TOWARD THE DEVELOPMENT OF A BIOCATALYTIC PROCESS FOR THE
PRODUCTION OF β-ARYL-β-AMINO ACIDS
By
Nishanka Dilini Ratnayake

A DISSERTATION
Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of
Chemistry - Doctor of Philosophy
2015

ntermediates

ABSTRACT
TOWARD THE DEVELOPMENT OF A BIOCATALYTIC PROCESS FOR THE
PRODUCTION OF β-ARYL-β-AMINO ACIDS
By
Nishanka Dilini Ratnayake
While not abundant in nature, aromatic β-amino acids are found occasionally in
pharmacologically important natural products such as andrimid, bestatin, and Taxol. In addition, they
are used as important precursors for the synthesis of β-peptides and pharmaceuticals. Given the
importance of β-amino acids, several methods are explored for the stereoselective synthesis of βamino acids. Phenylalanine aminomutases, which isomerize (2S)-α-phenylalanine to βphenylalanine with >99.9% ee, are efficient biocatalysts to produce enantiopure β-aryl-β-amino
acids. The ultimate goal of this work was to use the knowledge from mechanistic investigations
to rationally design Pantoea agglomerans phenylalanine aminomutase (PaPAM) as a biocatalyst
to produce industrially relevant compounds on scale.
The 4-methylidene-1H-imidazol-5(4H)-one (MIO) cofactor is believed to serve as the
electrophile in the isomerization reactions performed by phenylalanine aminomutases. Over
several decades, debates have centered on whether π-electrons at the phenyl ring or the amino
✦

group of the substrate act as the nucleophile. The structure of PaPAM determined at 1.7 Å

Enzyme Engineering: F455 Mutants

resolution revealed that the substrate (α-phenylalanine) and the product (β-phenylalanine)
covalently bind to the MIO via the amino group. This data confirmed a PaPAM mechanism
where the amino group of the substrate acts as the nucleophile.
TcPAM

PaPAM

ysis

0% cinnamate
0% cinnamate; 2% relative activity

PaPAM was incubated with various isotopically labeled substrates to establish the
F455 is important for the
proper alignment of
cinnamate intermediate for
C to C amine transfer

stereochemistry and mechanism of the reaction. The reaction was found to proceed through
removal, and interchange of the pro-(3S) hydrogen and the α-amino-group of the substrate,

e, N. D; Walker, K. D; Geiger, J. H. Angew. Chem. Int. Ed. 2012, 51, 2898-2902

followed by intramolecular reattachment of the migration partners on the vicinal carbon from the
same stereo-face that they originated from. Thus, PaPAM catalyzes its isomerization reaction
with the inversion-of-configuration at both reaction termini.
Variously ring-substituted α-phenylalanine analogues (19) were used to probe the
substituent effects on the intermediate steps of PaPAM isomerization mechanism. Influence of
the substituents on the catalytic rate (kcat) of PaPAM revealed concave-down correlations with
Hammett substituent constants (σ). This trend suggested the rate-determining step changes from
the step that eliminates hydrogen and amino group from the substrate to the 'hydroamination'
step, based on the direction and magnitude of the electronic properties of the substituent.
Theoretical calculations on analyzing enzyme-substrate interaction energies revealed that the
steric effects in the protein-ligand adduct and within the ligand are dominant over electrostatic
interactions when the substrate binds.
The PaPAM E. coli whole-cell biocatalyst was shown to produce several non-natural (3S)-βaryl-β-amino acids at >99.9% ee, with the highest turnover rate in M9 minimal medium at 16 C.
The whole-cell biocatalyst biosynthesized 18 β-arylalanines with moderate to excellent converted
yield (4-96%) at production levels of 8.5 – 235 mg·L–1 over 6 h, respectively. More notably, E. coli
cells are reusable over at least five reaction cycles without a noticeable loss in activity and cell
viability. This biocatalyst offers notable advantages over conventional synthetic methods and other
biocatalysts because of its excellent enantioselectivity, broad substrate scope, single-step conversion,
and sustainability.

Copyright by
NISHANKA DILINI RATNAYAKE
2015

I dedicate this dissertation to my loving parents, siblings and my husband who made countless
sacrifices to make me who I am today.....

v

ACKNOWLEDGEMENTS
I am yet to fully comprehend that I have come to the end of a long, challenging, but
rewarding five-year journey through graduate school. I could never have come this far without
the guidance, inspiration, and support of many wonderful people.
First, I sincerely thank my advisor Prof. Kevin D. Walker for his guidance,
understanding, patience, and support all through my graduate studies at Michigan State
University. I am forever grateful to him for his continuous motivation and encouragement, and
for guiding me to grow as an independent scientist. Without his guidance and persistent help, this
dissertation and all the publications would not have been possible.
To the members of my guidance committee, I am sincerely thankful to all of you. I thank
Prof. Babak Borhan for being a great mentor, teacher, and for his tremendous advice on
preparing my second year literature seminar. Many thanks to Prof. Leslie Kuhn for introducing
me to computational docking and for providing excellent guidance while I was learning the
SLIDE docking tool. I am really thankful for her suggestions on various other research questions,
and most importantly for writing impactful recommendation letters on my behalf. I am thankful
to Prof. Dana Spence for the suggestions on improving the in vivo biocatalytic investigations of
PaPAM. Also, I cannot forget late Prof. Gregory Baker, a former member of my guidance
committee who was very supportive during my early years at Michigan State University.
I would like to thank Prof. Dr. J. H. Gieger and Dr. Susan Strom for collaborating with us
to obtain the structure of PaPAM, which provided a lot of insight into my research project. My
gratitude extends to Prof. Leslie Kuhn and Ms. Nan Liu for their support on computational
modeling investigations of phenylalanine analogues. Your findings added a significant value to
our publication.
vi

My sincere appreciation goes to Prof. Daniel Jones and members of the technical staff at
the mass spectrometry and metabolomics core, Dr. Daniel Holmes at the NMR facility, Dr.
Kaillathe (Pappan) Padmanabhan at the macromolecular computing facility, Dr. Kathy Severin,
and many other faculty members in the Departments of Chemistry and Biochemistry who went
above and beyond their ways to assist me. I would also like to thank Dr. Amy Pollock and Dr.
Ardeshir Azadnia for being supportive while I was teaching undergraduate Chemistry courses.
Further, many thanks to all the administrative personnel in the Department of Chemistry for your
continuous support during my time at Michigan State University.
I am indebted to Department of Chemistry and College of Natural Sciences for
supporting me with graduate teaching assistantships and fellowships, which greatly helped me
financially over the past five years. I am also thankful to American Chemical Society local
section, College of Natural Sciences, Graduate School, and Council of Graduate Students
(COGS) at the Michigan State University for the travel funds to attend various scientific
conferences.
All the people who I met in Walker lab were very resourceful in making my graduate
school career a success. Many thanks to Dr. Mark Ondari for not only helping me to get started
in the lab, but also continuing to help me until I graduate; I thank you for encouraging me to do
my best in graduate school, and most importantly, for being a wonderful friend and a brother. I
would like to thank Dr. Irosha Nawarathne, Dr. Danielle Nevarez Mcbride, Ms. Getrude Dibo,
Mr. Sean Sullivan, Dr. Udayanga Wanninayake, and Dr. Washington Mutatu for your help to
learn many techniques, and providing suggestions on improving my research presentations. My
sincere appreciation goes to Chelsea, Ruth, Tyler, and Prakash for making a nice, friendly
working environment in the lab. I would also like to thank Chelsea for being a great friend and

vii

for being there with me during the hard times throughout all these years. I will miss you all so
much. Many thanks to all the undergraduates who worked in Walker lab; Olivia, Chelsea T.,
Jenna, Yvonne, Manisha, Doug, and Aws: thank you very much for all your technical support.
You all were great to work with and I would not forget that you helped me somewhere along the
way.
I am thankful to all my college teachers and mentors who guided me towards a fruitful
career in science. Many thanks to my pre-undergraduate research mentor, Prof. Lalith Jayasinghe
at the Institute of Fundamental Studies, Sri Lanka, for identifying my potential in pursuing a
career in Chemistry. I sincerely thank Prof. Veranja Karunaratne and Prof. B. M. R. Bandara for
patiently guiding me while working on my undergraduate research project.
To my extended family in USA; words cannot convey how grateful I am for helping me
and my husband to resettle in the USA, and for continuously maintaining a sense of home here in
the USA. We never felt alone, although our home country is thousands of miles away. My
appreciation extends to all the Sri Lankan friends in East Lansing for all your support, and
making my years at the Michigan State University a great experience. I will miss you all!
None of the achievements I have made in my life would have been possible without the
great support from my loving family. Appachchi (dad) and ammi (mom), I am truly grateful for
recognizing my strengths, believing in me, and for giving me the support to chase after my career
goals. Without you two, I would be nowhere near the person I am today. I cannot find words to
appreciate enough for the sacrifices you both made to give us the best always; I owe you a lot
and love you always. I sincerely thank my sister and brother for being there always whenever I
needed help, and for looking out for me. Regardless how much we fought and argued, and where
we are today; you two always hold a special place in my heart. Last, but not least, I thank my

viii

husband Salinda for his immense support, love, and care throughout past five years. Words
cannot express how thankful I am for selflessly sacrificing your own dreams and goals to support
my career goals. Thank you very much for keeping me sane when the things were not going
well. I truly appreciate everything you have done for me and I love you always!

ix

TABLE OF CONTENTS
LIST OF TABLES ....................................................................................................................... xiv
LIST OF FIGURES .......................................................................................................................xv
KEY TO ABBREVIATIONS .................................................................................................... xxiii
CHAPTER 1: Overview of Biocatalysis in Organic Chemistry and Biotechnology ......................1
REFERENCES ................................................................................................................................8
CHAPTER 2: Stereochemistry and Mechanism of Phenylalanine Aminomutase from Pantoea
Agglomerans ..................................................................................................................................11
2.1. Introduction .........................................................................................................................11
2.1.1. Aminomutases: Enzyme Class of 1,2-Amino Isomerases ............................................11
2.1.2. Mechanistic Diversity of Aminomutases .....................................................................12
2.1.3. Homolytic Mechanisms of Aminomutases ..................................................................12
2.1.3.1. Adenosylcobalamin-dependent Mechanisms........................................................12
2.1.3.2. S-Adenosyl Methionine-dependent Mechanisms .................................................15
2.1.4. Heterolytic Mechanisms of Aminomutases: MIO-dependent Mechanisms.................18
2.1.5. Other Aminomutases ....................................................................................................23
2.2. Experimental ......................................................................................................................26
2.2.1. Chemicals and Reagents ...............................................................................................26
2.2.2. Instrumentation .............................................................................................................26
2.2.3. Characterization of Racemate [ring, 3-2H6]-(2R,3S)/(2S,3R)-α-Phenylalanine and
[ring, 2,3-2H7]-(2S,3S)-α-Phenylalanine .................................................................................27
2.2.3.1. GC/EI-MS Analysis ..............................................................................................27
2.2.3.2. 1H- and 2H-NMR Analyses ...................................................................................27
2.2.3.3. Analysis of Stereochemistry by the Phenylalanine Ammonia Lyase (PAL)
Reaction Stereospecificity .................................................................................................28
2.2.4. Protein Expression and Purification .............................................................................28
2.2.5. Assessment of Absolute Stereochemistry of the β-Phenylalanine Product Produced by
PaPAM ...................................................................................................................................29
2.2.6. Enzymatic assays ..........................................................................................................30
2.2.7. Derivatization and Analysis of Amino Acids ...............................................................30
2.2.8. Determining the Stereochemistry of Hydrogen Rebound ............................................31
2.2.9. Kinetic Parameters of PaPAM with o-Methyl-(2S)-α-Phenylalanine ..........................31
2.3. Results and Discussion ........................................................................................................33
2.3.1. Absolute Configuration of the Biosynthetic β-phenylalanine ......................................33
2.3.2. Mechanism of Amino Transfer ....................................................................................33
2.3.3. Stereochemistry of Amino Transfer .............................................................................36
2.3.4. Stereochemistry of Hydrogen Rebound .......................................................................39
2.3.5. Overall Mechanism and Stereochemistry of PaPAM Isomerization ...........................41
2.4. Conclusion ...........................................................................................................................45
APPENDIX ................................................................................................................................46
x

REFERENCES ...........................................................................................................................58
CHAPTER 3: Structural and Mutational Insights into PaPAM Isomerization Reaction ..............62
3.1. Introduction .........................................................................................................................62
3.1.1. Discovery of the Electrophilic MIO Group ..................................................................62
3.1.1.1. Dehydroalanine as the Electrophile ......................................................................62
3.1.1.2. Identification of the Precursor of Active Site Dehydroalanine .............................65
3.1.1.3. Discovery of the 4-Methylidene-1H-imidazol-5(4H)-one (MIO) Group .............66
3.1.2. Characteristics and Formation of the Electrophilic MIO Group ..................................66
3.1.2.1. Mutants of PpHAL ...............................................................................................69
3.1.2.2. Mutants of TcPAM ...............................................................................................71
3.1.3. Mechanism of MIO-dependent Enzymes .....................................................................73
3.1.3.1. Mechanistic Proposals on Ammonia Lyase Elimination Reaction .......................73
3.1.3.2. Mechanism of MIO-dependent Aminomutases ....................................................75
3.2. Experimental .......................................................................................................................80
3.2.1. Mutagenesis of PaPAM cDNA ....................................................................................80
3.2.2. Expression and Purification of PaPAM Mutants .........................................................81
3.2.3. Assessing the Absolute Stereochemistry of the β-Phenylalanine Product Catalyzed by
PaPAM Mutants .....................................................................................................................83
3.2.4. Total Cell Protein (TCP) Fraction Analysis .................................................................83
3.2.5. Circular Dichroism (CD) Spectroscopic Analysis .......................................................84
3.3. Results and Discussion ........................................................................................................85
3.3.1. Structural Insights into Mechanism of PaPAM Isomerization ....................................85
3.3.1.1. Characteristics of the Overall Structure and the Active Site ................................85
3.3.1.2. Identification of Mechanism-based Intermediates ................................................87
3.3.2. Mutational Insights into Mechanism and Stereochemistry of PaPAM Isomerization
. ...............................................................................................................................................89
3.3.2.1 Comparison of the PaPAM Active Site with other MIO-dependent Aminomutases
............................................................................................................................................89
3.3.2.2. Mutational Analysis to Assess the Role of Phe455 in the Isomerization Reaction
............................................................................................................................................92
3.3.2.3. Mutational Analysis of Thr167 to Assess its Dependence on MIO Construction
............................................................................................................................................95
3.3.2.3.1. Expression of Mutant Genes ........................................................................95
3.3.2.3.2.Activity and Stability of Expressible Thr167 Mutants .................................99
3.4. Conclusion .........................................................................................................................104
REFERENCES .........................................................................................................................105
CHAPTER 4: Substrate Scope of PaPAM and Effect of Ring Substituents on the Isomerization
Mechanism ...................................................................................................................................110
4.1. Introduction .......................................................................................................................110
4.1.1. Biocatalytic Production of Novel β-Aryl-β-Amino Acids .........................................110
4.1.2. MIO-dependent Isomerization Mechanism of PaPAM .............................................112
4.1.2.1. Elimination of NH2/H from α-Arylalanine Substrate .........................................113
4.1.2.2. Hydroamination of the Cinnamate Intermediate.................................................116
4.2. Experimental .....................................................................................................................119
xi

