conformational changes of LgtC along its reaction pathway (Figure 4.29). In apo LgtC, the two flexible loops exist in an equilibrium between “opened” and “closed” states. These states give rise to the approximately equal intensity “a” and “b” peaks of several isoleucine residues in the methyl-TROSY spectra of apo LgtC (Figure 3.27A). Interconversion between these two states in the apo enzyme occurs on the seconds time-scale, as evidenced by their distinct chemical shifts and by weak magnetization transfer in 1H/13C-HMQC-detected exchange experiments (Figure 4.18). Relaxation dispersion measurements indicated that side-chains within each state also undergo msec-μsec time-scale motions (Figure 4.23). Furthermore, this dynamic behaviour extends to regions of LgtC distal from its active site.  208  Figure 4.29: Proposed conformational states of LgtC along its reaction pathway. The flexible loops are colored in green and the scaffold of the protein is colored in yellow.  209  Mg2+ or Mn2+ and UDP-Gal bind LgtC simultaneously to form the LgtC binary complex. This complex is predominantly in a “closed” state with reduced msec-μsec time-scale motions, yet still exhibits some conformational heterogeneity as reflected by multiple methyl-TROSY signals. The spectra of LgtC binary and ternary complexes are similar (Figure 3.27C & D), indicating that their conformations are similar. However, addition of lactose caused the signals of Ile76, Ile79, and Ile104 to disappear. This might result from conformational exchange broadening due to the association/dissociation of lactose or motions in the lactose-bound state (Figure 3.27D). When both sugar donor and acceptor are present in the active site, galactosyltransfer reaction will occur to yield α-1,4-galactosyllactose, which is released. UDP remains bound to the active site, resulting in product inhibition. However, the flexible loops may be more open, according to STD-NMR data and the CSPs of methyl-TROSY of the LgtC•Mg2+•UDP complex (Figure 4.6 & Figure 4.10).  210  Chapter 5 General conclusions and future work  Structural studies of GTs are frequently hampered by their large sizes and conformational dynamics. The general observation that these enzymes are most frequently characterized by X-ray crystallography in their substrate-bound forms hints that flexibility precludes or at least disadvantages the crystallization of their apo-states. Thus, detailed (albeit static) “snapshots” of the enzymes along their reaction pathways are lacking. In principle, NMR spectroscopy could be used to characterize the structural and dynamic properties of GTs in solution. In practice, these enzymes are generally at the “size limit” of this technique and also exhibit substantial conformational exchange broadening. Due to new technologies, such as TROSY-based approaches, NMR spectroscopic studies of GTs, such as Alg13, NarE, and Ost4, are only now beginning to appear (Wang, Weldeghiorghis et al. 2008; Carlier, Koehler et al. 2011; Gayen and Kang 2011; Koehler, Carlier et al. 2011). In this thesis, we add the inverting (CstII) and retaining (LgtC) GTs to this list.  211  5.1 GENERAL SUMMARY OF THE SIALYLTRANSFERASE CSTII STUDIES CstII is a bifunctional inverting sialyltransferase that catalyzes the α-2,3-sialyl and α2,8-sialyl-transfers of Neu5Ac from sugar donor CMP-Neu5Ac onto the sugar acceptors lactose and sialyllactose, respectively. Previous kinetic studies showed that CstII follows an iso-ordered bi-bi-mechanism with sugar donor CMP-Neu5Ac binding before the sugar acceptor (Lairson, Henrissat et al. 2008; Lee, Lairson et al. 2011). The solved crystal structures of CstII complexed with CMP, CMP-3FNeu5Ac, and CMP•Neu5Ac-α-2,3-Gal-β1,3-GalNAc provided views of its substrate binding and some general insights into its enzymatic mechanism (Chiu, Watts et al. 2004; Lee, Lairson et al. 2011). In particular, based on the combination of crystallography, mutagenesis, and kinetic measurements, His188 was proposed to serve as the general base that abstracts the proton from the nucleophilic hydroxyl group of the sugar acceptor, thereby facilitating attack on CMP-Neu5Ac. This proposal was confirmed by a chemical rescue experiment and the similarity of the site-specific pKa value (~ 6.6) of His188 with the apparent pKa value (~ 6.5) of its pH-dependent enzymatic activity (Chan, Lairson et al. 2009). Native CstI/II is a homo-tetramer and thus too large for conventional NMR experiments. Accordingly, two active monomeric CstII mutants, CstII-F121D and CstIIY125Q, were successfully made by site-directed mutagenesis. The fact that the monomeric mutants remained active indicates that the tetramer of CstII is unnecessary for catalysis per se. Although the 1H/15N-HSQC spectrum of monomeric CstII (32 kDa) was significantly improved relative to that of the native tetramer, about half of the expected amide signals were still absent. We speculate that this might reflect msec-μsec time-scale motions of the enzyme, which lead to conformational exchange broadening. In contrast, most expected Ile, Leu, and Val signals were observed and well-dispersed in the methyl-TROSY spectrum of CstIIY125Q. Methyl relaxation dispersion experiments revealed that both apo CstII and its binary complex with CMP-3FNeu5Ac exhibited measurable conformational exchange broadening. This confirmed the existence of the postulated structural dynamics. Unfortunately, in the absence of spectral assignments (which could be obtained using a full set of mutants, as shown for LgtC), these data could not be interpreted in a more site-specific manner. Currently, there is still no crystal structure of CstII with both sugar donor and acceptor bound. In this thesis, STD-NMR spectroscopy was attempted in order to probe the 212  binding epitopes of lactose in the presence of CMP-3FNeu5Ac. However, no STD effect on lactose was observed, probably due to its weak binding. Nevertheless, this study provided insights into the motion of the CstII lid. The dynamic property of the flexible lid motif was suggested previously from the crystal structures of CstII complexed with CMP and CMP•Neu5Ac-α-2,3-Gal-β-1,3-GalNAc and its homolog, CstI, in its apo form, as there was no observable electron density of the lid in all cases (Chiu, Watts et al. 2004; Chiu, Lairson et al. 2007; Lee, Lairson et al. 2011). Even with the sugar donor analog CMP-3FNeu5Ac bound, electron density associated with the lid motif was present in only one of the four subunits. This phenomenon was also observed in the STD-NMR experiment of CMP3FNeu5Ac with CstII. The relatively small STD effect of the acetyl methyl group of the inhibitor suggests that it does not interact with the lid motif as extensively as implied by the solved crystal structure. Thus, the lid might be still partially in an “opened” conformational state even when sugar donor is bound.  