4.2.1. Substrates, Authentic Standards and Reagents ...........................................................119
4.2.2. General Instrumentation .............................................................................................119
4.2.3. Expression and Purification of PaPAM .....................................................................120
4.2.4. Assessing the Substrate Specificity of PaPAM for (2S)-α-Phenylalanine Analogues
..............................................................................................................................................121
4.2.5. Kinetic Parameters of PaPAM for (2S)-α-Phenylalanine Analogues ........................122
4.2.6. Inhibition Assays for Non-productive Substrates.......................................................123
4.3. Results and Discussion ......................................................................................................124
4.3.1. Substrate Scope of PaPAM and Kinetic Parameters of α-Phenylalanine Analogues
..............................................................................................................................................124
4.3.2. Kinetic Parameters of ortho-Substituted Analogues ..................................................126
4.3.3. Kinetic Parameters of meta-Substituted Analogues ...................................................127
4.3.4. Kinetic Parameters of para-Substituted Analogues ...................................................128
4.3.5. Substituent Effects on the PaPAM Isomerization Reaction.......................................130
4.3.5.1. Electronic Considerations on PaPAM Isomerization Reaction ..........................130
4.3.5.2. Substituent effects on Michaelis Parameters: meta-Substituents........................133
4.3.5.2.1. Electron-withdrawing Substituents ............................................................133
4.3.5.2.2. Electron-donating Substituents ..................................................................136
4.3.5.2.3. meta-Substituent Effects on Catalytic Efficiency ......................................137
4.3.5.3. Substituent effects on Michaelis Parameters: para-Substituents ........................138
4.3.5.3.1. Electron-withdrawing Substituents ............................................................139
4.3.5.3.2. Electron-donating Substituents ..................................................................140
4.3.5.3.3. para-Substituent Effects on Catalytic Efficiency ......................................141
4.3.6. Structure Activity Relationships of PaPAM Substrate Analogues ............................142
4.3.6.1. Structural Characteristics of PaPAM Active Site ...............................................142
4.3.6.2. Correlation between Substrate-PaPAM Interaction Energies and KM................143
4.3.6.2.1. Preferred Conformations of Ring-substituted Substrates ..........................146
4.3.7. Kinetic Parameters and Interaction Energies of Heteroaromatic Substrates ..............149
4.4. Conclusion .........................................................................................................................152
APPENDIX ..............................................................................................................................154
REFERENCES .........................................................................................................................185
CHAPTER 5: Biocatalytic Production of β-Aryl-β-Amino Acids using PaPAM ......................189
5.1. Introduction .......................................................................................................................189
5.1.1. Structural diversity and Significance of β-Amino Acids ...........................................189
5.1.2. Chemical Approaches for Asymmetric Synthesis of β-Amino Acids ........................192
5.1.2.1. Arndt-Eistert homologation ................................................................................192
5.1.2.2. Conjugate Addition Reactions ............................................................................194
5.1.2.2.1. Diastereoselective Methods .......................................................................195
5.1.2.2.2. Enantioselective Methods ..........................................................................196
5.1.2.3. Catalytic Asymmetric Hydrogenation Reactions................................................197
5.1.3. Biochemical Approaches for Synthesis of β-Amino Acids ........................................198
5.1.3.1. Enzymatic Resolution of β-amino acids .............................................................199
5.1.3.2. Aminomutases for the Production of β-Amino Acids ........................................202
5.1.4. Whole-cell Biocatalysis ..............................................................................................204
5.2. Experimental .....................................................................................................................209
xii

5.2.1. Substrates, Authentic Standards and Reagents ...........................................................209
5.2.2. Bacterial Strains, Plasmids and Culture Media ..........................................................209
5.2.3. General Instrumentation: GC/EI-MS Analysis ..........................................................210
5.2.4. General Procedure for Whole-cell Biocatalytic Incubations ......................................210
5.2.5. Derivatization and Quantification of Amino Acids....................................................211
5.2.6. Analysis of Substrate Uptake and Product Release by E.coli Cells ...........................212
5.2.7. Effect of Temperature, Time and Culture Medium Type...........................................213
5.2.8. Effect of Substrate Concentration ..............................................................................213
5.2.9. Effect of the Biocatalyst Amount ...............................................................................213
5.2.10. Assessing Substrate Scope of the Biocatalytic System ...........................................214
5.2.11. Sustainability of the Biocatalytic System .................................................................214
5.2.12. Calculation of Colony Forming Units (CFU) ...........................................................215
5.3. Results and Discussion ......................................................................................................216
5.3.1. General Assay Conditions ..........................................................................................216
5.3.2. Assessment of the Whole-cell Biocatalytic Properties of PaPAM ............................216
5.3.3. Enantiomeric Excess of the Biosynthetic (S)-β-Phenylalanine ..................................218
5.3.4. Effect of Temperature on β-Phenylalanine Production ..............................................219
5.3.5. The Effect of Reaction Medium on α- to β-Phenylalanine Isomerization .................221
5.3.6. The Effect of α-Phenylalanine Concentration on β-Phenylalanine Production .........222
5.3.7. Substrate Scope of the Whole-cell Biocatalytic System ............................................224
5.3.8. Sustainability of the PaPAM Whole-cell Biocatalytic System ..................................227
5.4. Conclusion .........................................................................................................................230
APPENDIX ..............................................................................................................................231
REFERENCES .........................................................................................................................233

xiii

LIST OF TABLES
Table 1.1. Biocatalytic processes in industrial applications ............................................................6
Table 2.1. Stereochemistry of β-arylalanine products produced from MIO-dependent
aminomutases and their corresponding biosynthetic products ......................................................22
Table 2.2. Mass spectral fragment ions of isotopomers of N-benzoyl-β-phenylalanine methyl
esters ..............................................................................................................................................38
Table 3.1. Mutations in 141Gly-142Ala-143Ser -144Gly-145Asp sequence in PpHAL and relative
catalytic activity of corresponding mutants ...................................................................................71
Table 4.1. Kinetic parameters of PaPAM for various substituted aryl and heteroaromatic
substrates ......................................................................................................................................129
Table A.2.1. Calculated E(p-l) and E(l) values, and preference for NH2-cis versus -trans
orientation. ...................................................................................................................................184
Table 5.1 Bioactive natural products, pharmaceuticals, and β-peptides based on β-amino acids
......................................................................................................................................................191
Table 5. 2. Colony Forming Units (CFU) of E. coli in different reaction media ........................222
Table 5.3. Production levels and conversions of PaPAM whole-cell biocatalytic system for
various α-arylalanine substrates ...................................................................................................227

xiv

LIST OF FIGURES
Figure 1.1. Comparison of the biocatalytic and chemical hydrogenation processes for sitagliptin
synthesis ...........................................................................................................................................4
Figure 1.2. Reactions catalyzed by phenylalanine aminomutases PaPAM and TcPAM in their
corresponding biosynthetic routes for andrimid and Taxol .............................................................7
Figure 2.1. 1,2-Amino shift reaction catalyzed by aminomutases .................................................11
Figure 2.2. Generation of the 5'-deoxyadenosyl radical (Ado-•CH2) species from the homolytic
Co-C bond cleavage of adenosylcobalamin ...................................................................................13
Figure 2.3. Mechanism of isomerization reaction catalyzed by adenosylcobalamin-dependent
aminomutases .................................................................................................................................14
Figure 2.4. Mechanism of isomerization reaction catalyzed by SAM-dependent aminomutases .16
Figure 2.5. Stereochemical course of lysine 2,3-aminomutase from Clostridium sp. ...................17
Figure 2.6. Mechanism of MIO cofactor formation from active site resides Ala/Thr-Ser-Gly .....19
Figure 2.7. Mechanistic proposals for MIO-dependant aminomutases .........................................20
Figure 2.8. Reaction catalyzed by glutamate 1-semialdehyde 2,1-aminomutase ..........................25
Figure 2.9. Stereochemical determination of biosynthetic β-phenylalanine..................................34
Figure 2.10. Diagnostic mass spectral fragments and mass spectrometric profile of authentic Nbenzoyl-(3S)-β-phenylalanine methyl ester ...................................................................................36
Figure 2.11. NMR spectra of N-acetyl methyl esters of various α- and β-phenylalanines ............41
Figure 2.12. The stereochemistry of the isomerization reaction catalyzed by PaPAM ................42
Figure 2.13. The stereochemistry and mechanism of isomerization reactions catalyzed by
PaPAM and TcPAM ......................................................................................................................43
Figure A.1.1. GC and mass spectrometry profiles of various N-[(1'S)-camphanoyl]-phenylalanine
methyl esters ..................................................................................................................................47
Figure A.1.2. NMR spectra of the racemate [ring, 3-2H6]-(2R,3S)/(2S,3R)-α-phenylalanine .......48
Figure A.1.3. NMR spectra of [ring, 2,3-2H7]-(2S,3S)-α-phenylalanine .......................................49

xv

Figure A.1.4. Mass spectrometry profile of cinnamic acid methyl ester derived from the racemate
[ring, 3-2H6]-(2R,3S)/(2S,3R)-α-phenylalanine by PAL catalysis.. ...............................................50
Figure A.1.5. Mass spectrometry profile of cinnamic acid methyl ester derived from [ring, 2,32
H7]-(2S,3S)-α-phenylalanine by PAL catalysis.. ..........................................................................51
Figure A.1.6. Mass spectrometry profile of the derivatized biosynthetic β-phenylalanine product
isolated after co-incubating PaPAM with unlabeled and [U-13C9, 15N]-α-phenylalanine (+98%
enriched) ........................................................................................................................................52
Figure A.1.7. Mass spectrometry profile of the derivatized biosynthetic β-phenylalanine product
isolated after co-incubating PaPAM with racemate [ring,3-2H6]-(2R,3S)/(2S,3R)-α-phenylalanine
(98+% enriched).. ..........................................................................................................................53
Figure A.1.8. Mass spectrometry profile of the derivatized biosynthetic β-phenylalanine product
isolated after incubating PaPAM with [ring, 2,3-2H7]-(2S,3S)-α-phenylalanine (90+% ee, 98+%
2
H-enriched) ...................................................................................................................................54
Figure A.1.9. Mass spectrometry profile of the N-acetyl methyl ester of dideuterio-βphenylalanine isotopomer biosynthesized from [3,3-2H2]-(2S)-α-phenylalanine.. ........................55
Figure A.1.10. Authentic racemate [2,3-2H2]-N-acetyl-(2S,3R)/(2R,3S)-β-phenylalanine methyl
ester analyzed by GC/EI-MS.. .......................................................................................................56
Figure A.1.11. Mass spectrometry profiles of o-methylcinnamate methyl ester (A) and N(ethoxycarbonyl) o-methyl-β-phenylalanine methyl ester (B) .......................................................57
Figure 3.1. First reported mechanism for MIO-dependent enzymes suggesting the involvement of
an enzyme bound carbonyl group as the electrophile ....................................................................63
Figure 3.2. . Formation of 4-amino-2-hydroxybutyric acid (2) and 2,4-diaminobutyric acid (3)
from the proposed electrophilic cofactor (1) in HAL ....................................................................64
Figure 3.3. Discovery of the MIO group and autocatalytic peptide modifications involved in the
MIO formation ...............................................................................................................................67
Figure 3.4. Partial CustalW2 multiple sequence alignment of class I lyase-like family enzymes
showing the divergence in MIO forming residues (highlighted) ...................................................68
Figure 3.5. Comparison of wt-PpHAL and PpHAL mutants lacking an MIO ..............................70
Figure 3.6. Comparison of wt-TchPAM and ThcPAM mutants lacking an MIO ..........................72
Figure 3.7. Proposed mechanisms of PAL reaction .......................................................................74

xvi

Figure 3.8. Amino-MIO adduct (A) and Friedel-Crafts-like (B) isomerization mechanisms of
MIO-dependent aminomutases based on the proposed mechanistic proposals for MIO-based
ammonia lyases ..............................................................................................................................76
Figure 3.9. wt-SgTAM active site (PDB 2OHY) showing the closed active site due to the
hydrogen bonding between Glu71 and Tyr303 .............................................................................77
Figure 3.10. Structure of SgTAM solved with various mechanism-based inhibitors and ligands
........................................................................................................................................................78
Figure 3.11. Overall structure and the active site characteristics of PaPAM ................................86
Figure 3.12. MIO-bound intermediates identified in PaPAM active sites ....................................89
Figure 3.13. Comparison of PaPAM (green) with TcPAM (yellow) and SgTAM (pink) active
sites. ...............................................................................................................................................91
Figure 3.14. Overlay of the TcPAM (yellow; showing the cinnamate and MIO) and the PaPAM
(green; showing the (S)-β-phenylalanine/MIO adduct) MIO groups ............................................92
Figure 3.15. Overlay of gas chromatography profiles of N-[(1'S)-camphanoyl] methyl ester of
authentic (S)-β-phenylalanine (14.47 min) (red), N-[(1'S)-camphanoyl] methyl ester of (S)-βphenylalanine derived from Phe455Ala (14.48 min; blue), and Phe455Asn (14.49 min; green)
........................................................................................................................................................94
Figure 3.16. Total cell protein (TCP) fraction analysis of Thr167Ala (A) and Thr167Ser (B)
mutants ...........................................................................................................................................96
Figure 3.17. Protein production levels of wt-PaPAM (A), Thr167Ser (B), and Thr167Ala (C)
mutants ...........................................................................................................................................99
Figure 3.18. Far- (A) and near- (B) UV CD spectra of wt-PaPAM and Thr167 mutants ............101
Figure 3.19. Local environment of PaPAM MIO group .............................................................103
Figure 4.1. β-Aryl-β-amino acids as key components of pharmaceutically important molecules
......................................................................................................................................................111
Figure 4.2. Mechanism of the MIO-dependent isomerization catalyzed by PaPAM ..................112
Figure 4.3. Proposed elimination mechanisms for displacement of the NH2-MIO adduct .........114
Figure 4.4. Hydroamination of the acrylate intermediate ............................................................116
Figure 4.5. Concerted (A) and stepwise (B) hydroamination mechanisms of PaPAM ...............117

xvii

Figure 4.6. Broad substrate scope of PaPAM..............................................................................124
Figure 4.7. Elimination (A) and Hydroamination (B) reaction sequences of PaPAM isomerization
......................................................................................................................................................132
Figure 4.8. Dependence of the observed
[designated as log(kx/kH)] (A) and
[designated as log(kcat/KM)] (B) on the Hammett substituent constant for the PaPAMcatalyzed isomerization of meta-substituted α-arylalanines ........................................................134
Figure 4.9. Dependence of the observed
[designated as log(kx/kH)] (A) and
[designated as log(kcat/KM)] (B) on the Hammett substituent constant for the PaPAMcatalyzed isomerization of para-substituted α-arylalanines ........................................................139
Figure 4.10. Structural characteristics of the PaPAM active site, and NH2-cis and NH2-trans
configurations of substrate analogues ..........................................................................................143
Figure 4.11. PaPAM active site residues that cause van der Waals overlap with the ringsubstituted substrates (A) and, the active site residues that sterically hinder the ortho- (B), meta(C), and para-(C) substituted substrates ......................................................................................148
Figure 4.12. Resonance hybrids of 3-thienylalanine (8) (A) and, composite resonance hybrids of
2-furylalanine (7) and 2-thienylalanine (12) (B) ..........................................................................150
Figure A.2.1. EI-MS spectra of the N-(ethoxycarbonyl) methyl ester derivatives of biosynthetic
β-phenylalanine made from PaPAM catalysis (top) and authentic β-phenylalanine (bottom). ..155
Figure A.2.2. EI-MS spectra of the N-(ethoxycarbonyl) methyl ester derivatives of biosynthetic
m-bromo-β-phenylalanine made from PaPAM catalysis (top) and authentic m-bromo-βphenylalanine (bottom) ................................................................................................................156
Figure A.2.3. EI-MS spectra of the N-(ethoxycarbonyl) methyl ester derivatives of biosynthetic
m-fluoro-β-phenylalanine made from PaPAM catalysis (top) and authentic m-fluoro-βphenylalanine (bottom) ................................................................................................................157
Figure A.2.4. EI-MS spectra of the N-(ethoxycarbonyl) methyl ester derivatives of biosynthetic
m-chloro-β-phenylalanine made from PaPAM catalysis (top) and authentic m-chloro-βphenylalanine (bottom) ................................................................................................................158
Figure A.2.5. EI-MS spectra of the N-(ethoxycarbonyl) methyl ester derivatives of biosynthetic
p-fluoro-β-phenylalanine made from PaPAM catalysis (top) and authentic p-fluoro-βphenylalanine (bottom)................................................................................................................159
Figure A.2.6. EI-MS spectra of the N-(ethoxycarbonyl) methyl ester derivatives of biosynthetic
o-methyl-β-phenylalanine made from PaPAM catalysis (top) and authentic o-methyl-βphenylalanine (bottom) ................................................................................................................160

xviii

Figure A.2.7. EI-MS spectra of the N-(ethoxycarbonyl) methyl ester derivatives of biosynthetic
2-furyl-β-alanine made from PaPAM catalysis (top) and authentic 2-furyl-β-alanine (bottom)
......................................................................................................................................................161
Figure A.2.8. EI-MS spectra of the N-(ethoxycarbonyl) methyl ester derivatives of biosynthetic
3-thienyl-β-alanine made from PaPAM catalysis (top) and authentic 3-thienyl-β-alanine (bottom)
......................................................................................................................................................162
Figure A.2.9. EI-MS spectra of the N-(ethoxycarbonyl) methyl ester derivatives of biosynthetic
m-nitro-β-phenylalanine made from PaPAM catalysis (top) and authentic m-nitro-βphenylalanine (bottom) ................................................................................................................163
Figure A.2.10. EI-MS spectra of the N-(ethoxycarbonyl) methyl ester derivatives of biosynthetic
o-fluoro-β-phenylalanine made from PaPAM catalysis (top) and authentic o-fluoro-βphenylalanine (bottom)................................................................................................................164
Figure A.2.11. EI-MS spectra of the N-(ethoxycarbonyl) methyl ester derivatives of biosynthetic
m-methoxy-β-phenylalanine made from PaPAM catalysis (top) and authentic m-methoxy-βphenylalanine (bottom) ................................................................................................................165
Figure A.2.12. EI-MS spectra of the N-(ethoxycarbonyl) methyl ester derivatives of biosynthetic
2-thienyl-β-alanine made from PaPAM catalysis (top) and authentic 2-thienyl-β-alanine (bottom)
......................................................................................................................................................166
Figure A.2.13. EI-MS spectra of the N-(ethoxycarbonyl) methyl ester derivatives of biosynthetic
m-methyl-β-phenylalanine made from PaPAM catalysis (top) and authentic m-methyl-βphenylalanine (bottom) ................................................................................................................167
Figure A.2.14. EI-MS spectra of the N-(ethoxycarbonyl) methyl ester derivatives of biosynthetic
p-chloro-β-phenylalanine made from PaPAM catalysis (top) and authentic p-chloro-βphenylalanine (bottom) ................................................................................................................168
Figure A.2.15. EI-MS spectra of the N-(ethoxycarbonyl) methyl ester derivatives of biosynthetic
p-bromo-β-phenylalanine made from PaPAM catalysis (top) and authentic p-bromo-βphenylalanine (bottom) ................................................................................................................169
Figure A.2.16. EI-MS spectra of the N-(ethoxycarbonyl) methyl ester derivatives of biosynthetic
p-methyl-β-phenylalanine made from PaPAM catalysis (top) and authentic p-methyl-βphenylalanine (bottom) ................................................................................................................170
Figure A.2.17. EI-MS spectra of the N-(ethoxycarbonyl) methyl ester derivatives of biosynthetic
p-nitro-β-phenylalanine made from PaPAM catalysis (top) and authentic p-nitro-β-phenylalanine
(bottom)........................................................................................................................................171

xix

Figure A.2.18. EI-MS spectra of the N-(ethoxycarbonyl) methyl ester derivatives of biosynthetic
p-methoxy-β-phenylalanine made from PaPAM catalysis (top) and authentic p-methoxy-βphenylalanine (bottom) ................................................................................................................172
Figure A.2.19. EI-MS spectra of the N-(ethoxycarbonyl) methyl ester derivatives of biosynthetic
o-methoxy-β-phenylalanine made from PaPAM catalysis (top) and authentic o-methoxy-βphenylalanine (bottom) ................................................................................................................173
Figure A.2.20. Hanes-Woolf plot of biosynthetic β-phenylalanine (designated as velocity, v)
catalyzed by PaPAM from α-phenylalanine (S) ..........................................................................174
Figure A.2.21. Hanes-Woolf plot of biosynthetic m-bromo-β-phenylalanine (designated as
velocity, v) catalyzed by PaPAM from m-bromo-α-phenylalanine (S) .......................................174
Figure A.2.22. Hanes-Woolf plots of biosynthetic m-fluoro-β-phenylalanine (designated as
velocity, v) catalyzed by PaPAM from m-fluoro-α-phenylalanine (S) .......................................175
Figure A.2.23. Hanes-Woolf plots of biosynthetic m-chloro-β-phenylalanine (designated as
velocity, v) catalyzed by PaPAM from m-chloro-α-phenylalanine (S) .......................................175
Figure A.2.24. Hanes-Woolf plots of biosynthetic p-fluoro-β-phenylalanine (designated as
velocity, v) catalyzed by PaPAM from p-fluoro-α-phenylalanine (S) ........................................176
Figure A.2.25. Hanes-Woolf plots of biosynthetic o-methyl-β-phenylalanine (designated as
velocity, v) catalyzed by PaPAM from o-methyl-α-phenylalanine (S) .......................................176
Figure A.2.26. Hanes-Woolf plots of biosynthetic 2-furyl-β-alanine (designated as velocity, v)
catalyzed by PaPAM from 2-furyl-α-alanine (S) ........................................................................177
Figure A.2.27. Hanes-Woolf plots of biosynthetic 3-thiophenyl-β-alanine (designated as
velocity, v) catalyzed by PaPAM from 3-thiophenyl-α-alanine (S) ............................................177
Figure A.2.28. Hanes-Woolf plots of biosynthetic m-nitro-β-phenylalanine (designated as
velocity, v) catalyzed by PaPAM from m-nitro-α-phenylalanine (S) ..........................................178
Figure A.2.29. Hanes-Woolf plots of biosynthetic o-fluoro-β-phenylalanine (designated as
velocity, v) catalyzed by PaPAM from o-fluoro-α-phenylalanine (S) ........................................178
Figure A.2.30. Hanes-Woolf plots of biosynthetic m-methoxy-β-phenylalanine (designated as
velocity, v) catalyzed by PaPAM from m-methoxy-α-phenylalanine (S) ...................................179
Figure A.2.31. Hanes-Woolf plots of biosynthetic 2-thiophenyl-β-alanine (designated as
velocity, v) catalyzed by PaPAM from 2-thiophenyl-α-alanine (S) ............................................179
Figure A.2.32. Hanes-Woolf plots of biosynthetic m-methyl-β-phenylalanine (designated as
velocity, v) catalyzed by PaPAM from m-methyl-α-phenylalanine (S) ......................................180
xx

Figure A.2.33. Hanes-Woolf plots of biosynthetic p-chloro-β-phenylalanine (designated as
velocity, v) catalyzed by PaPAM from p-chloro-α-phenylalanine (S) ........................................180
Figure A.2.34. Hanes-Woolf plots of biosynthetic p-bromo-β-phenylalanine (designated as
velocity, v) catalyzed by PaPAM from p-bromo-α-phenylalanine (S) ........................................181
Figure A.2.35. Hanes-Woolf plots of biosynthetic p-methyl-β-phenylalanine (designated as
velocity, v) catalyzed by PaPAM from p-methyl-α-phenylalanine (S) .......................................181
Figure A.2.36. Hanes-Woolf plots of biosynthetic p-nitro-β-phenylalanine (designated as
velocity, v) catalyzed by PaPAM from p-nitro-α-phenylalanine (S) ...........................................182
Figure A.2.37. Hanes-Woolf plots of biosynthetic p-methoxy-β-phenylalanine (designated as
velocity, v) catalyzed by PaPAM from p-methoxy-α-phenylalanine (S) ....................................182
Figure A.2.38. Hanes-Woolf plots of biosynthetic o-methoxy-β-phenylalanine (designated as
velocity, v) catalyzed by PaPAM from o-methoxy-α-phenylalanine (S) ....................................183
Figure A.2.39. H-bonding interaction of ortho-methoxy-α-phenylalanine (19) and active site
Tyr320 ..........................................................................................................................................183
Figure 5.1 Comparison between the substitution patterns of α- and β-amino acids (A) and
fundamental constitutional-isomers of β-amino acids (B) ...........................................................189
Figure 5.2. Synthesis of β-amino acids from α-amino acids via Arndt-Eistert homologation ....192
Figure 5.3. Continuous four-step flow system for the production of N-protected β-amino acids
......................................................................................................................................................194
Figure 5.4. Various Conjugate addition approaches for the synthesis of β-amino acids .............194
Figure 5.5. Synthesis of chiral β-amino acids using diastereoselective Michael addition of a
chiral amine ..................................................................................................................................195
Figure 5.6. Enantioselective addition of N-benzylhydroxylamine to pyrrolidinone-derived
enoates in presence of a chiral Lewis acid ...................................................................................196
Figure 5.7. Hydrogenation of enamides using chiral bisphospholane ligand MalPHOS ............197
Figure 5.8. Penicillin G acylase route to β-aryl-β-amino acids ...................................................200
Figure 5.9. Amano lipase PS catalyzed production of β-aryl-β-amino acids ..............................200
Figure 5.10. Reactions catalyzed by phenylalanine aminomutases PaPAM and TcPAM in their
corresponding biosynthetic routes for andrimid and Taxol .........................................................203

xxi

Figure 5.11. Multiple-enzyme catalyzed catalyzed production of L-methionine from racemic
MTEH ..........................................................................................................................................205
Figure 5.12. Cofactor regeneration strategy using glucose dehydrogenase .................................206
Figure 5.13. Distribution of β-phenylalanine in culture medium and E. coli cells ......................218
Figure 5.14. Gas chromatogram profiles of the N-[(1'S)-camphanoyl] methyl esters of authentic
racemate 3R– (14.10 min) and 3S–β–phenylalanine (14.43 min) (solid line), and biosynthetic β–
phenylalanine (14.43 min) (dotted line).......................................................................................219
Figure 5.15. Effect of reaction temperature and reaction medium ..............................................220
Figure 5.16. Effect of increased substrate concentration and biocatalyst amount .......................223
Figure 5.17. Recycling of the E. coli whole-cell biocatalyst (OD600 ~35) for β-phenylalanine
production ....................................................................................................................................228
Figure A.3.1. Comparison of product accumulation and substrate depletion over time .............232
Figure A.3.2. Comparison of substrate depletion in different reaction media .............................232

xxii

KEY TO ABBREVIATIONS
ATP: Adenosine triphosphate
BME-vitamins: Eagle’s basal medium vitamins
Bu3N: Tributylamine
BzCl: Benzoyl chloride
CaCl2: Calcium chloide
CD: Circular dichroism
CFU: Colony-forming units
CH2N2: Diazomethane
CHCl3: Chloroform
ClCO2Et: Ethyl chloroformate
D2O: Deuterium oxide
DMSO: Dimethyl sulfoxide
DNA: Deoxyribonucleic acid
E. coli: Escherichia coli
EDTA: Ethylenediaminetetraacetic acid
EtOH: Ethanol
GC/EI-MS: Gas chromatography electron ionization mass spectrometry
HCl: Hydrochloric acid
IPTG: Isopropyl-β-d-thiogalactopyranoside
KCN: Potassium cyanide
KH2PO4: Monopotassium phosphate
KOH: Potassium hydroxide
xxiii