213  5.2 FUTURE DIRECTIONS - SIALYLTRANSFERASES Attempt to investigate the mode of sugar acceptor lactose binding on CstII in the presence of sugar donor analog CMP-3FNeu5Ac using STD-NMR spectroscopy failed probably due to the low affinity of lactose (Km =35 mM). An alternative sugar acceptor for CstII, 3’-sialyllactose, which has a stronger binding affinity (K m = 3.5 mM) than lactose, could be used (Chiu, Watts et al. 2004). In addition, CstI could also be used since the binding affinity of sugar acceptor lactose with CstI (Km = 500 μM) is stronger than that with CstII (Chiu, Lairson et al. 2007). This could allow STD-NMR spectroscopy to be applied to CstI to investigate the binding mode of its sugar acceptor in the presence of sugar donor. The experiments could be done with the newly equipped UBC 850 MHz spectrometer as the dispersion of the signals from lactose and 3’-sialyllactose should be better than with lower field instruments. Sialyltransferases are believed to have a flexible motif that becomes ordered upon sugar donor binding. Characterization of the conformational dynamics of this flexible motif in CstII had been done using methyl-TROSY based NMR experiments. However, due to the size and instability of the protein (even with monomeric mutants), no spectral assignments were obtained. In principle, this roadblock could be overcome using a mutational approach, as demonstrated for LgtC. It might be useful, however, to consider some different sialyltransferases to study. One candidate is the monomeric GT29 porcine ST3Gal-I whose crystal structure was solved recently (Rao, Rich et al. 2009). Just as was seen with CstII, the electron density from part of the flexible loop (305-316) that was proposed to cover the active site in the presence of sugar donor was missing in ST3Gal-I. Although the size of ST3Gal-I is ~34 kDa and thus it is rather too large for conventional NMR experiments, methyl-TROSY based experiments could be used to study the conformational dynamics of the structure and the flexible loop. More importantly, only 14 alanine residues are in the structure and 2 are in the flexible loop. All alanine residues in the protein could be selectively labeled and their spectral assignments could be achieved by using site-directed mutagenesis and NOESY methods (Ayala, Sounier et al. 2009; Godoy-Ruiz, Guo et al. 2011; Kato, van Ingen et al. 2011).  214  5.3 GENERAL  SUMMARY  OF  THE  GALACTOSYLTRANSFERASE  LGTC  STUDIES LgtC is a retaining galactosyltransferase that catalyzes α-1,4-galactosyl transfer from the sugar donor UDP-Gal to the sugar acceptor lactose. Crystal structures of the LgtC binary complex with Mn2+•UDP-2FGal and the ternary complex with Mn2+•UDP-2FGal•4’deoxylactose provided high resolution views of the bound substrates and general insights into its enzymatic mechanism (Persson, Ly et al. 2001). As frequently seen with GTs, Mg2+/Mn2+ is essential for the enzymatic activities of LgtC (Powell and Brew 1976; Tran-Dinh and Neumann 1977). Although MD simulations predicted that Mn 2+ can tightly bind in the active site of LgtC in the absence of sugar donor (Snajdrova, Kulhanek et al. 2004), our NMR titration experiments unequivocally demonstrated that Mg2+ (and likely Mn2+) and UDPGal/UDP-2FGal bind to LgtC simultaneously. This is also consistent with kinetic studies showing the same apparent Km value of 20 μM for either Mn2+ or UDP-Gal in the presence of saturating amounts of the other. More significantly, despite this modest K m value, substrate binding by LgtC occurs in the slow-exchange regime on the chemical shift timescale (msec-μsec). This reflects relatively slow association and dissociation kinetics, and suggests that a rate-limiting conformational change of LgtC is required for substrate binding. Initially, based on the X-ray crystallographic structure of LgtC and by analogy with hexosaminidases, the side chain amide of Gln189 was cautiously suggested as the catalytic residue in this glycosyltransferase (Zechel and Withers 2000; Persson, Ly et al. 2001). Through the 3% residual enzyme activity observed in a Q189A mutant lacking the glutamine side-chain, it caused the authors to question such a role and they ended up favouring an S Nilike process (Persson, Ly et al. 2001). Remarkably, in a subsequent study with the Q189E mutant, Asp190 was found by mass spectrometry to be covalently bound to a galactose moiety and to form a catalytically competent intermediate (Lairson, Henrissat et al. 2008). Although the 3D structure of this species was not determined, Asp190 of wild-type LgtC in its ternary complex is ~9 Å from the substrate anomeric carbon. This suggests either that trapping is an anomaly of the Q189E mutant or that a major conformational change occurs along the reaction pathway of the wild-type enzyme. Subsequently, many additional findings indicated that neither residue is the nucleophile and supported the proposal that LgtC uses an SNi or SNi-like reaction mechanism (Ly, Lougheed et al. 2002; Tvaroska 2004; Lairson, 215  Henrissat et al. 2008; Gomez, Polyak et al. 2012). The oxocarbenium ion-like transition state in a SNi reaction mechanism or the short-lived oxocarbenium-phosphate ion pair intermediate in the SNi-like reaction mechanism must be stabilized by charge-charge and van der Waals interactions with the active site residues (Gomez, Polyak et al. 2012). To investigate the structural dynamics of LgtC by NMR spectroscopy, I first attempted to analyze its 1H/15N-TROSY-HSQC spectrum. Unfortunately, despite enormous efforts, only half of the observable amide signals could be assigned in the spectrum of apo LgtC. Furthermore, the rapid aggregation of LgtC in the presence of Mg2+ precluded our efforts to assign the spectra of its substrate and product complexes. Although the size of LgtC is a challenge (i.e. broad line-widths and spectral overlap) even with the most sophisticated NMR approaches, the fact that the signals from amides closest to the active site of the protein were either unobserved or observed but not assigned is not a coincidence. Rather, this fits with the common observation that the “important” parts of a protein are frequently the most “unusual” and hence difficult to analyze. In the case of LgtC, conformational mobility likely caused exchange broadening of the signals from active site amides. In an alternative approach, methyl-TROSY spectra of LgtC in its apo and substratebound states were acquired and all the isoleucine signals successfully assigned by mutation. Strikingly, many of the isoleucines (mono  13  CH3) yielded two methyl signals, and these  signals were found to change in relative intensity with temperature and interconvert in magnetization transfer experiments. Thus, apo LgtC exists in an equilibrium of “opened” (yielding “a” peaks) and “closed” (yielding “b” peaks) states that interconvert on a seconds time-scale. Furthermore, methyl relaxation dispersion experiments revealed an additional layer of msec-μsec time-scale motions for several isoleucines that are distal from the active site. Binding of the sugar donor shifted the conformational equilibrium towards the “closed” state and generally dampened the motions of the enzyme. This conformational shift may be responsible for substrate binding in the slow exchange regime. However, multiple signals and conformational exchanges were still observed for some isoleucines in the substrate complex, indicating that LgtC does not exist in simply two conformational states. The subsequent binding of sugar acceptor only modestly perturbs the conformation of the binary complex. After hydrolysis, the product UDP remains bound, creating yet another conformational state of LgtC. 216  Based on the results from Chapter 4, I proposed a few key conformational changes that are essential for the proposed SNi-like reaction mechanism of LgtC. In particular, the movements of the dynamic catalytic loop containing the DXD motif and α-helix J containing Gln189 and Asp190 may provide a suitable electrostatic environment in the active site upon substrate binding and facilitate the galactosyl-transfer reaction by stabilizing the transition state. Second, the two flexible loops (residues 75-80 and 246-251) may also become more ordered and cover the active site of the enzyme upon binding of Mn 2+•UDP-Gal in order to prevent wasteful hydrolysis of the activated sugar donor.  