MeOH: Methanol
MgSO4: Magnesium sulfate
MIO: 4-methylidene-1H-imidazol-5(4H)-one
mRNA: Messenger ribonucleic acid
MTBE: Methyl tert-butyl ether
MWCO: Molecular weight cutoff
Na2HPO4.7H2O: Sodium monohydrogen phosphate heptahydrate
NaBH4: Sodium borohydrate
NaCl: Sodium chloride
NaCN: Sodium cyanide
NAD(P): Nicotinamide adenine dinucleotide phosphate
NaHSO3: Sodium bisulfite
NaOH: Sodium hydroxide
NH4Cl: Ammonium chloride
NH4OH: Ammonium hydroxide
NMR: Nuclear magnetic resonance
n-PrOH: Propyl alcohol
OD: Optical density
PCR: Polymerase chain reaction
Pd(OH)2: Palladium hydroxide
PLP: Pyridoxal 5'-phosphate
SAM: S-adenosyl methionine
SDS−PAGE: Sodium dodecyl sulfate Polyacrylamide gel electrophoresis

xxiv

SOCl2: Thionyl chloride
TCP: Total cell protein
THF: Tetrahydrofuran
TMSCH2N2: Trimethylsilyldiazomethane
UV: Ultraviolet

xxv

CHAPTER 1: Overview of Biocatalysis in Organic Chemistry and Biotechnology
Biocatalysis, the use of isolated enzymes or microbial whole-cells as catalysts, offers
various unique advantages over conventional chemical catalysis.1,2 The most important
advantages of a biocatalyst include, excellent stereo-, regio- and chemo-selectivity. Additionally,
protection/deprotection steps are can often be avoided, and therefore, the reactions are generally
atom- and reaction-step economical. Other advantages, such as mild operational conditions and
reduced hazardous waste generation are also very attractive in large scale applications.
Traditionally, the enzymes were directly isolated from their corresponding sources such
as microbes, plants, insects, and mammalian species. Consequently, the main drawback in
enzyme catalysis until the end of the 1970s was the enzyme production in large enough
quantities for practical applications.3 However, with advances in recombinant DNA technology,
the DNA sequences could be manipulated and proteins overexpressed in non-native host
organisms.2 This provided a means to produce biocatalysts in large quantities in fast-growing
organisms.
Nonetheless, narrow substrate scope, substrate or product inhibition and low stability of
the biocatalysts limit the application of enzymes in industrial scale applications.2 Developments
in protein engineering methods using rational design or directed evolution (by error-prone PCR
or gene shuffling) and advances in understanding protein structure-function relationships are
enabling scientists to rapidly tailor the properties of biocatalysts for particular chemical
processes.1

1

Generally, improvements in enzyme thermal stability often results in loss of the catalytic
activity at lower temperatures.4 However, the thermal stability of Bacillus subtilis, p-nitrobenzyl
esterase was increased significantly without reducing the catalytic activity at lower temperature.4
Six generations of random mutagenesis increased the melting temperature of the esterase by 14
C.
Rational design based on site-specific mutagenesis has also been instrumental in
enhancing the thermostability of enzymes. For example, a moderately stable protease from
Bacillus stearothermophilus (TLP-ste) was made hyper-stable by rationally designed mutations.5
By comparing the structure of TLP-ste with naturally occurring more thermostable homologues,
eight point mutations and a disulfide bridge were introduced in TPL-ste. These modifications
resulted in a 340-fold more stable protein compared to wt- TLP-ste at 100 C.5 Additionally, the
8-fold mutant retained its activity at high concentrations of denaturing agents.
Although it is more challenging, rational design of proteins has been used to improve
their stereoselectivity.3 For example, the enantioselectivity and catalytic activity of Burkholderia
cepacia lipase toward poor substrates bearing bulky substituents was improved by site-specific
mutagenesis.6 Based on molecular docking, a double mutant (I290A/I287F) was designed to
create more space for bulkier substrates and to improve the catalytic activity of the (R)enantiomer of the racemic substrates. Consequently, the E value (E = ln[1-c(1+ee)]/ln[1-c(1-ee)],
where c = conversion; ee = enantiomeric excess) for bulkier substrates increased from five (wildtype) to 200 for the double mutant.6
Merck and Codexis researchers recently engineered an (R)-selective transaminase (ATA117) to increase the efficiency of a large-scale manufacturing process of the antidiabetic
2

compound sitagliptin (Figure 1.1).7 Starting from ATA-117, which was active only toward
methyl- and small cyclic- ketones, a combination of computational modeling and site-saturation
mutagenesis was used to create a biocatalyst with marginal activity for prositaglipin ketone (4%
conversion). The ATA-117 variant with the highest activity was further engineered via 11 rounds
of directed evolution to optimize the activity, tolerance for organic cosolvents and higher
substrate concentrations, and stability at the elevated reaction temperatures.7 The evolved
transaminase contained 27 mutations in the active site and the dimer interface, which improved
the substrate binding as well as protein stability. In the optimized process, 6 g/L of the best
transaminase variant tolerates a 200 g/L concentration of prositagliptin ketone in 50% DMSO at
40 C, and produces sitagliptin in 92% yield at >99.95% ee.7 This biocatalytic route replaced the
high pressure (250 psi), Rh-catalyzed enamine asymmetric hydrogenation for the large-scale
manufacture of sitagliptin (Figure 1.1). The biocatalytic route not only reduced the total waste
(19% reduction) and eliminated all transition metals, but also increased the overall yield (10-13%
increase) and the productivity (53% increase) compared to the metal-catalyzed process.7
Biocatalysis became a mature technology through several protein engineering
innovations, and currently several biocatalytic processes are operational in pharmaceutical,
chemical, agricultural, and food industries.8,9 Hydrolases10 and ketoreductases11 are the most
frequently used biocatalysts in industrial organic synthesis. For example, Candida antarctica
lipase B led to a more efficient and less expensive manufacturing process for the key
intermediate L-fluoroleucine (1) (Table 1.1A) in the synthesis of antiresorptive agent odanacatib
(Merck).12 This process was demonstrated at 100 kg scale with >90% yield and 88% ee.
Additionally, enzymatic kinetic resolution by Thermomyces lanuginosus lipase was used in
preparative scale synthesis of (1S,3S)-3-aminocyclohexanol (2) (Table 1.1B) with >96% ee
3

(Novartis).13 An engineered ketoreductase in combination with a glucose dehydrogenase for
cofactor regeneration has been widely investigated for the manufacture of atorvastatin, the active
ingredient of cholesterol lowering drug Lipitor (Codexis).14 Biocatalytic reduction of ethyl-4chloroacetoacetate by the ketoreductase produced ethyl-4-chloro-3-hydroxybutyrate (3) (Table
1.1C), a key intermediate of atorvastin, in 96% isolated yield and >99.5% ee.

Figure 1.1. Comparison of the biocatalytic and chemical hydrogenation processes for sitagliptin
synthesis. The enzymatic route using an engineered transaminase is superior (higher yield and
enantioselectivity) to the metal-catalyzed asymmetric hydrogenation method.

4

In addition to lipases and ketoreductases, various other classes of enzymes are also
emerging and expanding as industrially relevant biocatalysts. Examples include transaminases in
the sitagliptin manufacture (Merck and Codexis) (cf. Figure 1.1),7 aldolases in the production of
the natural sweetener monatin (4) (CSIR Biosciences) (Table 1.1D),15 an (R)-amino acid oxidase
used in (S)-2-amino-3-(6-o-tolylpyridin-3-yl)propanoic acid (5) production (Bristol-Myers
Squibb) (Table 1.1E),16 and proteolytic enzyme thermolysin in the production of low-calorie
sweetener aspartame (6) (DSM) (Table 1.1F).8,17
In view of increasing demand for renewable sources for energy and chemical feedstock,
biocatalysis is becoming an attractive process.18 Although many biocatalysts are implemented in
the syntheses of various fine chemicals, there is still a growing need for more efficient or novel
biocatalysts to supply the increasing demand of chiral compounds for industrial purposes. The
impact of emerging technologies in bio-informatics, next generation sequencing, and highthroughput screening open up new avenues to discover novel biocatalysts and to create efficient
enzyme variants.19,20 De novo design of enzymes with pre-determined characteristics,21 powerful
microorganism screening methods such as metagenomics,22 as well as the exploration of the
catalytic flexibility and efficiency of existing biocatalysts offer new approaches for the
exploration of enzyme catalysts.

5

Table 1.1. Biocatalytic processes in industrial applications

Enzyme

Product

Source

Reference

Merck

12

A.

Lipase
(Candida antarctica)

B.

Lipase
(Thermomyces lanuginosus)

Novartis

13

C.

Ketoreductase and glucose
dehydrogenase

Codexis

14

D.

Aldolase

CSIR
Biosciences

15

E.

(R)-amino acid oxidase

Bristol-Myers
Squibb

16

F.

Protease thermolysin

DSM

8, 17

Phenylalanine aminomutases (PAM) are responsible for the biosynthesis of enantiopure
β-phenylalanines in structurally diverse bioactive natural products.23,24 In Taxus plants, TcPAM
catalyzes the conversion of (2S)-α-phenylalanine to (3R)-β-phenylalanine, a proposed
biosynthetic precursor of the phenylisoserine side chain of Taxol (Figure 1.2).24 A homologous
enzyme from Pantoea agglomerans (PaPAM) produces the enantiomeric (3S)-β-phenylalanine
in the biosynthetic pathway of the antibiotic andrimid (Figure 1.2).23
6

Figure 1.2. Reactions catalyzed by phenylalanine aminomutases PaPAM and TcPAM in their
corresponding biosynthetic routes for andrimid and Taxol.

The excellent product enantioselectivity (99.9% ee) and non-reliance on external
cofactors make PAMs attractive biocatalysts for producing β-amino acids, which are highly
valuable building blocks of peptidomimetics and pharmaceutics.25 While the potential of TcPAM
for the synthesis of β-amino acids has been demonstrated in various investigations in this area,2630

PaPAM has not been evaluated for its biocatalytic potential and applications. The main

objective of this work was to investigate the mechanistic, stereochemical, kinetic and whole-cell
biocatalytic characteristics aimed at developing PaPAM as a biocatalyst for producing βarylalanines. Chapter 2, 3, and 4 of this dissertation will highlight the stereochemical,
mechanistic, kinetic, and structural investigations that will guide the rational design of PaPAM
as a better biocatalyst. Further, the initial investigations on developing PaPAM as a whole-cell
biocatalyst will be described in chapter 5.

7

REFERENCES

8

REFERENCES
(1)

de Carvalho, C. C. Biotechnol. Adv. 2011, 29, 75-83.

(2)

Arnold, F. H. Nature 2001, 409, 253-257.

(3)

Reetz, M. T. J. Am. Chem. Soc. 2013, 135, 12480-12496.

(4)

Giver, L.; Gershenson, A.; Freskgard, P. O.; Arnold, F. H. Proc. Natl. Acad. Sci. USA
1998, 95, 12809-12813.

(5)

Van den Burg, B.; Vriend, G.; Veltman, O. R.; Venema, G.; Eijsink, V. G. Proc. Natl.
Acad. Sci. USA 1998, 95, 2056-2060.

(6)

Ema, T.; Kamata, S.; Takeda, M.; Nakano, Y.; Sakai, T. Chem. Commun. 2010, 46, 54405442.

(7)

Savile, C. K.; Janey, J. M.; Mundorff, E. C.; Moore, J. C.; Tam, S.; Jarvis, W. R.;
Colbeck, J. C.; Krebber, A.; Fleitz, F. J.; Brands, J.; Devine, P. N.; Huisman, G. W.;
Hughes, G. J. Science 2010, 329, 305-309.

(8)

Schmid, A.; Dordick, J. S.; Hauer, B.; Kiener, A.; Wubbolts, M.; Witholt, B. Nature
2001, 409, 258-268.

(9)

Breuer, M.; Ditrich, K.; Habicher, T.; Hauer, B.; Kesseler, M.; Sturmer, R.; Zelinski, T.
Angew. Chem. Int. Ed. 2004, 43, 788-824.

(10)

Gotor-Fernandez, V.; Brieva, R.; Gotor, V. J. Mol. Catal. B: Enzym. 2006, 40, 111-120.

(11)

Moore, J. C.; Pollard, D. J.; Kosjek, B.; Devine, P. N. Acc. Chem. Res. 2007, 40, 14121419.

(12)

Truppo, M. D.; Hughes, G. Org. Process Res. Dev. 2011, 15, 1033-1035.

(13)

Brocklehurst, C. E.; Laumen, K.; La Vecchia, L.; Shaw, D.; Vogtle, M. Org. Process
Res. Dev. 2011, 15, 294-300.

(14)

Ma, S. K.; Gruber, J.; Davis, C.; Newman, L.; Gray, D.; Wang, A.; Grate, J.; Huisman,
G. W.; Sheldon, R. A. Green Chemistry 2010, 12, 81-86.

(15)

L., R. A.; Buddoo, S. R.; Gordon, G. E. R.; Beemadu, S.; Kupi, B. G.; Lepuru, M. J.;
Maumela, M. C.; Parsoo, A.; Sibiya, D. M.; Brady, D. Org. Process Res. Dev. 2011, 15,
249–257.

9

(16)

Chen, Y.; Goldberg, S. L.; Hanson, R. L.; Parker, W. L.; Gill, I.; Tully, T. P.; Montana,
M. A.; Goswami, A.; Patel, R. N. Org. Process Res. Dev. 2011, 15, 241–248.

(17)

Schulze, B.; Wubbolts, M. G. Curr. Opin. Biotechnol. 1999, 10, 609-615.

(18)

Pollard, D. J.; Woodley, J. M. Trends Biotechnol. 2007, 25, 66-73.

(19)

Turner, N. J.; Truppo, M. D. Curr. Opin. Chem. Biol. 2013, 17, 212-214.

(20)

Davids, T.; Schmidt, M.; Bottcher, D.; Bornscheuer, U. T. Curr. Opin. Chem. Biol. 2013,
17, 215-220.