217  5.4 FUTURE DIRECTIONS - GALACTOSYLTRANSFERASES The structural dynamics of LgtC were measured by  13  C relaxation dispersion  experiment using three spectrometers with different fields (500 MHz, 600 MHz, and 850 MHz). The 600 MHz spectrometer equipped with a cryo-probe yielded data with the best signal-to-noise ratio. During the time period required for this thesis preparation, a new cryoprobe has been installed on our 850 MHz spectrometer. Accordingly, a set of high quality methyl relaxation dispersion experiments will be acquired at multiple fields in order to obtain more quantitative insights (populations, exchange rates, and chemical shift differences) ino the dynamic properties of LgtC. With the improved sensitivity of the 850 MHz spectrometer, it may also be feasible to characterize the binary complex of LgtC with Mg 2+•UDP-2FGal and the product complex with Mg2+•UDP. Methyl-TROSY based experiments could be used to investigate possible differences in the dynamic properties of wild-type LgtC and LgtC-Q189E. Although a SNi-like reaction mechanism has been reasonably well established for LgtC (Lairson, Henrissat et al. 2008; Ardevol and Rovira 2011; Gomez, Polyak et al. 2012), it would still be worthwhile to determine whether the covalent glycosyl-enzyme intermediate with Asp190 in LgtC-Q189EGal arose from artefacts due to the experimental conditions or to a change in enzymatic mechanism introduced by a mutation (Goedl and Nidetzky 2009). If it is the latter case, the structure and active site of LgtC-Q189E would be expected to be more dynamic than that of the wild-type enzyme.  218  References Abbott,  D.  W.,  M.  S.  Macauley,  et  al.  (2009).  "Streptococcus  pneumoniae  endohexosaminidase D, structural and mechanistic insight into substrate-assisted catalysis in family 85 glycoside hydrolases." J Biol Chem 284(17): 11676-89. Aharoni, A., K. Thieme, et al. (2006). "High-throughput screening methodology for the directed evolution of glycosyltransferases." Nat Methods 3(8): 609-14. Aleshin, A. E., H. J. Fromm, et al. (1998). "Multiple crystal forms of hexokinase I: new insights regarding conformational dynamics, subunit interactions, and membrane association." FEBS Lett 434(1-2): 42-6. Amaya, M. F., A. Buschiazzo, et al. (2003). "The high resolution structures of free and inhibitor-bound Trypanosoma rangeli sialidase and its comparison with T. cruzi transsialidase." J Mol Biol 325(4): 773-84. Amero, C., M. Asuncion Dura, et al. (2011). "A systematic mutagenesis-driven strategy for site-resolved NMR studies of supramolecular assemblies." J Biomol NMR 50(3): 229-36. Ang, C. W., B. C. Jacobs, et al. (2004). "The Guillain-Barre syndrome: a true case of molecular mimicry." Trends Immunol 25(2): 61-6. Angulo, J., C. Rademacher, et al. (2006). "NMR analysis of carbohydrate-protein interactions." Methods Enzymol 416: 12-30. Arakawa, T., D. Ejima, et al. (2007). "Suppression of protein interactions by arginine: a proposed mechanism of the arginine effects." Biophys Chem 127(1-2): 1-8. Ardevol, A. and C. Rovira (2011). "The molecular mechanism of enzymatic glycosyl transfer with retention of configuration: evidence for a short-lived oxocarbenium-like species." Angew Chem Int Ed Engl 50(46): 10897-901. Ayala, I., O. Hamelin, et al. (2012). "An optimized isotopic labelling strategy of isoleucinegamma2 methyl groups for solution NMR studies of high molecular weight proteins." Chem Commun (Camb) 48(10): 1434-6. Ayala, I., R. Sounier, et al. (2009). "An efficient protocol for the complete incorporation of methyl-protonated alanine in perdeuterated protein." J Biomol NMR 43(2): 111-9.  219  Bardgett, R. D., C. Freeman, et al. (2008). "Microbial contributions to climate change through carbon cycle feedbacks." ISME J 2(8): 805-14. Barlow, J. N. and J. S. Blanchard (2000). "Enzymatic synthesis of UDP-(3-deoxy-3-fluoro)D-galactose and UDP-(2-deoxy-2-fluoro)-D-galactose and substrate activity with UDP-galactopyranose mutase." Carbohydr Res 328(4): 473-80. Bennett, W. S. and R. Huber (1984). "Structural and functional aspects of domain motions in proteins." CRC Crit Rev Biochem 15(4): 291-384. Blume, A., M. Berger, et al. (2008). "Characterization of ligand binding to Nacetylglucosamine kinase studied by STD NMR." Biochemistry 47(49): 13138-46. Blume, A., M. Fitzen, et al. (2009). "Specificity of ligand binding to yeast hexokinase PII studied by STD-NMR." Carbohydr Res 344(12): 1567-74. Brecker, L., A. Schwarz, et al. (2008). "Studying non-covalent enzyme carbohydrate interactions by STD NMR." Carbohydr Res 343(12): 2153-61. Breton, C., L. Snajdrova, et al. (2006). "Structures and mechanisms of glycosyltransferases." Glycobiology 16(2): 29R-37R. Bundi, A. and K. Wuthrich (1979). "H-1-Nmr Parameters of the Common Amino-Acid Residues Measured in Aqueous-Solutions of the Linear Tetrapeptides H-Gly-Gly-XL-Ala-Oh." Biopolymers 18(2): 285-297. Burmeister, W. P., R. W. Ruigrok, et al. (1992). "The 2.2 A resolution crystal structure of influenza B neuraminidase and its complex with sialic acid." Embo J 11(1): 49-56. Campbell, R. E., R. F. Sala, et al. (1997). "Properties and kinetic analysis of UDP-glucose dehydrogenase from group A streptococci. Irreversible inhibition by UDPchloroacetol." J Biol Chem 272(6): 3416-22. Cantarel, B. L., P. M. Coutinho, et al. (2009). "The Carbohydrate-Active EnZymes database (CAZy): an expert resource for Glycogenomics." Nucleic Acids Res 37(Database issue): D233-8. Carlier, L., C. Koehler, et al. (2011). "NMR resonance assignments of NarE, a putative ADPribosylating toxin from Neisseria meningitidis." Biomol NMR Assign 5(1): 35-8. Cazet, A., S. Julien, et al. (2010). "Consequences of the expression of sialylated antigens in breast cancer." Carbohydr Res 345(10): 1377-83.  220  Cellitti, S. E., D. H. Jones, et al. (2008). "In vivo incorporation of unnatural amino acids to probe structure, dynamics, and ligand binding in a large protein by nuclear magnetic resonance spectroscopy." J Am Chem Soc 130(29): 9268-81. Chan, P. H., L. L. Lairson, et al. (2009). "NMR spectroscopic characterization of the sialyltransferase CstII from Campylobacter jejuni: histidine 188 is the general base." Biochemistry 48(47): 11220-30. Chen, H. and W. Qiu (2010). "Key technologies for bioethanol production from lignocellulose." Biotechnol Adv 28(5): 556-62. Chiu, C. P., L. L. Lairson, et al. (2007). "Structural analysis of the alpha-2,3-sialyltransferase Cst-I from Campylobacter jejuni in apo and substrate-analogue bound forms." Biochemistry 46(24): 7196-204. Chiu, C. P., A. G. Watts, et al. (2004). "Structural analysis of the sialyltransferase CstII from Campylobacter jejuni in complex with a substrate analog." Nat Struct Mol Biol 11(2): 163-70. Close, B. E., S. S. Mendiratta, et al. (2003). "The minimal structural domains required for neural cell adhesion molecule polysialylation by PST/ST8Sia IV and STX/ST8Sia II." J Biol Chem 278(33): 30796-805. Collaborative Computational Project, n. (1994). "The CCP4 suite: programs for protein crystallography." Acta Crystallogr D Biol Crystallogr 50(Pt 5): 760-3. Crennell, S., T. Takimoto, et al. (2000). "Crystal structure of the multifunctional paramyxovirus hemagglutinin-neuraminidase." Nat Struct Biol 7(11): 1068-74. Crespo, A., D. A. Scherlis, et al. (2003). "A DFT-based QM-MM approach designed for the treatment of large molecular systems: Application to chorismate mutase." Journal of Physical Chemistry B 107(49): 13728-13736. Crick, F. (1970). "Central dogma of molecular biology." Nature 227(5258): 561-3. Crick, F. H. (1958). "On protein synthesis." Symp Soc Exp Biol 12: 138-63. Davis, I. W., A. Leaver-Fay, et al. (2007). "MolProbity: all-atom contacts and structure validation for proteins and nucleic acids." Nucleic Acids Res 35(Web Server issue): W375-83. Delaglio, F., S. Grzesiek, et al. (1995). "NMRPipe: a multidimensional spectral processing system based on UNIX pipes." J Biomol NMR 6(3): 277-93. 221  DeLano, W. L. and J. W. Lam (2005). "PyMOL: A communications tool for computational models." Abstracts of Papers of the American Chemical Society 230: U1371-U1372. Doniach, S. and P. Eastman (1999). "Protein dynamics simulations from nanoseconds to microseconds." Curr Opin Struct Biol 9(2): 157-63. Dwek, R. A. (1996). "Glycobiology: Toward Understanding the Function of Sugars." Chem Rev 96(2): 683-720. Emsley, P. and K. Cowtan (2004). "Coot: model-building tools for molecular graphics." Acta Crystallogr D Biol Crystallogr 60(Pt 12 Pt 1): 2126-32. Endtz, H. P., C. W. Ang, et al. (2000). "Molecular characterization of Campylobacter jejuni from patients with Guillain-Barre and Miller Fisher syndromes." J Clin Microbiol 38(6): 2297-301. Farrow, N. A., O. Zhang, et al. (1994). "A heteronuclear correlation experiment for simultaneous determination of 15N longitudinal decay and chemical exchange rates of systems in slow equilibrium." J Biomol NMR 4(5): 727-34. Fernandez, C. and G. Wider (2003). "TROSY in NMR studies of the structure and function of large biological macromolecules." Curr Opin Struct Biol 13(5): 570-80. Flint, J., E. Taylor, et al. (2005). "Structural dissection and high-throughput screening of mannosylglycerate synthase." Nat Struct Mol Biol 12(7): 608-14. Fritz, T. A., J. Raman, et al. (2006). "Dynamic association between the catalytic and lectin domains  of  human  UDP-GalNAc:polypeptide  alpha-N-  acetylgalactosaminyltransferase-2." J Biol Chem 281(13): 8613-9. Gardner, K. H. and L. E. Kay (1998). "The use of 2H, 13C, 15N multidimensional NMR to study the structure and dynamics of proteins." Annu Rev Biophys Biomol Struct 27: 357-406. Gayen, S. and C. Kang (2011). "Solution structure of a human minimembrane protein Ost4, a subunit of the oligosaccharyltransferase complex." Biochem Biophys Res Commun 409(3): 572-6. Gelis, I., A. M. Bonvin, et al. (2007). "Structural basis for signal-sequence recognition by the translocase motor SecA as determined by NMR." Cell 131(4): 756-69. Geraldes, C. F. and M. M. Castro (1989). "Multinuclear NMR study of the interaction of vanadate with mononucleotides, ADP, and ATP." J Inorg Biochem 37(3): 213-32. 222  Gibson, R. P., J. P. Turkenburg, et al. (2002). "Insights into trehalose synthesis provided by the structure of the retaining glucosyltransferase OtsA." Chem Biol 9(12): 1337-46. Gilbert, M., J. R. Brisson, et al. (2000). "Biosynthesis of ganglioside mimics in Campylobacter jejuni OH4384. Identification of the glycosyltransferase genes, enzymatic synthesis of model compounds, and characterization of nanomole amounts by 600-mhz (1)h and (13)c NMR analysis." J Biol Chem 275(6): 3896-906. Gilbert, M., M. F. Karwaski, et al. (2002). "The genetic bases for the variation in the lipooligosaccharide of the mucosal pathogen, Campylobacter jejuni. Biosynthesis of sialylated ganglioside mimics in the core oligosaccharide." J Biol Chem 277(1): 32737. Goddard, T. D., and Kneeler, D. G. (1999). Sparky 3. San Francisco, University of California. Godoy-Ruiz, R., C. Guo, et al. (2011). "Alanine methyl groups as NMR probes of molecular structure and dynamics in high-molecular-weight proteins." J Am Chem Soc 132(51): 18340-50. Godschalk, P. C., M. L. Kuijf, et al. (2007). "Structural characterization of Campylobacter jejuni lipooligosaccharide outer cores associated with Guillain-Barre and Miller Fisher syndromes." Infect Immun 75(3): 1245-54. Goedl, C. and B. Nidetzky (2009). "Sucrose phosphorylase harbouring a redesigned, glycosyltransferase-like active site exhibits retaining glucosyl transfer in the absence of a covalent intermediate." Chembiochem 10(14): 2333-7. Gomez, H., I. Polyak, et al. (2012). "Retaining glycosyltransferase mechanism studied by QM/MM method: LgtC transfers alpha-galactose via an oxocarbenium ion-like transition state." J. Am. Chem. Soc. Gordon, R. D., P. Sivarajah, et al. (2006). "X-ray crystal structures of rabbit Nacetylglucosaminyltransferase I (GnT I) in complex with donor substrate analogues." J Mol Biol 360(1): 67-79. Gosselin, S., M. Alhussaini, et al. (1994). "A continuous spectrophotometric assay for glycosyltransferases." Anal Biochem 220(1): 92-7. Graziano, J. J., W. Liu, et al. (2008). "Selecting folded proteins from a library of secondary structural elements." J Am Chem Soc 130(1): 176-85. 223  Griffiss, J. M., H. Schneider, et al. (1988). "Lipooligosaccharides: the principal glycolipids of the neisserial outer membrane." Rev Infect Dis 10 Suppl 2: S287-95. Guerry, P., C. M. Szymanski, et al. (2002). "Phase variation of Campylobacter jejuni 81-176 lipooligosaccharide affects ganglioside mimicry and invasiveness in vitro." Infect Immun 70(2): 787-93. Haselhorst, T., A. C. Lamerz, et al. (2009). "Saturation transfer difference NMR spectroscopy as a technique to investigate protein-carbohydrate interactions in solution." Methods Mol Biol 534: 375-86. Hayashi, T., B. W. Murray, et al. (1997). "A chemoenzymatic synthesis of UDP-(2-deoxy-2fluoro)-galactose and evaluation of its interaction with galactosyltransferase." Bioorg Med Chem 5(3): 497-500. Henzler-Wildman, K. and D. Kern (2007). "Dynamic personalities of proteins." Nature 450(7172): 964-972. Horenstein, B. A. (1997). "Quantum mechanical analysis of an alpha-carboxylate-substituted oxocarbenium ion. Isotope effects for formation of the sialyl cation and the origin of an unusually large secondary C-14 isotope effect." J Am Chem Soc 119(5): 11011107. Horenstein, B. A. and M. Bruner (1998). "The N-acetyl neuraminyl oxocarbenium ion is an intermediate in the presence of anionic nucleophiles." J Am Chem Soc 120(7): 13571362. Hwang, P. M., W. Y. Choy, et al. (2002). "Solution structure and dynamics of the outer membrane enzyme PagP by NMR." Proc Natl Acad Sci U S A 99(21): 13560-5. Igura, M., N. Maita, et al. (2008). "Structure-guided identification of a new catalytic motif of oligosaccharyltransferase." Embo J 27(1): 234-43. Jackson, J. C., J. T. Hammill, et al. (2007). "Site-specific incorporation of a (19)F-amino acid into proteins as an NMR probe for characterizing protein structure and reactivity." J Am Chem Soc 129(5): 1160-6. Jeanneau, C., V. Chazalet, et al. (2004). "Structure-function analysis of the human sialyltransferase ST3Gal I: role of n-glycosylation and a novel conserved sialylmotif." J Biol Chem 279(14): 13461-8.  224  Jones, D. H., S. E. Cellitti, et al. (2010). "Site-specific labeling of proteins with NMR-active unnatural amino acids." J Biomol NMR 46(1): 89-100. Joseph, D., G. A. Petsko, et al. (1990). "Anatomy of a conformational change: hinged "lid" motion of the triosephosphate isomerase loop." Science 249(4975): 1425-8. Kato, H., H. van Ingen, et al. (2011). "Architecture of the high mobility group nucleosomal protein 2-nucleosome complex as revealed by methyl-based NMR." Proc Natl Acad Sci U S A 108(30): 12283-8. Kay, L. E. (2011). "Solution NMR spectroscopy of supra-molecular systems, why bother? A methyl-TROSY view." J Magn Reson. Kay, L. E., P. Keifer, et al. (1992). "Pure Absorption Gradient Enhanced Heteronuclear Single Quantum Correlation Spectroscopy with Improved Sensitivity." J Am Chem Soc 114(26): 10663-10665. Keeler, J. (2010). Understanding NMR Spectroscopy (2nd edition), John Wiley & Sons, Ltd. Kleckner, I. R. and M. P. Foster (2012). "GUARDD: user-friendly MATLAB software for rigorous analysis of CPMG RD NMR data." J Biomol NMR 52(1): 11-22. Koehler, C., L. Carlier, et al. (2011). "Structural and biochemical characterization of NarE, an iron-containing ADP-ribosyltransferase from Neisseria meningitidis." J Biol Chem 286(17): 14842-51. Korzhnev, D. M., K. Kloiber, et al. (2004). "Probing slow dynamics in high molecular weight proteins by methyl-TROSY NMR spectroscopy: application to a 723-residue enzyme." J Am Chem Soc 126(12): 3964-73. Krezel,  A.,  R.  Latajka,  et  al.  (2003).  "Coordination  properties  of  tris(2-  carboxyethyl)phosphine, a newly introduced thiol reductant, and its oxide." Inorg Chem 42(6): 1994-2003. Kumar, S., B. Ma, et al. (2000). "Folding and binding cascades: dynamic landscapes and population shifts." Protein Sci 9(1): 10-9. Lairson, L. L., C. P. Chiu, et al. (2004). "Intermediate trapping on a mutant retaining alphagalactosyltransferase identifies an unexpected aspartate residue." J Biol Chem 279(27): 28339-44. Lairson, L. L., B. Henrissat, et al. (2008). "Glycosyltransferases: structures, functions, and mechanisms." Annu Rev Biochem 77: 521-55. 