(21)

Kries, H.; Blomberg, R.; Hilvert, D. Curr. Opin. Chem. Biol. 2013, 17, 221-228.

(22)

Lorenz, P.; Eck, J. Nat. Rev. Microbiol. 2005, 3, 510-516.

(23)

Magarvey, N. A.; Fortin, P. D.; Thomas, P. M.; Kelleher, N. L.; Walsh, C. T. ACS Chem.
Biol. 2008, 3, 542-554.

(24)

Walker, K. D.; Klettke, K.; Akiyama, T.; Croteau, R. J. Biol. Chem. 2004, 279, 5394753954.

(25)

Lelais, G.; Seebach, D. Biopolymers 2004, 76, 206-243.

(26)

Klettke, K. L.; Sanyal, S.; Mutatu, W.; Walker, K. D. J. Am. Chem. Soc. 2007, 129, 69886989.

(27)

Wanninayake, U.; DePorre, Y.; Ondari, M.; Walker, K. D. Biochemistry 2011, 50,
10082-10090.

(28)

Szymanski, W.; Wu, B.; Weiner, B.; de Wildeman, S.; Feringa, B. L.; Janssen, D. B. J.
Org. Chem. 2009, 74, 9152-9157.

(29)

Wu, B.; Szymanski, W.; Wietzes, P.; de Wildeman, S.; Poelarends, G. J.; Feringa, B. L.;
Janssen, D. B. ChemBioChem 2009, 10, 338-344.

(30)

Wu, B.; Szymanski, W.; Wybenga, G. G.; Heberling, M. M.; Bartsch, S.; de Wildeman,
S.; Poelarends, G. J.; Feringa, B. L.; Dijkstra, B. W.; Janssen, D. B. Angew. Chem. Int.
Ed. 2012, 51, 482-486.

10

CHAPTER 2: Stereochemistry and Mechanism of Phenylalanine Aminomutase from
Pantoea Agglomerans

2.1. Introduction

2.1.1. Aminomutases: Enzyme Class of 1,2-Amino Isomerases
Intramolecular transferase (EC 5.4) family contains enzymes that transfer acyl-, amino-,
phospho- or other functional groups from one position of a molecule to another.1 Aminomutases
(EC 5.4.3) comprise a subclass of the intramolecular isomerase family and catalyze the
chemically challenging exchange of an amino group and a proton present on the vicinal carbons
of a substrate (Figure 2.1).2

Figure 2.1. 1,2-Amino shift reaction catalyzed by aminomutases.

Aminomutases have gained much attention in recent years due to their potential
application as biocatalysts in the production of pharmacologically significant molecules and
intermediates.3 The aminomutase family consists of lysine 2,3- (EC 5.4.3.2),4,5 β-lysine 5,6- (EC
5.4.3.3),6,7

D-lysine

5,6- (EC 5.4.3.4),6,7

D-ornithine

4,5- (EC 5.4.3.5),8 tyrosine 2,3- (EC

5.4.3.6),9,10 leucine 2,3- (EC 5.4.3.7),11 glutamate-1-semialdehyde 2,1- (EC 5.4.3.8),12,13
11

glutamate 2,3- (EC 5.4.3.9),14 and phenylalanine- (EC 5.4.3.1015 and EC 5.4.3.1116)
aminomutase enzymes. Additionally, a very recently discovered 2-aza-L-tyrosine aminomutase
catalyzes a 2,3-isomerization reaction.17

2.1.2. Mechanistic Diversity of Aminomutases
Aminomutases use various cofactors to catalyze the challenging vicinal exchange of the
amino group and a proton. The aminomutase reaction bifurcate into homolytic and heterolytic
mechanistic classes. The mechanisms of lysine 2,3-,4,5 β-lysine 5,6-,6,7

D-lysine

5,6-,6,7

D-

ornithine 4,5-,8 leucine 2,3-,11 and glutamate 2,3-aminomutases14 involve radical intermediates.
These enzymes use either S-adenosyl methionine (SAM), pyridoxal 5'-phosphate (PLP) and a
[4Fe-4S]+ cluster or adenosylcobalamin (vitamin B12) and PLP as cofactors. Tyrosine-9,10 and
phenylalanine-aminomutases15,16 catalyze their reactions by a heterolytic isomerization
mechanism that requires the 4-methylidene-1H-imidazol-5(4H)-one (MIO) prosthetic group.

2.1.3. Homolytic Mechanisms of Aminomutases
2.1.3.1. Adenosylcobalamin-dependent Mechanisms
Adenosylcobalamin-dependent aminomutases rely on coenzyme B12 and PLP cofactors
are mainly involved in catabolic pathways of amino acids. For example, lysine 5,6-6,7 and
ornithine 4,5-aminomutase8,18 isolated from Clostridium sticklandii participates in fermentation
of DL-lysine and L-ornithine, respectively.
12

Lysine 5,6-aminomutases catalyze the migration of the -amino group of either D-lysine
or

L-β-lysine

to the δ-carbon, respectively producing 2,5-diminohexanoate and (3S,5S)-

diaminohexanoate.7 Similarly, D-ornithine aminomutase catalyzes the conversion of D-ornithine
to (2R,4S)-2,4-diaminopentanoic acid by exchanging the δ-amino group and a γ-proton.8 These
enzymes follow a radical-mediated mechanism using adenosylcobalamin as the source of
radicals (Figure 2.2). Binding of the substrate to the enzyme induces the homolytic cleavage of
the weak Co-C bond in adenosylcobalamin, generating cob(II)alamin and 5'-deoxyadenosyl
radical (Ado-•CH2) (Figure 2.2).7,8

Figure 2.2. Generation of the 5'-deoxyadenosyl radical (Ado-•CH2) species from the homolytic Co-C
bond cleavage of adenosylcobalamin.

During the isomerization reactions (Figure 2.3), PLP cofactor covalently bound as a
schiff base to an active site lysine residue (1) makes an imine (2) with the /δ-amino group of the
substrate via a transaminase reaction. Then the highly reactive Ado-•CH2 species abstracts a
13

hydrogen from the PLP-bound substrate producing 5'-deoxyadenosine (Ado-CH3) and a carboncentered substrate radical intermediate 3. The PLP-bound substrate radical in turn rearranges to
the product radical 5, which regains a proton from Ado-CH3 generating the PLP-bound product
6. Finally, the isomerized amino acid 7 is released while PLP rebinds with the enzymatic lysine
residue as a schiff base. Investigation of the steric course of β-lysine 5,6-aminomutase indicated
that the overall substitution at δ-C occurs with inversion of configuration.19,20 However,
stereochemical control of other enzymes of this class has not been evaluated so far.

Figure 2.3. Mechanism of isomerization reaction catalyzed by adenosylcobalamin-dependent
aminomutases. 5'-Deoxyadenosyl radical (Ado-•CH2) formed from adenosylcobalamin and PLP cofactor
assists the chemically challenging 1,2-amino shift.

14

2.1.3.2. S-Adenosyl Methionine-dependent Mechanisms
The reactions catalyzed by S-adenosyl methionine (SAM)-dependent isomerases are
closely related to the coenzyme B12-dependent rearrangements.21 The [4Fe-4S]+ cluster functions
analogous to adenosylcobalamin and mediates the formation of highly reactive 5'-deoxyadenosyl
radical from SAM (Figure 2.4A).21 Furthermore, similar to adenosylcobalamin-dependent
aminomutases, radical-stabilizing PLP serves as a cofactor in this class of enzyme catalysis.
However, in contrast to adenosylcobalamin-dependent enzymes, these enzymes catalyze the α- to
β-amino acid conversions. For example, lysine 2,3-, and glutamate 2,3-aminomutases isomerize
(2S)-α-lysine and -glutamate, respectively, to their corresponding β-amino counterparts.4,14 In
anaerobic bacteria such as Clostridium subterminale and Porphyromonas gingivalis, which
utilize L-lysine for growth as a carbon and nitrogen source, lysine 2,3-aminomutase plays a role
in L-lysine metabolism.4 Furthermore, the conversions catalyzed by lysine 2,3-aminomutases
provide a pool of β-lysine for the biosynthesis of antibiotics such as streptothricin F, viomycin
and Capreomycin IB.5,22,23 A recently characterized glutamate 2,3-aminomutase found in
Clostridium difficile is believed to participate in glutamate metabolism.14 Additionally, the βglutamate likely serves as an osmolyte in bacterial and archeal species.24
The mechanism of the 1,2-amine shift catalyzed be SAM-dependent aminomutases is
similar to the adenosylcobalamin-dependent mechanism (Figure 2.4B).21 PLP reacts initially
with the α-amino group of the substrate (7) to form an external aldimine. In this mechanism,
Ado-•CH2 mediated abstraction of a prochiral proton from β-carbon of the PLP bound substrate
followed by radical rearrangements produce the product-like radical intermediate 11. A proton is

15

transferred to intermediate 11 from Ado-CH3, and then the β-amino acid 13 is released from a
transamination reaction.

A

B

Figure 2.4. Mechanism of isomerization reaction catalyzed by SAM-dependent aminomutases. A)
[4Fe-4S]+ mediated formation of Ado-•CH2 species from SAM. B) PLP-dependent mechanism of α- to βamino acid isomerization.

Although the stereochemical control of recently discovered glutamate 2,3-aminomutase is
still unknown,14 the stereochemistry of lysine 2,3-aminomutase has been studied extensively.24 In
16

Clostridium sp., the conversion of (2S)-α-lysine to (3S)-β-lysine proceeds with inversion of
configuration at both C-2 and C-3 (Figure 2.5).25 Ado-•CH2 species abstracts the pro-(3R)
hydrogen from the lysyl-side chain of the PLP bound substrate and transfers it to the pro-(2R)position of β-lysine. Furthermore, amino group transfer occurs intramolecularly to the original
position of pro-(3S)-proton in (2S)-α-lysine. In addition, the interconversion of α- to β-lysine was
shown to occur with little or no exchange of substrate hydrogens with solvent hydrogens.26

Figure 2.5. Stereochemical course of lysine 2,3-aminomutase from Clostridium sp.

Lysine 2,3-aminomutase (LAM) from E. coli produces (3R)-β-lysine which is the
enantiomer produced by the analogous enzymes from Clostridium and Bacillus sp.4 Most of the
active site residues that make contacts with the substrate are conserved in both enzymes, except
the Asp330 in the Clostridial lysine aminomutase.4 The E. coli homologue has a glutamate
residue in the corresponding position, and these residues make ionic contacts with the -amino
group of the substrate. It was suggested that the steric bulk of the glutamate residue in the E. coli
LAM forces the lysyl-side chain of the substrate to bind in a more restricted conformation. The
variant configuration of the lysyl-side chain allows the abstraction of the pro-(3R) hydrogen
from (2S)-α-lysine.4 Furthermore, This configuration leads to an (R)-configuration in the
azacyclopropylcarbinyl radical (10) and in the β-lysine product (Figure 2.4).

17

2.1.4. Heterolytic Mechanisms of Aminomutases: MIO-dependent Mechanisms
In contrast to other aminomutases, recently characterized MIO-dependent aminomutases
catalyze a heterolytic isomerization mechanism.9,15 In 1998, Walker and coworkers reported an
unusual aminomutase from Taxus brevifolia, which was the first aminomutase found in a higher
plant.27 This enzyme is the first ever reported example of a phenylalanine aminomutase, and it
was proposed to catalyze the production of (3R)-β-phenylalanine for the biosynthesis of anticancer drug Taxol.27 Furthermore, the reaction was shown to proceed with intramolecular amino
group migration with retention of configuration at the β-carbon. The retention-of-configuration
observed for this enzyme was distinctly different from inversion-of-configuration of all other
microbial aminomutases characterized at that time. However, the MIO-dependency of
phenylalanine aminomutases was not established from the assays carried out with crude cell-free
extracts of T. brevifolia bark.27
After a span of five years, in 2003, Christenson et. al. discovered the novel MIOdependent class of aminomutases, which are principally homologous to a family of ammonia
lyases.9 They identified a tyrosine aminomutase (TAM) form Streptomyces globisporos
(SgTAM) that is homologous to MIO-dependent Streptomyces griseus histidine ammonia lyase28
(39% identity and 56% similarity) and Streptomyces maritimus phenylalanine ammonia lyase29
(38% identity and 56% similarity). The MIO cofactor in ammonia lyases was previously
identified in the X-ray structure of a histidine ammonia lyase (EC 4.3.1.3).30 This cofactor was
proposed to form autocatalytically by cyclization three active-site residues (Ala-Ser-Gly)
followed by dehydration (Figure 2.6). The Taxus phenylalanine aminomutase and tyrosine
aminomutase consist of an MIO group formed from the signature Ala-Ser-Gly motif.9,15 By

18

contrast, the MIO moiety of a recently characterized phenylalanine aminomutase from Pantoea
agglomerans is formed from a Thr-Ser-Gly sequence.16

Figure 2.6. Mechanism of MIO cofactor formation from active site residues Ala/Thr-Ser-Gly.

Pre-incubation of SgTAM with NaBH4 and KCN abolished the activity by presumed
nucleophilic inactivation of the MIO cofactor.9 In addition, the Ser153Ala mutation of Ala-SerGly catalytic triad decreased the activity of SgTAM 340-fold compared to the wild-type
enzyme.9 Taken together, these results strongly suggested the dependency of SgTAM catalysis
on the MIO cofactor.
MIO-dependent aminomutases occur on various biosynthetic pathways and catalyze the
2,3-amino group migration to make non-proteinogenic β-amino acids from the corresponding
(2S)-α-aromatic amino acid (Table 2.1). Phenylalanine 2,3-aminomutases from Taxus sp.
(TcPAM) and Pantoea agglomerans bacteria (PaPAM) catalyze the production of βphenylalanine in Taxol and antibiotic andrimid biosynthesis, respectively.16,20 In addition to
SgTAM on the enediyne antitumor antibiotic C-1027 biosynthetic pathway, another TAM from
Chondromyces cracatus (CcTAM) is involved in the production of β-tyrosine for cytotixic
chondramides biosynthesis.31 Furthermore, MdpC4 TAM from Actinomadura madurae, MfTAM
19

from Myxococcus fulvus, and MxTAM from Myxococcus sp. Mx-BO have also been
characterized.10,32 A recently discovered KedY4 aminomutase on the antitumor antibiotic
kedarcidin biosynthetic pathway in Streptoalloteichus sp. now increases the breadth of MIOdependent aminomutases.17 The latter aminomutase stereospecifically catalyzes the conversion
of 2-aza-L-tyrosine to (R)-2-aza-β-tyrosine and is the first MIO-dependant aminomutase found to
accept a natural product heteroaromatic substrate.17
The MIO cofactor is believed to serve as the electrophile in the isomerization reactions
performed by this class of enzymes.30 However, an active debate for decades centers on whether
the amino group or the aromatic ring of the substrate acts as the nucleophile. Two mechanisms
have been proposed for MIO-catalyzed isomerization of α- to β-amino acids (Figure 2.1).33

A

B

Figure 2.7. Mechanistic proposals for MIO-dependant aminomutases. A) Amino-alkylation and. B)
Friedel-Crafts aryl-alkylation pathway.