225  Lairson, L. L., W. W. Wakarchuk, et al. (2007). "Alternative donor substrates for inverting and retaining glycosyltransferases." Chem Commun (Camb)(4): 365-7. Lariviere, L., N. Sommer, et al. (2005). "Structural evidence of a passive base-flipping mechanism for AGT, an unusual GT-B glycosyltransferase." J Mol Biol 352(1): 13950. Lazarus, M. B., Y. Nam, et al. (2011). "Structure of human O-GlcNAc transferase and its complex with a peptide substrate." Nature 469(7331): 564-7. Lee, H. J., L. L. Lairson, et al. (2011). "Structural and kinetic analysis of substrate binding to the sialyltransferase Cst-II from Campylobacter jejuni." J Biol Chem 286(41): 3592232. Legler, P. M., M. A. Massiah, et al. (2002). "Mutational, kinetic, and NMR studies of the mechanism of E. coli GDP-mannose mannosyl hydrolase, an unusual Nudix enzyme." Biochemistry 41(35): 10834-48. Leslie, A. G. (2006). "The integration of macromolecular diffraction data." Acta Crystallogr D Biol Crystallogr 62(Pt 1): 48-57. Liu, C. C. and P. G. Schultz (2010). "Adding new chemistries to the genetic code." Annu Rev Biochem 79: 413-44. Liu, J. and A. Mushegian (2003). "Three monophyletic superfamilies account for the majority of the known glycosyltransferases." Protein Sci 12(7): 1418-31. Liu, S., L. Meng, et al. (2009). "Nuclear magnetic resonance structural characterization of substrates bound to the alpha-2,6-sialyltransferase, ST6Gal-I." Biochemistry 48(47): 11211-9. Liu, T. and C. Khosla (2010). "Genetic engineering of Escherichia coli for biofuel production." Annu Rev Genet 44: 53-69. Lloyd, K. O. and K. Furukawa (1998). "Biosynthesis and functions of gangliosides: recent advances." Glycoconj J 15(7): 627-36. Lobsanov, Y. D., P. A. Romero, et al. (2004). "Structure of Kre2p/Mnt1p: a yeast alpha1,2mannosyltransferase involved in mannoprotein biosynthesis." J Biol Chem 279(17): 17921-31. Loria, J. P., M. Rance, et al. (1999). "Transverse-relaxation-optimized (TROSY) gradientenhanced triple-resonance NMR spectroscopy." J Magn Reson 141(1): 180-4. 226  Lougheed, B. (1998). Neisseria meningitidis lipopolysaccharide galactosyl transferase: mechanistc investigations and applications for oligosaccharide synthesis. Department of Chemistry. Vancouver, University of British Columbia. Lougheed, B., H. D. Ly, et al. (1999). "Glycosyl fluorides can function as substrates for nucleotide phosphosugar-dependent glycosyltransferases." J Biol Chem 274(53): 37717-22. Lovering, A. L., L. H. de Castro, et al. (2007). "Structural insight into the transglycosylation step of bacterial cell-wall biosynthesis." Science 315(5817): 1402-5. Luo, Y., S. C. Li, et al. (1998). "The crystal structure of an intramolecular trans-sialidase with a NeuAc alpha2-->3Gal specificity." Structure 6(4): 521-30. Ly, H. D., B. Lougheed, et al. (2002). "Mechanistic studies of a retaining alphagalactosyltransferase from Neisseria meningitidis." Biochemistry 41(16): 5075-85. MacLeod, A. M., T. Lindhorst, et al. (1994). "The acid/base catalyst in the exoglucanase/xylanase from Cellulomonas fimi is glutamic acid 127: evidence from detailed kinetic studies of mutants." Biochemistry 33(20): 6371-6. Martinez-Fleites, C., M. Proctor, et al. (2006). "Insights into the synthesis of lipopolysaccharide and antibiotics through the structures of two retaining glycosyltransferases from family GT4." Chem Biol 13(11): 1143-52. Mayer, M. and B. Meyer (2001). "Group epitope mapping by saturation transfer difference NMR to identify segments of a ligand in direct contact with a protein receptor." J Am Chem Soc 123(25): 6108-17. McCoy, A. J. (2007). "Solving structures of protein complexes by molecular replacement with Phaser." Acta Crystallogr D Biol Crystallogr 63(Pt 1): 32-41. McIntosh, L. P., A. J. Wand, et al. (1990). "Assignment of the backbone 1H and 15N NMR resonances of bacteriophage T4 lysozyme." Biochemistry 29(27): 6341-62. Meng, F., Y. Park, et al. (2001). "Role of proline, glycerol, and heparin as protein folding aids during refolding of rabbit muscle creatine kinase." Int J Biochem Cell Biol 33(7): 701-9. Meyer, B. and T. Peters (2003). "NMR spectroscopy techniques for screening and identifying ligand binding to protein receptors." Angew Chem Int Ed Engl 42(8): 864-90.  227  Monegal, A. and A. Planas (2006). "Chemical rescue of alpha3-galactosyltransferase. Implications in the mechanism of retaining glycosyltransferases." J Am Chem Soc 128(50): 16030-1. Moran, A. P., M. M. Prendergast, et al. (1996). "Molecular mimicry of host structures by bacterial lipopolysaccharides and its contribution to disease." FEMS Immunol Med Microbiol 16(2): 105-15. Mulder, F. A. A., C. A. E. M. Spronk, et al. (1996). "Improved HSQC experiments for the observation of exchange broadened signals." J Biomol NMR 8(2): 223-228. Murshudov, G. N., A. A. Vagin, et al. (1997). "Refinement of macromolecular structures by the maximum-likelihood method." Acta Crystallogr D Biol Crystallogr 53(Pt 3): 24055. Ni, L., H. A. Chokhawala, et al. (2007). "Crystal structures of Pasteurella multocida sialyltransferase complexes with acceptor and donor analogues reveal substrate binding sites and catalytic mechanism." Biochemistry 46(21): 6288-98. Nothaft, H. and C. M. Szymanski (2010). "Protein glycosylation in bacteria: sweeter than ever." Nat Rev Microbiol 8(11): 765-78. Ollerenshaw, J. E., V. Tugarinov, et al. (2003). "Methyl TROSY: explanation and experimental verification." Magnetic Resonance in Chemistry 41(10): 843-852. Ou, W. B., Y. D. Park, et al. (2002). "Effect of osmolytes as folding aids on creatine kinase refolding pathway." Int J Biochem Cell Biol 34(2): 136-47. Pace, C. N. (1990). "Measuring and increasing protein stability." Trends Biotechnol 8(4): 938. Pedersen, L. C., J. Dong, et al. (2003). "Crystal structure of an alpha 1,4-Nacetylhexosaminyltransferase (EXTL2), a member of the exostosin gene family involved in heparan sulfate biosynthesis." J Biol Chem 278(16): 14420-8. Penner, J. L. and G. O. Aspinall (1997). "Diversity of lipopolysaccharide structures in Campylobacter jejuni." J Infect Dis 176 Suppl 2: S135-8. Persson, K., H. D. Ly, et al. (2001). "Crystal structure of the retaining galactosyltransferase LgtC from Neisseria meningitidis in complex with donor and acceptor sugar analogs." Nat Struct Biol 8(2): 166-75.  228  Pervushin, K., R. Riek, et al. (1997). "Attenuated T2 relaxation by mutual cancellation of dipole-dipole coupling and chemical shift anisotropy indicates an avenue to NMR structures of very large biological macromolecules in solution." Proc Natl Acad Sci U S A 94(23): 12366-71. Powell, J. T. and K. Brew (1976). "Metal ion activation of galactosyltransferase." J Biol Chem 251(12): 3645-52. Rao, F. V., J. R. Rich, et al. (2009). "Structural insight into mammalian sialyltransferases." Nat Struct Mol Biol 16(11): 1186-8. Razi, N. and A. Varki (1998). "Masking and unmasking of the sialic acid-binding lectin activity of CD22 (Siglec-2) on B lymphocytes." Proc Natl Acad Sci U S A 95(13): 7469-74. Reinert, D. J., T. Jank, et al. (2005). "Structural basis for the function of Clostridium difficile toxin B." J Mol Biol 351(5): 973-81. Reynolds, C., D. Damerell, et al. (2009). "ProtorP: a protein-protein interaction analysis server." Bioinformatics 25(3): 413-4. Rosen, M. L., M. Edman, et al. (2004). "Recognition of fold and sugar linkage for glycosyltransferases by multivariate sequence analysis." J Biol Chem 279(37): 38683-92. Schubert, M., D. K. Y. Poon, et al. (2007). "Probing electrostatic interactions along the reaction pathway of a glycoside hydrolase: Histidine characterization by NMR spectroscopy." Biochemistry 46(25): 7383-7395. Schuttelkopf, A. W. and D. M. van Aalten (2004). "PRODRG: a tool for high-throughput crystallography of protein-ligand complexes." Acta Crystallogr D Biol Crystallogr 60(Pt 8): 1355-63. Shugar, D. and J. J. Fox (1952). "Spectrophotometric studies of nucleic acid derivatives and related compounds as a function of pH. I. Pyrimidines." Biochim Biophys Acta 9(2): 199-218. Snajdrova, L., P. Kulhanek, et al. (2004). "Molecular dynamics simulations of glycosyltransferase LgtC." Carbohydr Res 339(5): 995-1006.  229  Sommer, N., R. Depping, et al. (2004). "Bacteriophage T4 alpha-glucosyltransferase: a novel interaction with gp45 and aspects of the catalytic mechanism." Biochem Biophys Res Commun 323(3): 809-15. Soya, N., Y. Fang, et al. (2010). "Trapping and Characterization of Covalent Intermediates of Mutant Retaining Glycosyltransferases." Glycobiology. Sprangers, R. and L. E. Kay (2007). "Probing supramolecular structure from measurement of methyl (1)H-(13)C residual dipolar couplings." J Am Chem Soc 129(42): 12668-9. Sprangers, R. and L. E. Kay (2007). "Quantitative dynamics and binding studies of the 20S proteasome by NMR." Nature 445(7128): 618-22. Sprangers, R., X. Li, et al. (2008). "TROSY-based NMR evidence for a novel class of 20S proteasome inhibitors." Biochemistry 47(26): 6727-34. Sprangers, R., A. Velyvis, et al. (2007). "Solution NMR of supramolecular complexes: providing new insights into function." Nat Methods 4(9): 697-703. Steitz, T. A., M. Shoham, et al. (1981). "Structural dynamics of yeast hexokinase during catalysis." Philos Trans R Soc Lond B Biol Sci 293(1063): 43-52. Strajbl, M., A. Shurki, et al. (2003). "Apparent NAC effect in chorismate mutase reflects electrostatic transition state stabilization." J Am Chem Soc 125(34): 10228-37. Todd, A. E., C. A. Orengo, et al. (1999). "Evolution of protein function, from a structural perspective." Curr Opin Chem Biol 3(5): 548-56. Tossavainen, H. and P. Permi (2004). "Optimized pathway selection in intraresidual tripleresonance experiments." J Magn Reson 170(2): 244-51. Tran-Dinh, S. and J. M. Neumann (1977). "A 31P-NMR study of the interaction of Mg2+ ions with nucleoside diphosphates." Nucleic Acids Res 4(2): 397-403. Tugarinov, V., P. M. Hwang, et al. (2003). "Cross-correlated relaxation enhanced 1H[bond]13C NMR spectroscopy of methyl groups in very high molecular weight proteins and protein complexes." J Am Chem Soc 125(34): 10420-8. Tugarinov, V. and L. E. Kay (2003). "Ile, Leu, and Val methyl assignments of the 723residue malate synthase G using a new labeling strategy and novel NMR methods." J Am Chem Soc 125(45): 13868-78. Tugarinov, V. and L. E. Kay (2004). "An isotope labeling strategy for methyl TROSY spectroscopy." J Biomol NMR 28(2): 165-72. 230  Tvaroska, I. (2004). "Molecular modeling insights into the catalytic mechanism of the retaining galactosyltransferase LgtC." Carbohydr Res 339(5): 1007-14. Tzeng, Y. L. and D. S. Stephens (2000). "Epidemiology and pathogenesis of Neisseria meningitidis." Microbes Infect 2(6): 687-700. Velyvis, A., H. K. Schachman, et al. (2009). "Assignment of Ile, Leu, and Val methyl correlations  in  supra-molecular  systems:  an  application  to  aspartate  transcarbamoylase." J Am Chem Soc 131(45): 16534-43. Venter, H., A. E. Ashcroft, et al. (2002). "Molecular dissection of membrane-transport proteins: mass spectrometry and sequence determination of the galactose-H+ symport protein, GalP, of Escherichia coli and quantitative assay of the incorporation of [ring2-13C]histidine and (15)NH(3)." Biochem J 363(Pt 2): 243-52. Viladot, J. L., E. de Ramon, et al. (1998). "Probing the mechanism of Bacillus 1,3-1,4-betaD-glucan 4-glucanohydrolases by chemical rescue of inactive mutants at catalytically essential residues." Biochemistry 37(32): 11332-42. Vimr, E. R., K. A. Kalivoda, et al. (2004). "Diversity of microbial sialic acid metabolism." Microbiol Mol Biol Rev 68(1): 132-53. Vincentelli, R., S. Canaan, et al. (2004). "High-throughput automated refolding screening of inclusion bodies." Protein Sci 13(10): 2782-92. Vogel, H. J. and W. A. Bridger (1982). "Phosphorus-31 nuclear magnetic resonance studies of the methylene and fluoro analogues of adenine nucleotides. Effects of pH and magnesium ion binding." Biochemistry 21(2): 394-401. Vyas, A. A. and R. L. Schnaar (2001). "Brain gangliosides: functional ligands for myelin stability and the control of nerve regeneration." Biochimie 83(7): 677-82. Wakarchuk, W. W., A. Cunningham, et al. (1998). "Role of paired basic residues in the expression of active recombinant galactosyltransferases from the bacterial pathogen Neisseria meningitidis." Protein Eng 11(4): 295-302. Wand, A. J. (2001). "Dynamic activation of protein function: a view emerging from NMR spectroscopy." Nat Struct Biol 8(11): 926-31. Wang, X., T. Weldeghiorghis, et al. (2008). "Solution structure of Alg13: the sugar donor subunit of a yeast N-acetylglucosamine transferase." Structure 16(6): 965-75.  231  Waugh, D. S. (1996). "Genetic tools for selective labeling of proteins with alpha-15N-amino acids." J Biomol NMR 8(2): 184-92. Whiteson, K. L., Y. Chen, et al. (2007). "Identification of a potential general acid/base in the reversible phosphoryl transfer reactions catalyzed by tyrosine recombinases: Flp H305." Chem Biol 14(2): 121-9. Wilkins, M. R., E. Gasteiger, et al. (1999). "Protein identification and analysis tools in the ExPASy server." Methods Mol Biol 112: 531-52. Wilson, D. B. (2009). "Cellulases and biofuels." Curr Opin Biotechnol 20(3): 295-9. Yang, D. and L. E. Kay (1999). "Improved 1HN-detected triple resonance TROSY-based experiments." J Biomol NMR 13(1): 3-10. Yeoman, C. J., Y. Han, et al. (2010). "Thermostable enzymes as biocatalysts in the biofuel industry." Adv Appl Microbiol 70: 1-55. Yon, J. M., D. Perahia, et al. (1998). "Conformational dynamics and enzyme activity." Biochimie 80(1): 33-42. Yonetani, T. and M. Laberge (2008). "Protein dynamics explain the allosteric behaviors of hemoglobin." Biochimica Et Biophysica Acta-Proteins and Proteomics 1784(9): 1146-1158. Yuan, Y., D. Barrett, et al. (2007). "Crystal structure of a peptidoglycan glycosyltransferase suggests a model for processive glycan chain synthesis." Proc Natl Acad Sci U S A 104(13): 5348-53. Yuki, N. (2010). "Human gangliosides and bacterial lipo-oligosaccharides in the development of autoimmune neuropathies." Methods Mol Biol 600: 51-65. Yuki, N. and M. Odaka (2005). "Ganglioside mimicry as a cause of Guillain-Barre syndrome." Curr Opin Neurol 18(5): 557-61. Yuki, N., T. Taki, et al. (1993). "A bacterium lipopolysaccharide that elicits Guillain-Barre syndrome has a GM1 ganglioside-like structure." J Exp Med 178(5): 1771-5. Yuzwa, S. A., M. S. Macauley, et al. (2008). "A potent mechanism-inspired O-GlcNAcase inhibitor that blocks phosphorylation of tau in vivo." Nat Chem Biol 4(8): 483-90. Zechel, D. L. and S. G. Withers (2000). "Glycosidase mechanisms: anatomy of a finely tuned catalyst." Acc Chem Res 33(1): 11-8.  232  Zechel, D. L. and S. G. Withers (2001). "Dissection of nucleophilic and acid-base catalysis in glycosidases." Curr Opin Chem Biol 5(6): 643-9. Zhao, Q. (2010). "Protein flexibility as a biosignal." Crit Rev Eukaryot Gene Expr 20(2): 157-70. Zhu, G., Y. Xia, et al. (2000). "Protein dynamics measurements by TROSY-based NMR experiments." J Magn Reson 143(2): 423-6. Zhu, P., M. J. Klutch, et al. (2002). "Genetic diversity of three lgt loci for biosynthesis of lipooligosaccharide (LOS) in Neisseria species." Microbiology 148(Pt 6): 1833-44. Zinoviev, S., F. Muller-Langer, et al. (2010). "Next-generation biofuels: Survey of emerging technologies and sustainability issues." ChemSusChem 3(10): 1106-33.  233  Appendices  Appendix A  Enzymatic synthesis of UDP-2FGal  A.1 PREPARATION OF 2FGAL Enzymatic synthesis of UDP-2FGal was done with modified protocol from literature (Barlow and Blanchard 2000). 