20

In the amino-alkylation mechanism, the electrophilic methylidene of the MIO cofactor
was postulated to react with the α-amino group of the substrate (Figure 2.7A). The second
proposed mechanism suggests π-electrons of the aromatic ring act as the nucleophile in an
Friedel-Crafts-type addition to attack the ortho-carbon to the MIO (Figure 2.7B). Both
mechanisms are proposed to reduce the pKa of the β-hydrogens and facilitate their removal by an
enzymatic base (Tyr).33 Subsequent elimination of ammonia generates an aryl acrylate
intermediate that is occasionally released, yet also serves as a scaffold for stereospecific
amination at the β-carbon to produce the corresponding β-amino acid.
The stereochemical course of tyrosine- (TAMs) and phenylalanine-aminomutases
(PAMs) are extensively studied (Table 2.1), and these enzymes can be categorized based on the
enantioselectivity for the β-amino acid. SgTAM9 and PaPAM16 produce (3S)-β-arylalanines
while TcPAM20 and CcTAM31 make (3R)- β-isomers. However, TAMs are less enantioselective
compared to PAMs. Both TAMs produce one enantiomer under kinetic control, but the opposite
stereoisomer is also produced upon prolonged incubation of the enzyme with the substrate.10,33
For SgTAM, (3R)-β-tyrosine is produced more slowly, but eventually reaches a 1:1 ratio with the
(3S)-isomer due to the β-tyrosine racemase activity of TAMs.33 By comparison, CcTAM
produces 85% of (3R)-β-tyrosine and 15% of (3S)-β-tyrosine at pH 8.5, and this ratio varies with
the pH.34 Consequently, TAMs catalyze the isomerization reaction with both inversion and
retention of configuration at the α- and β-carbons.

21

Table 2.1. Stereochemistry of β-arylalanine products
aminomutases and their corresponding biosynthetic products

Enzyme

Substrate

produced

Product

Tyrosine Aminomutase
(Streptomyces globisporus)

from

Configuration
at C-α & C-β

MIO-dependent

Biosynthetic
product

Inversion
Antitumor
antibiotic C-1027
1:1 (3S):(3R)-βtyrosine at equilibrium
conditions

Inversion and
retention

Tyrosine Aminomutase
(Chondromyces cracatus)
Retention
Cytotoxic
chondramides
6:1 (3R):(3S)-βtyrosine at equilibrium
conditions

Inversion and
retention

Tyrosine Aminomutase
(Actinomadura madurae)

Antitumor
antibiotic
maduropeptin

Tyrosine Aminomutase
(Myxococcus sp.)

Unknown

Bacterial protein
biosynthesis
inhibitor
myxovalargin

Retention

Anti-cancer Taxol

Unknown

Antibiotic
andrimid

Unknown

Antitumor
antibiotic
kedarcidin

Phenylalanine Aminomutase
(Taxus sp.)

Phenylalanine Aminomutase
(Pantoea agglomerans)

2-Aza-tyrosine Aminomutase
(Streptoalloteichus sp.)

22

Interestingly, isomerization reactions of PAMs are highly enantioselective (>99.9%) and
preclude enantiomeric mixtures. Thus, PAMs are potential sources for the scalable biocatalytic
production of enantiomerically pure β-arylalanines as synthetic building blocks of
pharmaceuticals.35 The formation of (3R)-β-phenylalanine catalyzed by TcPAM occurs with
intramolecular exchange of the pro-(3S)-proton and the α-amino group with the retention-ofconfiguration at both reaction termini.15,20 Consequently, this process demands the removal and
reattachment of the migrating partners from the opposite stereo-faces of the cinnamic acid
intermediate. The homologous PaPAM enzyme makes the opposite β-phenylalanine
stereoisomer and its stereochemical course is described herein.

2.1.5. Other Aminomutases
Few initial reports were published about the occurrence of a coenzyme B12-dependent
leucine 2,3-aminomutase involved in leucine catabolism in various organisms including animals,
plants, and microorganisms.36-38 Freer and co-workers in 1981 reported on an Andrographis
panicdata leucine 2,3-aminomutase that converted (2S)-α- to (3R)-β-leucine, yet coenzyme B12
was not required.11 In addition, the existence of leucine 2,3-aminomutase in mammalian tissues
has been questioned since β-leucine has not yet detected in rat liver or human blood.39 Further,
there are no detailed studies describing the mechanism and stereochemistry of the characterized
leucine 2,3-aminomutase.
In 1980, Parry and coworkers suggested the occurrence of an ATP-dependent Bacillus
brevis Vm4 tyrosine 2,3-aminomutase (BbTAM) in biosynthetic pathway to the antibiotic

23

edeine-A and -B.40 From a recent (2004) personal communication with Parry, this original
obeservation was likely an error (K.D.W.). However, this investigation did substantiate that
BbTAM does not require PLP, S-adenosylmethionine, and adenosylcobalamin as a cofactor, and
that the enzyme was inhibited with reagents that react with carbonyl groups.40 While the exact
carbonyl reacting reagents are not described, this data suggest the dependency of BbTAM on an
MIO, which bears an α,β-unsaturated carbonyl group that reacts with reducing agents (NaBH4)
and nucleophiles (KCN).9,20 Detailed mechanistic analyses carried out with cultures of B. brevis
suggested that (3S)-β-tyrosine was formed with the inversion-of-configuration at Cβ after the
pro-(3S) proton is removed.40 However, the configuration of Cα was not analyzed in this study
and no further investigations are reported.
The mechanism and stereochemistry of the (3S)-β-arginine made on the biosynthetic
pathway of antibiotic blasticidin S biosynthesis was examined by feeding variously labeled αarginines to whole-cell cultures of Streptomyces griseochromogenes.41 The aminomutase
reaction was observed to proceed with inversion-of-configuration at the reaction termini. The αamino group was transferred intramolecularly to the β-carbon of the substrate. This arginine 2,3aminomutase is 48% identical and 65% similar to C. subterminale lysine aminomutase, and
includes the conserved lysine for PLP attachment;42 however, the cofactor requirement of this
enzyme could not assessed in this in vivo experiments. But based on sequence similarity alone,
the arginine aminomutase likely catalyzes the isomerization reaction via a radical-dependent
homolytic mechanism. Unfortunately, recent efforts on detecting the aminomutase activity of
arginine 2,3-aminomutases from E. coli and Streptomyces lividans were not successful.42

24

Glutamate 1-semialdehyde 2,1-aminomutase uses only PLP as a cofactor for the amino
group shift, without the involvement of free radical intermediates.12 In cyanobacteria, green and
purple sulphur bacteria, and higher plants, glutamate 1-semialdehyde 2,1-aminomutase catalyzes
the conversion of glutamate 1-semialdehyde to δ-aminolevulinic acid (Figure 2.8), the universal
precursor for the biosynthesis of heme, chlorophyll, and other tetrapyrroles.12,13 However, these
enzymes demonstrate a clear sequence similarity to aminotransferases of the α-family of B6dependent enzymes, and mechanistically act as an aminotransferase rather than an
aminomutase.43

Figure 2.8. Reaction catalyzed by glutamate 1-semialdehyde 2,1-aminomutase.

25

2.2. Experimental

2.2.1. Chemicals and Reagents
Unlabeled (2R)- and (2S)-α-phenylalanines were obtained from Sigma-Aldrich; [3,32

H2]-(2S)-α-phenylalanine,

[U-13C,

2-15N]-(2S)-α-phenylalanine,

[ring,

β-C-2H6]-trans-

cinnamate, and [2-15N]-(2S)-α-phenylalanine were purchased from Cambridge Isotope
Laboratories (Andover, MD). ortho-Methyl-(2S)-α-phenylalanine was obtained from Peptech
(Burlington, MA). [2,3-2H2]-N- Acetyl (2S,3R)/(2R,3S)-β-phenylalanine methyl ester, [ring, 2,32

H7]-(2S,3S)-α-phenylalanine (90% ee, 98+% enriched), and racemate [ring, 3-2H6]-

(2R,3S)/(2S,3R)-α-phenylalanine (98+% enriched) were sample stocks from the Walker
laboratory that had been synthesized according to reported literature20,44 and were characterized
as described below. All other chemicals and reagents were obtained from Sigma-Aldrich and
utilized without further purification, unless noted otherwise.

2.2.2. Instrumentation
A gas chromatograph (model 6890N, Agilent) coupled to a mass selective detector
(model 5973 inert®, Agilent) was used for the analysis of derivatized amino acids. Samples were
loaded onto a 5HS GC column (0.25-mm inner diameter × 30 m, 0.25-μm film thickness)
mounted in a GC oven. The GC oven conditions were as follows: column temperature was
programmed from 70 C to 250 C or 300 °C at 10 °C/min (23 min and 18 min total runtime,

26

respectively), splitless injection was selected, and helium was used as the carrier gas (1.2
ml/min). The MS conditions were set with an ion scan mode from 100 - 400 atomic mass units at
70 eV ionization voltage. All 1H- (500 MHz) and 2H-NMR (76.7 MHz) spectra were obtained on
a Varian superconducting NMR-Spectrometer using standard acquisition parameters.

2.2.3. Characterization of Racemate [ring, 3-2H6]-(2R,3S)/(2S,3R)-α-Phenylalanine and
[ring, 2,3-2H7]-(2S,3S)-α-Phenylalanine
2.2.3.1. GC/EI-MS Analysis
The racemate and the enantiopure amino acid (~0.25 mg) were separately dissolved in 50
mM phosphate buffer (pH 8.5) and adjusted to pH > 10 (6 M NaOH). Camphanoyl chloride (0.5
mg) was added to derivatize the amino acids to their N-[(1'S)-camphanoyl] analogues. The
solution was acidified to pH 2 (6 M HCl) and extracted with ethyl acetate, the organic fractions
were combined (2 mL), to which methanol (700 μL) was added. A solution TMS-diazomethane
was added dropwise to the ethyl acetate/methanol solution to convert the acids to their methyl
esters. Each sample was separately analyzed by GC/EI-MS. (Figure A.1.1).
2.2.3.2. 1H- and 2H-NMR Analyses
The racemic mixture and the enantiopure amino acid (~0.5 mg of each, as their free
amino acids) were separately dissolved in H2O (600 μL) and D2O (600 μL), and analyzed by 2HNMR (3,500 scans) and 1H-NMR (200 scans), respectively (Figure A.1.2 and Figure A.1.3,
respectively).

27

2.2.3.3. Analysis of Stereochemistry by the Phenylalanine Ammonia Lyase (PAL) Reaction
Stereospecificity
Phenylalanine ammonia lyase (Sigma-Aldrich, St. Louis, MO; PAL) catalyzes the nonoxidative deamination of (2S)-α-phenylalanine to trans-cinnamic acid and ammonia. This
deamination mechanism was shown to proceed stereospecifically with removal of the pro-(3S)
proton and the ammonia from the substrate.45 PAL was used to assess the stereochemistry at Cβ
of racemate [ring, 3-2H6]-(2R,3S)/(2S,3R)-α-phenylalanine and the enantiopure [ring, 2,3-2H7](2S,3S)-α-phenylalanine. Each compound at ~3 mM in 1 mL of 50 mM phosphate buffer (pH
7.5), were incubated with PAL (20 μg; from Rhodotorula glutinis; Sigma-Aldrich) for 1 h at 31
°C. The reactions were acidified to pH 2 (6 M HCl), extracted with ethyl acetate (2 × 1 mL), and
the organic fractions were combined. Methanol (700 µL) was added to the ethyl acetate, and a
TMS-diazomethane solution was added dropwise to methyl esterify the carboxylic acids. The
cinnamic acid methyl esters were analyzed by GC/EI-MS (Figure A.1.4 and Figure A.1.5).

2.2.4. Protein Expression and Purification
E. coli (BL21) cells transformed to express the papam gene were grown in six 1-L
cultures of Luria-Bertani medium. The media was supplemented with kanamycin (50 μg/mL) at
37 °C and grown until the cell density reached an optical density of A600 ~ 0.6, after which
protein expression was induced with isopropyl-β-D-thiogalactopyranoside (100 μM) at 16 °C.
After 16 h, the cells were harvested by centrifugation at 6,000g (15 min) and the cell pellet was
resuspended in lysis buffer (50 mM sodium phosphate containing 5% (v/v) glycerol, 300 mM

28

NaCl and 10 mM imidazole, pH 8.0), and the suspension was lysed by sonication. The cellular
debris and light membranes were removed by centrifugation at 9,700g (45 min), and 102,000g
(1h), respectively, to provide soluble enzyme extract.
The resultant crude aminomutase in the soluble fraction was purified by Nickelnitrilotriacetic acid (Ni-NTA) affinity chromatography according to the protocol described by the
manufacturer (Qiagen, Valencia, CA). PaPAM fractions eluted in 250 mM imidazole were
combined and loaded into a size selective centrifugal filtration unit (Centriprep centrifugal filter
units, 30,000 MWCO; Millipore). Then the protein solution was concentrated and the buffer was
exchanged with 50 mM sodium phosphate containing 5% (v/v) glycerol (pH 8.0) using the same
Centriprep centrifugal filter unit. The purity (>95%) of the concentrated enzyme was assessed by
SDS−PAGE with Coomassie Blue staining, and the quantity was determined by the Bradford
protein assay.