2,3,4,6-tetra-O-acetyl α-D-galactosyl fluoride was obtained from Withers group and used as the starting material to synthesize UDP-2FGal (Figure A.1). By using Zemplén transesterification, 2,3,4,6-tetra-O-acetyl α-D-galactosyl fluoride was deacetylated to form α-D-galactosyl fluoride (2FGal).  Figure A.1: Reaction scheme of enzymatic synthesis of UDP-2FGal.  A.2 EXPRESSION AND PURIFICATION OF FUSION ENZYME GALK/GALT The construct His6-tagged GalK/GalT (galactokinase and galactose-1-phosphate uridyltransferase) was provided by Prof. Warren Wakarchuk. The His6-tagged proteins were expressed in E. coli BL21 (DE3) cells. The cells were grown in LB broth media at 37 °C to an OD600 of 0.8, and then induced with IPTG at a final concentration of 1 mM. After further growth at 30 °C for 16 hours, the cells were harvested by centrifugation and lysed by sonication in the presence of 50 mg/L lysozyme (Sigma). The cell debris was removed by centrifuging at 15,000 rpm in a Sorvall SS32 rotor and the protein was isolated from the supernatant using a HisTrap HP column (GE Healthcare). The purified protein was concentrated to 500 μM and buffer-exchanged into 20 mM Tris, 5 mM TCEP, and pH 8.5 using an Amicon Ultra-15 Centrifugal Filter Devices. Protein concentrations were determined by UV absorbance using the predicted 280 value (90,300 M-1 cm-1) (Wilkins, Gasteiger et al. 1999).  234  A.3 SYNTHESIS OF UDP-2FGAL UDP was incorporated onto 2FGal by GalK/GalT. GalK/GalT (10 μM final) was added to a solution (1 mL) containing 10 mM ATP, 10 mM MgCl 2, 2 mM UDP-glucose, 2 mM TCEP, 15 mM 2FGal, and 25 mM HEPES, pH = 7.0. The reaction was incubated at 37 °C for 3 days. The formation of UDP-2FGal was checked by ESI-MS, 1H and  19  F NMR  spectroscopy. GalK/GalT was isolated from the reaction mixture by using Amicon Ultra-15 Centrifugal Filter Devices. Filtrate was kept for further purification. A.4 PURIFICATION OF UDP-2FGAL In order to facilitate the purification, UDP-glucose dehydrogenase (UDPGlcDH) from E. coli (Calbiochem) was used to catalyze excess UDP-glucose to UDP-glucuronic acid, which has different charge from UDP-2FGal. 46 mU of UDPGlcDH was added to a solution (1 mL) containing 1.5 mM NAD+, 2 mM DTT, 1.6 mg lyophilized mixture from GalK/GalT reaction, and 50 mM Tris, pH = 8.7. The reaction was incubated at room temperature for 5 days. UDPGlcDH was isolated from the reaction mixture by using Amicon Ultra-15 Centrifugal Filter Devices. The filtrate was first purified on a HiPrep 16/10 Q FF column with gradient of 0 – 200 mM KCl. Elution of substances was monitored at A262. The fractions containing UDP-2FGal were identified by TLC with solvent composed of EtOAc: MeOH: ddH2O: NH4OH (4:2:1:0.1). These fractions were lyophilized and dissolved in 2 mL of ddH2O. The solution was purified by P2 gel filtration column (BioRad). The fractions containing UDP-2FGal were identified by TLC and confirmed by 1H and  19  F NMR  spectroscopy. The concentration of UDP-2FGal was determined by UV absorbance using the predicted 262 value (9,900 M-1 cm-1) of UDP (Shugar and Fox 1952).  235  Appendix B  Characterization of UDPGlcDH  UDPGlcDH was used to catalyze excess UDP-glucose in the UDP-2FGal synthesis reaction to UDP-glucuronic acid, which could facilitate the purification by using anionexchange chromatography. However, UDPGlcDH was previously mentioned to be able to bind with UDP-Gal (Lougheed 1998). Thus, it can potentially bind with UDP-2FGal and catalyze it into UDP-2FGalA. Before using UDPGDH in the purification of UDP-2FGal synthesis, its substrate binding specificity has to be determined. Among the literatures about UDPGlcDH, the substrate specificity of the enzyme has not been investigated yet. Therefore, steady-state kinetic parameters for UDPGDH with UDP-Glc and UDP-Gal were determined using previously described method (Campbell, Sala et al. 1997) (Table B.1).  Table B.1: Steady state kinetic parameters for UDPGDH. Km (μM)  kcat (s-1)  kcat / Km (μM-1 s-1)  UDP-Glc  367  1.9  0.1  UDP-Gal  402  0.04  0.0001  Data for UDP-Glc and UDP-Gal in the presence of 500 μM NAD+ at pH 8.3 and 30 °C.  From the kinetic parameters, UDPGlcDH binds to UDP-Glc and UDP-Gal with same specificity due to similar Km values. This result is a little surprising because 4’-OH of galactose moiety of UDP-Gal would have steric hindrance effect with the side-chain of Asn208 based on the crystal structure of UDPGlcDH ternary complex of UDP-GlcA and NADH (Figure B.2). Fortunately, the turnover rate (kcat) of UDP-Gal is about 50-fold slower than that of UDP-Glc. Thus, UDP-2FGal would not be catalyzed by UDPGDH in a control reaction conditions. The residual UDP-2FGalA and excess UDP-GlcA could be isolated from the UDP2FGal by anion-exchange chromatography due to the charge difference.  236  Figure B.2: Crystal structure of UDGlcDH ternary complex with UDP-GlcA. Ribbon diagram of UDPGlcDH ternary complex with UDP-glucuronic acid (UDP-GlcA) and NADH (PDB code: 1DLJ) shows the interaction of Asn208 with the glucuronic acid moiety of UDP-GlcA (carbon, grey; oxygen, red; nitrogen, blue; phosphorus, orange).  237  Appendix C  1  H/15H-TROSY-HSQC NMR chemical shift table  of apo LgtC Table C.1: NMR chemical shift table of apo 15N-labeled LgtC. Cα (ppm)  Cβ (ppm)  56.170  36.240  54.135  44.232  174.114  57.331  37.714  127.733  172.853  58.837  33.713  7.989  120.354  174.550  54.567  A6  9.823  124.678  175.902  50.714  22.037  A7 D8 V22  8.660 6.888  120.348 117.651  173.313 176.278  51.653  22.997  66.684  30.601  E23  8.408  121.566  179.843  60.451  29.399  A24  8.612  121.874  179.091  54.071  17.069  A25 H26 P27  7.270 7.317  116.848 116.299  175.935 172.669  51.695  18.265  62.061  29.493  D28  8.136  114.042  176.523  54.272  39.745  T29  7.409  120.142 174.452  55.948  28.400  1  HN (ppm)  NH (ppm)  C' (ppm)  D2  5.523  122.564  173.214  I3  8.751  126.640  V4  8.710  F5  Residue  15  13  M1  E30  13  13  I31  7.848  127.018  174.844  58.920  38.293  R32  8.141  127.074  173.726  54.273  30.389  F33  8.411  120.106  175.221  56.427  41.344  H34  9.459  125.918  174.022  54.079  32.481  V35  8.880  122.699  175.330  59.978  32.795  L36  9.575  128.734  173.023  54.019  37.506  D37  9.091  129.050  176.038  53.295  42.643  A38  8.446  129.343  176.782  50.621  18.379  G39  9.145  109.835  173.559  46.890  -  I40  11.240  130.469  62.211  36.157  S41  9.344  126.167  175.099  57.355  64.392  E42  8.809  122.303  179.489  59.696  28.457  A43  8.460  121.253  180.932  54.334  17.337  N44  7.936  119.261  178.059  54.926  37.913  R45  8.747  123.364  178.818  59.859  30.138  A46  7.547  119.582  180.452  54.044  16.946  A47  7.705  122.264  179.555  54.389  18.107  238  Table C.1: NMR chemical shift table of apo 15N-labeled LgtC. Residue  HN (ppm)  NH (ppm)  C' (ppm)  Cα (ppm)  Cβ (ppm)  V48  8.298  118.532  177.783  66.703  30.773  A49  7.694  118.558  179.231  54.946  17.624  A50  7.949  116.496  178.732  53.376  17.658  N51  7.315  114.163  174.128  53.968  41.791  L52  7.260  121.168  178.112  53.650  40.984  R53  10.800  127.221  178.966 45.644  -  1  15  13  G57  13  13  N58  8.218  117.812  173.068  52.341  37.799  I59  7.302  117.864  172.174  58.590  40.175  R60  7.864  125.816  173.159  53.310  32.051  F61  8.880  123.739  175.016  57.290  38.625  I62  9.562  128.751  174.753  59.979  39.191  D63  8.389  127.046  175.329  56.368  40.666  V64  8.117  123.677  174.573  60.107  33.000  N65  8.509  125.548  174.873 58.056  27.988  E67 D68  7.942  120.162  176.796  55.842  39.079  F69  7.485  116.345  175.360  56.654  38.237  A70  7.235  123.774  178.229 44.403  44.400  G71 F72  7.751  123.223  175.036  L74  12.345  129.976  174.123  53.973  N75  7.276  121.624  178.123  53.666  P73  57.580  I76  41.810  178.975  L87 K88  7.619  114.401  L89  7.685  121.986  177.274  G90  56.687  40.658  55.374  32.381  45.642  E91  7.484  117.892  178.112  56.723  29.338  Y92  7.604  117.697  176.529  57.785  39.094  I93  7.916  117.937  174.002 52.237  19.271  53.718  39.894  55.317  40.696  54.895  34.730  A94 D95  8.446  114.868  175.932  C96  6.873  115.195  174.857  D97 K98  7.314  121.141  174.529  239  Table C.1: NMR chemical shift table of apo 15N-labeled LgtC. Residue  HN (ppm)  NH (ppm)  C' (ppm)  Cα (ppm)  Cβ (ppm)  V99  7.584  112.977  172.495  58.481  35.818  L100  9.068  127.476  174.454  53.576  43.518  Y101  9.506  127.208  173.067  56.719  40.717  L102  7.333  128.696  178.538  S111  7.993  110.567  57.539  63.496  L112  12.211  127.429  53.384  41.391  T113  8.854  121.100 56.769  40.472  1  15  13  D110  13  13  52.371  L115 W116  8.596  120.355  175.237  59.646  29.658  D117  8.140  114.292  176.541  54.247  39.861  T118  7.530  120.100  172.880  64.377  69.004  D119  8.414  126.609  176.413  52.565  39.606  L120  8.351  126.861  177.947  54.917  41.895  G121  8.220  109.997  174.034 52.728  40.879  D122 N123  8.206  117.493  175.051  53.632  38.052  W124  8.980  119.385  177.755  59.676  30.054  L125  7.386  109.580  173.737  53.584  42.550  G126  9.689  109.748  I129  8.715  125.380  D130  9.752  129.645  V133  8.275  E134  S128  57.241  64.249  175.604  63.386  38.430  122.572  179.941  65.898  30.997  8.126  119.122  177.961  58.695  29.373  R135  6.971  114.800  176.175  55.784  29.328  Q136  7.414  123.290  175.230  54.222  25.978  E137  8.040  127.253  171.770 45.200  -  Y139  8.033  123.009  177.722  62.606  37.424  K140  8.658  117.805  177.952  58.314  29.842  Q141  8.178  121.363  180.767  58.879  25.607  K142  7.774  121.929  177.870  58.579  31.080  I143  6.583  108.106  174.730  59.990  35.871  G144  7.148  105.065  45.482  45.496  M145  7.909  119.849  174.878  55.765  32.835  A146  9.011  126.000  177.966  49.883  19.743  D147  8.248  118.495  178.114  G138  240  Table C.1: NMR chemical shift table of apo 15N-labeled LgtC. 1  Residue  HN (ppm)  15  NH (ppm)  G148  C' (ppm)  Cα (ppm)  Cβ (ppm)  173.271  44.189  -  13  13  13  E149  7.236  121.978  175.113  54.363  28.391  Y150  7.153  122.892  175.289  56.784  38.083  Y151  9.968  130.770  175.464  57.251  38.462  F152  9.718  129.107  174.043  56.549  41.183  N153  8.810  119.198  176.360  55.137  40.758  A154  9.352  129.167  175.609  52.912  G155  46.406  -  V156  6.358  114.743  176.190  61.217  34.111  L157  9.084  124.489  174.883  52.384  47.330  L158  8.990  126.273  175.213  53.088  42.774  I159  9.287  127.586  175.001  61.514  38.883  N160  7.107  124.686  174.030  49.954  35.616  L161  7.207  127.451  177.475  56.465  41.089  K162  7.824  118.138  177.937  59.166  31.304  K163  6.856  118.982  179.571  58.456  32.849  W164  8.556  120.702 175.898  51.295  38.722  D168 I169  7.879  124.052  179.591  61.482  34.856  F170  9.151  124.690  178.053  62.327  38.165  K171  8.017  124.390  179.338  59.026  31.678  M172  8.430  118.456  180.128  58.738  33.769  S173  8.746  117.661  177.430 61.792  25.604  S174 E175  7.785  121.066  178.823  58.583  28.653  W176  8.013  123.156  179.236  62.762  29.613  V177  8.575  117.896  176.936  65.927  30.744  E178  7.277  117.771  177.865  58.690  28.307  Q179  6.825  114.978  177.938  56.847  28.271  Y180  7.102  115.862  176.561  59.138  37.963  K181  8.072  121.206  177.004  59.223  30.393  D182  8.409  117.211  176.897  55.496  39.998  V183  7.651  112.143  176.085  60.763  32.810  M184  7.080  119.832  176.934  57.119  34.278  Q185  8.722  123.600  176.608  55.691  30.004  Y186  10.057  121.117  55.799  37.051  Q187  8.571  116.882  57.502  41.111  D190  175.616  241  Table C.1: NMR chemical shift table of apo 15N-labeled LgtC. Residue  HN (ppm)  NH (ppm)  C' (ppm)  Cα (ppm)  Cβ (ppm)  I191  6.365  116.610  176.450  63.122  37.439  L192  7.384  114.564  177.491  56.722  40.907  177.298  55.769  38.846  45.738  -  1  15  13  13  13  N193  6.462  116.988  G194  7.990  106.092  L195  7.511  117.566  180.334  55.811  41.055  F196  6.802  114.970  174.652  53.636  36.905  K197  6.816  118.854  176.855  56.064  30.780  G198  9.611  114.178  174.587  45.277  -  G199  8.859  112.110  172.831  46.391  -  V200  7.609  119.776  175.173  59.798  34.541  C201  8.294  125.201  173.720  56.387  26.769  Y202 A203 A218  8.639 8.669  126.120 129.742  176.284  57.735  38.966  51.650  22.975  N219 R220 P228 L229  6.990 8.046  117.714 117.640  176.270 179.136  51.401  8.247  119.906  177.527  55.155  41.728 28.430 31.244 40.674  Y230  7.703  119.629  55.384  37.668  R231 D232 G269  7.678 7.114  120.847 119.721  59.229  30.189  45.525  -  S270  7.464  114.876  57.675  63.870  L271  6.924  122.816  F291 Q292  7.673  126.352  174.444  58.746 174.467 180.154  57.048  38.770  242  Appendix D  Methyl-TROSY NMR chemical shift tables of  LgtC Table D.1: NMR chemical shift table of apo LgtC. Residue Hm (ppm)  Cm (ppm)  I3a  11.576  0.480  I3b  11.400  0.511  V4i  21.474  1.103  V22i  23.612  1.351  I31a  12.755  0.353  I31b  12.113  0.284  V35i  20.897  0.792  I40  13.913  0.768  I59  14.413  0.356  I62  14.865  0.850  V64i  19.957  0.548  L74i  23.224  0.857  I76a  13.379  0.726  I76b  13.179  0.659  I79  12.848  0.659  I81  13.723  0.749  I93  13.676  0.483  V99i  18.929  0.525  L100i  25.926  0.817  I104a  14.472  0.871  I104b  13.140  0.891  L120  25.891  0.462  L125ia  25.245  0.120  L125ib  24.994  0.098  L125ii  28.387  0.947  I129a  12.957  0.159  I129b  12.480  -0.011  V133ia  21.441  0.273  V133ib  21.081  0.248  V133ii  20.535  -0.222  I143a  15.198  1.084  I143b  15.257  1.110  L157i  25.899  1.146  243  Table D.1: NMR chemical shift table of apo LgtC. Residue Hm (ppm)  Cm (ppm)  L157ii  26.898  0.803  I159  13.388  0.374  I169  9.500  0.416  V177i  22.476  0.917  V177ii  24.654  1.256  V183i  19.447  0.976  I191  15.914  1.043  L192  24.968  0.557  L192i  26.162  1.053  L192ii  20.564  0.727  L195i  25.094  0.023  L195ii  22.436  0.309  V200ia  20.696  0.106  V200ib  20.808  0.151  244  Table D.2: NMR chemical shift table of LgtC binary complex with Mg2+•UDP-2FGal. Residue  Hm (ppm)  Cm (ppm)  I3a  11.592  0.478  I3b  11.610  0.521  I31b  12.047  0.259  I40  13.636  0.766  I59  14.358  0.369  I62  14.812  0.846  I76a  13.350  0.734  I76b  13.521  1.080  I79a  12.827  0.652  I79b  15.387  0.821  I79c  13.892  0.766  I81  14.657  1.024  I93a  13.727  0.483  I93b  13.522  0.479  I104a  14.477  0.862  I104b  11.061  0.858  V106i  18.769  0.999  V106ii  27.285  1.018  I129a  12.956  0.160  I129b  12.502  0.020  V133ib  21.115  0.234  V133ii  20.374  -0.231  I143a  15.219  1.088  I143b  15.258  1.122  I159a  13.332  0.371  I159b  13.450  0.344  I169a  9.541  0.419  I169b  9.353  0.404  I191a  15.970  1.032  I191b  16.992  0.995  245  Table D.3: NMR chemical shift table of LgtC ternary complex with Mg2+•UDP2FGal•lactose. Residue Hm (ppm)  Cm (ppm)  I3a  11.529  0.505  I31b  12.033  0.246  I40  13.668  0.753  I59  14.359  0.350  I62  14.807  0.831  I76a  13.302  0.706  I79a  12.809  0.625  I79c  13.890  0.757  I81  14.492  0.948  I93a  13.741  0.472  I93b  13.524  0.471  I129a  12.992  0.152  I129b  12.461  -0.021  I143a  15.215  1.074  I143b  15.256  1.113  I159a  13.312  0.356  I159b  13.456  0.329  I169  9.447  0.395  I191a  15.942  1.016  I191c  16.630  0.986  246  Table D.4: NMR chemical shift table of LgtC product complex with Mg2+•UDP. Residue  Hm (ppm)  Cm (ppm)  I3a  11.588  0.489  I3b  11.610  0.526  I31a  12.810  0.376  I31b  12.218  0.298  I40  13.727  0.790  I59  14.377  0.385  I62  14.806  0.863  I76a  13.332  0.737  I79a  12.848  0.657  I79c  14.004  0.774  I81  13.688  0.757  I93a  13.723  0.502  I93b  13.468  0.503  I104a  14.440  0.865  I104b  12.539  0.937  I104d  14.317  0.888  I129a  12.968  0.173  I129b  12.503  0.029  I143a  15.222  1.101  I143b  15.285  1.138  I159  13.332  0.384  I169  9.499  0.427  I191a  15.981  1.047  I191d  16.411  1.026  247  

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
http://iiif.library.ubc.ca/presentation/dsp.24.1-0072879/manifest

Comment

Related Items