2.2.5. Assessment of Absolute Stereochemistry of the β-Phenylalanine Product Produced by
PaPAM
(2S)-α-Phenylalanine (500 μM) was incubated with PaPAM (100 μg) in 1.0 mL of
phosphate buffer (pH 8.0) at 31 °C. After 3 h, the reaction was terminated by basification to pH
10 (6 M NaOH). (1S)-(−)-camphanic chloride (0.5 mg) was then added and allowed to react for
30 min. After derivatization, the mixture was acidified to pH 2-3 (6 M HCl) and extracted with
ethyl acetate (2 x 1 mL). Then the organic solvent was evaporated under vacuum, the residue
was dissolved in ethyl acetate/methanol (3:1, v/v) (200 μL), and the solution was treated with

29

excess (trimethylsilyl)diazomethane until a persistent faint yellow color persisted. The
derivatized β-amino acid was identified by GC/EI-MS analysis and compared against the
retention time and mass spectral fragmentations of authentic N-[(1'S)-camphanoyl]-(3S)-βphenylalanine methyl ester derivatized similarly.

2.2.6. Enzymatic assays
α-Phenylalanine isotopes (each at 500 μM) were incubated with PaPAM (50 μg) in 1-mL
assays for 1 h. When applicable, [ring, β-C-2H6]-(E)-cinnamic acid (500 μM) was added to the
reaction in competitive assays. In each experiment, control experiments were run in parallel by
incubating substrate without enzyme; enzyme without substrate; and a single substrate in
competitive assays. After 1 h, the reaction was terminated by basification to pH 10 (6 M NaOH),
and the resulting amino acids were derivatized for analysis by GC/EI-MS.

2.2.7. Derivatization and Analysis of Amino Acids
The amino acids in all assays were generally derivatized as follows: To each terminated
assay at pH 10, benzoyl chloride (50 μl) was added, and the solution was thoroughly stirred.
After 10 min, the solution was again basified to pH 10 (6 M NaOH) and a second batch of
benzoyl chloride (50 μl) was added, stirred and allowed to react for 10 min. Then the mixture
was acidified to pH 2-3 (6 M HCl) and extracted with ethyl acetate (2 x 1 mL). The organic
solvent was evaporated under vacuum, and the resulting residue was dissolved in ethyl

30

acetate/methanol

(3:1,

v/v)

(200

μL).

The

solution

was

treated

with

excess

(trimethylsilyl)diazomethane until a faint yellow color persisted to convert the derivatized amino
acids to their methyl esters. The derivatized β-amino acids were analyzed and identified by
GC/EI-MS analysis.

2.2.8. Determining the Stereochemistry of Hydrogen Rebound
[3,3-2H2]-α-Phenylalanine (5 mM) was incubated with PaPAM (2 mg) in 2.5 mL of
phosphate buffer (pH 8.0) at 31 °C for 2 h. The amino acid isotopomers were derivatized as
described before except the N-acetyl methyl esters were made using acetic anhydride instead of
benzoyl chloride. The mixture of the derivatized deuterium-labeled α- and β-amino acids from
the assay was dissolved in CHCl3 (600 μL) and analyzed by 2H-NMR (18,700 scans). Similarly,
[3,3-2H2]-α-phenylalanine and a mixture of unlabeled (2S)-α- and (3S)-β-phenylalanine isolated
from phosphate buffer (pH 8.0) were derivatized as their N-acetyl methyl esters and analyzed by
2

H-NMR (76.7 MHz, CHCl3) and 1H-NMR, respectively (500 MHz, CDCl3). A 2H-NMR

spectrum of synthetically-derived authentic standard [2,3-2H2]-N-acetyl (2S,3R)/(2R,3S)-βphenylalanine methyl ester was used for reference.

2.2.9. Kinetic Parameters of PaPAM with o-Methyl-(2S)-α-Phenylalanine
PaPAM (10 μg) was incubated with o-methyl-(2S)-α-phenylalanine (1 mM) to establish
steady-state conditions with respect to a fixed protein concentration and time at 31°C. Under

31

steady-state conditions, the substrate at 10, 20, 40, 80, 150, 300, 500, 750 and 1000 μM were
separately incubated with enzyme (10 μg) for 1 h in single stopped-time assays. Then the omethylcinnamic acid product was derivatized as the methyl ester and the amino acids were
derivatized as the N-(ethoxycarbonyl) methyl ester as described previously using ethyl
chloroformate and (trimethylsilyl)diazomethane. Derivatized products were then quantified by
GC/EI-MS analysis. The kinetic parameters (KM and kcat) were determined from the HanesWoolf plot (R2 ~0.99).

32

2.3. Results and Discussion

2.3.1. Absolute Configuration of the Biosynthetic β-phenylalanine
Absolute configuration of biosynthetic β-phenylalanine derived from PaPAM was
assigned by incubating PaPAM with (2S)-α-phenylalanine under standard assay conditions and
converting the products to N-(1(S)-camphanoyl) methyl esters for subsequent GC/EI-MS
analysis. The derivatized β-amino acid showed an identical retention time (16.02 min) and
corresponding fragments ions to the authentic N-[(1’S)-camphanoyl]-(3S)-β-phenylalanine
methyl ester (Figure 2.9). There was no detectable product at 15.86 min corresponding to the
retention time of authentic N-[(1’S)-camphanoyl]-(3R)-β-phenylalanine methyl ester. Therefore
the absolute configuration of the biosynthetic product was established as (3S)-β-phenylalanine.

2.3.2. Mechanism of Amino Transfer
Based on the reaction of the homologous PAM from Taxus plants, (E)-cinnamate is a
proposed intermediate of the forward reaction during isomerization of α- to β-phenylalanine.46 In
addition, (E)-cinnamate is reported as a substrate used to make a mixture of α- and βphenylalanines in presence of an amino source (6 M NH3 and styrylalanine).47,48 Thus, to
evaluate whether PaPAM could transfer the amino group from phenylalanine to an exogenous
(E)-cinnamate, a mixture of [ring, β-C-2H6]-(E)-cinnamate (98+% 2H-enriched, 500 μmol) and
[15N]-(2S)-α-phenylalanine (98+%

15

N-enriched, 500 μmol) were incubated with PaPAM under

33

standard assay conditions. Resulting product mixtures were derivatized as N-benzoyl methyl
esters and analyzed by GC/EI-MS. For quantitative analysis of isotopic enrichment and
distribution, authentic standard (3S)-β-phenylalanine was also derivatized as N-benzoyl methyl
ester and analyzed by GC/EI-MS. Authentic standard was comprised of low abundance
molecular ion ([M]+) of m/z = 283 and two diagnostic mass fragments ions at m/z = 210 {[M −
CH2C(O)OCH3]+} and m/z = 178 {[M − PhC(O)]+} (Figure 2.10 and Table 2.2A).

A

B
178

100
16.02

0.8

Relative Abundance

Relative Total Ion Count

1.0

0.6
15.86

0.4
0.2

15.78

80
60
40
20

121

286
146162

0.0
15

16

Time (min)

17

18

0
100

150

328 359

200

250

300

350

400

m/z

Figure 2.9 Stereochemical determination of biosynthetic β-phenylalanine. (A) Overlay of gas
chromatography profiles of N-[(1'S)-camphanoyl] methyl esters of authentic racemic (3R)- (15.86 min)
and (3S)-β-phenylalanine (16.02 min) (solid red line), and of the N-[(1'S)-camphanoyl] methyl ester of
biosynthetic (3S)-β-phenylalanine (16.02 min) derived from PaPAM catalysis (dashed line). The latter
chromatogram shows the N-[(1'S)-camphanoyl] methyl ester of (2S)-α-phenylalanine (15.78 min) used as
the substrate. (B) The mass spectrometry profile of the N-[(1'S)-camphanoyl] methyl ester of (3S)-βphenylalanine (16.02 min) derived from PaPAM catalysis.

34

Mass spectral analysis of products formed from the above assay with [ring, β-C-2H6]-(E)cinnamate and [15N]-(2S)-α-phenylalanine showed a molecular ion {[(3-15N)-M]+} of m/z = 284
and two diagnostic mass fragments ions at m/z = 211 {[(3-15N) − CH2C(O)OCH3]+} and m/z =
179 {[(3-15N) − PhC(O)]+} (Table 2.2B). No isotopic enrichment was observed for these mass
fragments compared to the authentic standard, which suggests that the additional mass unit was
derived from the

15

N-atom. Furthermore, there was no molecular ion at m/z = 290 nor a base

peak at m/z = 185 with additional seven mass units, which indicates that no biosynthetically
derived β-phenylalanine from [ring, β-C-2H6]-(E)-cinnamate is present in the product mixture.
This shows that PaPAM doesn’t transfer the amino group to an exogenous (E)-cinnamate, a
proposed intermediate of this isomerization reaction,16 but it transfers the amino group
intramolecularly to the same carbon skeleton to afford the isomerized amino acid.
Further proof for the intramolecular amino transfer from Cα to Cβ was provided by
incubating PaPAM with a mixture of [U-13C9, 2-15N]-(2S)-α-phenylalanine and unlabeled 2S-αphenylalanine and analyzing the derivatized amino acids by GC/EI-MS. Mass spectral analysis
of resulting β-phenylalanines showed molecular ions at m/z = 283 and 293 indicating the
presence of unlabeled and [13C9, 3-15N]-(3S)-β-phenylalanine (with ten additional mass units)
respectively. This was further confirmed by the presence of diagnostic mass fragments at m/z =
210 {unlabeled, [M – CH2C(O)OCH3]+}, m/z = 218 {[(U-13C9, 3-15N)M – CH2C(O)OCH3]+} and
base peaks at m/z 178 {unlabeled, [M – PhC(O)]+} and m/z 188 [(U-13C9, 3-15N)M – PhC(O)]+
(Table 2.2C). In addition, calculated ratios of ion abundance of m/z 178:m/z 179 (P+1) and m/z
188:187 (P−1) for derivatized biosynthetic amino acids were identical to that of authentic
standard. The lack of enrichment for m/z 179 and m/z 187 in above mixed substrate assay

35

compared to authentic standards suggests that that intermolecular amino transfers did not occur,
which further suggests an intramolecular amino transfer during PaPAM aminomutase catalysis.

178

Relative Ion Abundance (%)

100

80

105

60

40

20
131

210

146

222

252

0
100

150

200

250

283
300

m/z

Figure 2.10. Diagnostic mass spectral fragments and mass spectrometric profile of authentic Nbenzoyl-(3S)-β-phenylalanine methyl ester. Fragment ions m/z 178 [M – PhC(O)]+ (base peak) and m/z
210 [M – CH2C(O)OCH3]+ were diagnostic and present at 100% and 10% abundance, respectively.

2.3.3. Stereochemistry of Amino Transfer
Deuterium labeled α-amino acids, [ring, 2,3-2H7]-(2S,3S)-α-phenylalanine (90+% ee,
98+% 2H-enriched) and racemate [ring, 3-2H6]-(2R,3S)/(2S,3R)-α-phenylalanine (98+% 2Henriched) were employed to assess the stereochemical course of this isomerization reaction.
Stereospecifically deuterium labeled α-amino acids were separately incubated with PaPAM
under standard assay conditions and the resulting β-amino acids were converted to their N36

benzoyl methyl esters for subsequent GC/EI-MS analysis. The diagnostic mass fragments ions at
m/z = 210 {[M − CH2C(O)OCH3]+} and m/z = 178 {[M − PhC(O)]+} obtained for the authentic
unlabeled (2S)-α-phenylalanine was used to identify the fate of deuteriums during the
isomerization reaction.
Mass spectral analysis of the derivatized biosynthetic β-phenylalanine obtained from
[ring, 2,3-2H7]-(2S,3S)-α-phenylalanine showed diagnostic mass fragments (Table 2.2E) at m/z =
215 {[(ring-2H5)M – CH2C(O)OCH3]+} with no m/z = 216 enrichment compared to authentic
standard indicating that no deuterium retained on Cβ. The base peak of the product with m/z =
185 (no m/z = 184 enrichment), consisted of seven additional mass units (seven deuteriums)
compared to the unlabeled authentic standard, thus confirming that the deuterium removed from
β-carbon migrated to Cα during the isomerization.
Incubation of PaPAM with (2S)-α-phenylalanine and (2R)-α-phenylalanine separately
indicated that only the 2S-isomer is a productive substrate for this aminomutase. Thus, the [ring,
3-2H6]-(2S,3R)/(2R,3S)-α-phenylalanine racemate contains only [ring, 3-2H6]-(2S,3R)-αphenylalanine as a productive substrate. Derivatized β-phenylalanine obtained after incubation of
[ring, 3-2H6]-(2S,3R)/(2R,3S)-α-phenylalanine racemate showed a mass fragment [(2H1)M –
CH2C(O)OCH3]+ at m/z 216 (no m/z 215 enrichment) and a base peak [(2H1)M – PhC(O)]+ at m/z
184 (no m/z 183 enrichment) (Table 2.2D) indicating that the deuterium atom at β-position is
retained during the enzymatic reaction.
These results confirm that pro-(3S) hydrogen is removed and transferred intramolecularly
to the Cα during the isomerization reaction while the pro-(3R) Hydrogen is retained at the βposition. This observation coupled with the known 3S stereochemistry of the β-phenylalanine
37

product,16 provides support for an inversion-of-configuration at Cβ where the amino group
reattaches after removal from the Cα. That is, after the pro-(3S) hydrogen is removed from βposition, amino group attaches to the position of pro-(3R) hydrogen and moves the latter to the
position originally occupied by pro-(3S) hydrogen.
Table 2.2. Mass spectral fragment ions of isotopomers of N-benzoyl-β-phenylalanine methyl esters

Biosynthetic products, and diagnostic mass
fragments 'a' and 'b' in GC/EI-MS

Substrate(s) incubated with PaPAM

A
'a'

m/z = 178
X, Y, Z = H

'b'

m/z = 210
X=H

'a'

m/z = 179 (15N)
X, Y, Z = H

'b'

m/z = 211 (15N)
X=H

'a'

m/z = 178
X, Y, Z = H
m/z = 188 (13C9, 15N)
X, Y, Z = H

'b'

m/z = 210
X=H
m/z = 218 (13C7, 15N)
X=H

'a'

m/z = 184 (D6)
X = D; Y, Z = H

'b'

m/z = 216
X=D

'a'

m/z = 185 (D7)
X = H; Y, Z = D

'b'

m/z = 215
X=H

B

C

D

E

38

2.3.4. Stereochemistry of Hydrogen Rebound
The stereochemistry of hydrogen rebound at Cα during the aminomutase reaction was
assessed by incubating PaPAM with [3,3-2H2]-α-phenylalanine under standard assay conditions.
The biosynthetic product was converted to the N-acetyl methyl ester and analyzed using 2HNMR and GC/EI-MS. 2H-NMR spectra of synthetic authentic racemate [2,3-2H2]-N-acetyl
(2S,3R)/(2R,3S)-β-phenylalanine methyl ester and N-acetyl [3,3-2H2]-α-phenylalanine methyl
ester, and 1H-NMR spectra of a mixture of (2S)-α- and (3S)-β-phenylalanine N-acetyl methyl
esters were also acquired for referencing the deuterium signals observed for the
substrate/biosynthetic product mixture.
Proton-decoupled 2H-NMR spectrum of synthetically-derived authentic racemate, [2,32

H2]-N-acetyl (2S,3R)/(2R,3S)-β-phenylalanine methyl ester with spectroscopically equivalent

deuteriums, showed two singlets at δ 2.83 and δ 5.40 for deuteriums at Cα and Cβ, respectively
(Figure 2.11C). The reason for observing singlets instead of doublets is due to the smaller
deuterium-deuterium spin coupling compared to hydrogen-hydrogen spin couplings (about 40
times smaller).49 Diagnostic 1H-NMR chemical shifts of unlabeled N-acetyl (3S)-β-phenylalanine
methyl ester were observed at δ 2.82 (dd) and δ 2.92 (dd) for the protons at Cα. By comparing the
Cα deuterium chemical shift (δ 2.83) of authentic product standard, the chemical shifts δ 2.82 and
δ 2.92 of N-acetyl (3S)-β-phenylalanine methyl ester were assigned for pro-(2R) and pro-(2S)
hydrogens, respectively (Figure 2.11D). Similarly, pro-(3S) and pro-(3R) hydrogens of
derivatized (2S)-α-phenylalanine showed chemical shifts at δ 3.13 and δ 3.07, respectively
(Figure 2.11D). The chemical shift resonances observed for authentic N-acetyl [3,3-2H2]-αphenylalanine methyl ester were at δ 3.10 and δ 3.08 (Figure 2.11A) and they were virtually

39

identical to the chemical shifts observed for the unlabeled N-acetyl (2S)-α-phenylalanine methyl
ester. 2H-NMR profile of the assay mixture obtained after incubating PaPAM with [3,3-2H2]-αphenylalanine (Figure 2.11B and inset) showed two unresolved singlets (δ 3.10, 3.08), which
correspond to the unreacted substrate and two singlets [δ 2.83 (HCαD), 5.40 (NCβD)], which
correspond to biosynthetic (3S)-β-phenylalanine. The chemical shift resonance at δ 2.83 of the
biosynthetic β-phenylalanine was identical to the chemical shift resonance observed for pro-(2R)
hydrogen of the N-acetyl (3S)-β-phenylalanine methyl ester. Therefore this indicates that the pro(3S) deuterium that was removed during the reaction rebinds at the α-carbon on the original
spatial position of pro-(2R) hydrogen. These NMR data coupled with the known (3S)stereochemistry16 (also found herein) of the biosynthetically derived [2H]-β-phenylalanine
product established the biosynthetic product as the (2R,3S)-enantiomer.
Further analysis of the 2H-NMR spectrum of the derivatized biosynthetic product
obtained from the above assay showed that the integral of the peak area (set to 1.0 deuterium) for
the resonance signal at δ 5.40 (CβD) and for the signal at δ 2.83 (CαD) were equal. Furthermore,
GC/EI-MS analysis of N-acetyl methyl ester of biosynthetic [2H]-labeled β-phenylalanine
showed a mass fragment ion at m/z = 180 {[PhCβD(NH)CαDHCO2CH3]+} without isotopic
enrichment for the P−1 (monodeuterated product; [PhCβD(NH)CαH2CO2CH3]+). The
biosynthesized dideuterio-β-phenylalanine isotopomer was present at ~99% compared to the
monodeuterio-β-isotopomer. These observations suggest that no hydrogen exchange occurs with
the buffer protons during the isomerization catalyzed by this enzyme. In contrast, TcPAM, which
produces 3R-β-phenylalanine, has shown significant hydrogen exchange (~60%) over 1 h20 and
this suggests that the PaPAM active site excludes water more effectively than the TcPAM active
site.
40

Pro-3S

A

B

C

pro-2R pro-2S

D

Figure 2.11. NMR spectra of N-acetyl methyl esters of various α- and β-phenylalanines. A) A partial
2
H-NMR (76.7 MHz, CHCl3) spectrum of N-acetyl methyl ester of authentic [3,3-2H2]-(2S)-αphenylalanine. B) A partial 2H-NMR spectrum of N-acetyl methyl esters of unreacted [3,3-2H2]-(2S)-αphenylalanine and biosynthetic [2H]-labeled β-phenylalanine isotopomer. C) A partial 2H-NMR (76.7
MHz, CHCl3) spectrum of the authentic racemate [2,3-2H2]-N-acetyl (2S,3R)/(2R,3S)-β-phenylalanine
methyl ester. D) A partial 1H-NMR (500 MHz, CDCl3) spectrum of a mixture of unlabeled (2S)-N-acetyl
α- and (3S)-N-acetyl β-phenylalanine methyl esters.

2.3.5. Overall Mechanism and Stereochemistry of PaPAM Isomerization
Taken together, these GC/EI-MS and 2H-NMR data show that PaPAM isomerization
reaction occurs with inversion of configuration at both Cα and Cβ (Figure 2.12). The hydrogen
formerly at the pro-(2R) position is replaced by the pro-(3S) hydrogen, while the position
formerly occupied by pro-(3R) hydrogen is replaced by the amino group (Figure 2.12).
41

According to the accepted mechanism of MIO-based aminomutases, both the amino
group and hydrogen are removed heterolytically to form trans-cinnamic acid and then the amino
group and hydrogen are rebound to afford the corresponding β-phenylalanine. The observations
provided in the foregoing discussion suggest that the PaPAM likely removes from and reattaches
the amino and the hydrogen groups to the same stereo-face of the cinnamate intermediate from
which they originated (Figure 2.12).

Figure 2.12. The stereochemistry of the isomerization reaction catalyzed by PaPAM. During the αto β-isomerization, pro-(3S) hydrogen (blue) and the amino group remove and reattach from the same
stereo-face by inverting the stereochemistry at each reaction termini.

This observation is distinctly different than the retention-of-configuration at both reaction
termini observed for TcPAM from Taxus plants.20 In this isomerization reaction catalyzed by
TcPAM, the amino group removed from the Cα must be reattached at Cβ on the opposite face to
produce (3R)-β-phenylalanine. Several hypotheses have been proposed to explain the
stereochemical control of plant ortholog TcPAM. The earliest hypothesis was that the phenyl
ring and carboxylate group of (2S)-α-phenylalanine are arranged in a syn-periplanar orientation
in which the amine group and the leaving pro-(3S) hydrogen atom positioned on the same side of
the substrate, leading to the formation of cis-cinnamic acid.20 However, it was shown later, only
the trans-cinnamic is involved in PAM catalysis.35,46 Accordingly, TcPAM requires a
reorientation of the trans-cinnamic to expose the Cα re- and Cβ si-face respectively for the
42

amino- and proton-rebound. It has been proposed that the cinnamate intermediate undergoes a
180 rotation about the C1-Cα and Cipso-Cβ bonds prior to the reattachment of the amino group
from the MIO-NH2 adduct (Figure 2.13 path b).46 This bond rotation doesn’t alter the salt bridge
to Arg-325 of TcPAM, but cause a minor displacement of the aromatic ring from its original
position without resulting steric clashes in the active site. Alternatively, an internal rotation
around the Cipso-Cβ bond to position the carboxylate group in two hydrogen bonds with
neighboring active site residues (Asn355 and Asn231) is been proposed in a recent mutational
investigation on Taxus chinensis PAM.50 However, the latter proposal is energetically
unfavorable since it requires the breakage of the high energy salt-bridge between the substrate
carboxylate group and the active site Arg.

Figure 2.13. The stereochemistry and mechanism of isomerization reactions catalyzed by PaPAM
and TcPAM. Isomerization mechanism begins with attack on the MIO by the amine at Cα of the
substrate, followed by a concerted elimination of NH3 to produce a cinnamate intermediate. PaPAM (path
a), remove and reattach the migrating proton and the amino group from the same face. In contrast,
TcPAM (path b) involves a rotation of the acrylate intermediate about the C1-Cα and Cβ-Cipso bonds, prior
to the reattachment of the migrating partners. The open circles (○) and the filled circles (●) on the
cinnamate molecule are used to distinguish the si- and re-faces of the double bond, respectively, using Cβ
as the reference.

43

Incubation of PaPAM and TcPAM separately with o-methyl-(2S)-α-phenylalanine has
provided some evidence for the proposed difference in the mechanistic pathways of these two
enzymes that produce two enantiomeric forms of β-phenylalanine. When incubated with omethyl-(2S)-α-phenylalanine, PaPAM produces exclusively o-methyl-(3S)-β-phenylalanine,
while TcPAM produces 1% of o-methyl-(3R)-β-phenylalanine and 99% of o-methyl-(E)cinnamate intermediate.51 Steric barrier due to the methyl group on the aromatic ring likely
prevents the proposed bond rotation in TcPAM, and therefore the catalytic reaction cannot
proceed beyond o-methyl-(E)-cinnamate intermediate to produce the corresponding βphenylalanine enantiomer. In contrast, PaPAM can produce o-methyl-(3S)-β-phenylalanine as it
doesn’t require any additional rotomeric forms of the intermediate. Furthermore, kinetic
parameters of PaPAM for o-methyl-(3S)-β-phenylalanine (kcat = 0.061 s−1; KM = 0.05 mM)
reveals that the catalytic efficiency (kcat/KM = 1.2 s−1•mM−1) of PaPAM is 6-fold greater than the
value reported for TcPAM (0.2 s−1•mM−1)51 likely due to the use of a single rotamer instead of
two rotamers as in TcPAM catalysis.

44

2.4. Conclusion

PaPAM catalytic mechanism likely proceeds through a stepwise mechanism where the
migratory hydrogen and amino group are eliminated heterolytically from the substrate and held
by the enzyme followed by an intramolecular transfer on to the same carbon skeleton. During the
PaPAM aminomutase catalysis, the pro-(3S) hydrogen and the α-amino group are eliminated
from the substrate to produce cinnamic acid intermediate. Then, the migrating partners exchange
their positions intramolecularly from the same stereo-face that they originated. For example, The
amino group and the pro-(3S) hydrogen reattach respectively to original positions of pro-(3R)and pro-(2R)-hydrogens of (2S)-α-phenylalanine. Consequently, the configuration of both α- and
β-carbons are inverted during the isomerization. Thus, PaPAM catalyzes the isomerization of
(2S)-α-phenylalanine to (3S)-β-phenylalanine with inversion-of-configuration at both reaction
termini.
The mechanism is clearly different from the retention of configuration observed for the
plant ortholog TcPAM, where the migrating partners exchange their positions in opposite stereofaces. Thus, the mechanistic information obtained from this study revealed two classes of
enzymes in MIO-family of PAMs. However, the exact mechanism responsible for the opposite
stereocontrol origin in these homologous PAMs is not well understood from this study.
Therefore, further future investigations are required to dissect the enantioselectivity origin of
mechanistically similar MIO-dependant aminomutases.

45

APPENDIX

46

1.6

Relative Total Ion Count

B

15.85 min

100

N-[(1'S)-camphanoyl]-(2S)- -phenylalanine methyl ester
Authentic standard

1.4
2
(2R,3S)-[ring, 3- H6]derivative of racemate

1.2

2
(2S,3R)-[ring, 3- H6]-derivative
of racemate

1.0

Relative Ion Abundance

A

15.67 min

0.8

D H

O

H D O
OCH3

0.6

HN

D5

O

OCH3
HN

D5

O

0.4

O

O

O

O

168

80

60

40

20

0.2

306

136 137
240

0.0
15

C

167

16

365

0

17

100

D

1.2

150

100

H D

0.6

HN

D5

Relative Ion Abundance

Relative Total Ion Count

0.8

O
D

OCH3
O

O

250

300

350

169

2
(2S,3S)-[ring, 2,3- H7]-Derivative

1.0

200

80
168
60

40

O

0.4
2
(2R,3R)-[ring, 2,3- H7]Derivative

0.2

20
137

307

138
241

366

0
100

15

16

Time (min)

17

150

200

250

300

350

m/z

Figure A.1.1. GC and mass spectrometry profiles of various N-[(1'S)-camphanoyl]-phenylalanine
methyl esters. A) GC profile of authentic N-[(1'S)-camphanoyl]-(2S)-α-phenylalanine methyl ester (15.85
min) (dashed red line); GC profile of N-[(1'S)-camphanoyl] methyl esters of racemate [ring, 3-2H6](2R,3S)/(2S,3R)-α-Phenylalanine, 15.67 and 15.85 min, respectively (solid line), B) and the corresponding
mass spectrometry profile. C) GC profile of the N-[(1'S)-camphanoyl] methyl ester of [ring, 2,3-2H7](2S,3S)-α-Phenylalanine (90% ee.) [synthesized as shown above with the racemate], D) and its mass
spectrometry profile.

47

A

B

Figure A.1.2. NMR spectra of the racemate [ring, 3-2H6]-(2R,3S)/(2S,3R)-α-phenylalanine. A) 2HNMR spectrum (δ 3.08, br, CβDH, D is gauche relative to amino group; 7.39, br, D5 on phenyl). B) 1HNMR spectrum (δ 3.13, d, CβHD, H is anti relative to amino group, J = 5.7 Hz; 3.87, d, CαHNH2, J = 5.7
Hz).

48

A

B

Figure A.1.3. NMR spectra of [ring, 2,3-2H7]-(2S,3S)-α-phenylalanine. A) 2H-NMR spectrum (δ 3.19,
br, CβDH, D is anti relative to amino group; 3.25, br, CαDNH2; 7.40, br, D5 on phenyl). B) 1H-NMR
spectrum (δ 2.94, s, CβH, H is gauche relative to amino group).

49

100

137

Relative Ion Abundance

80

60

109

168

40

20

0
100

120

140

160

180

200

Time (min)

Figure A.1.4. Mass spectrometry profile of cinnamic acid methyl ester derived from the racemate
[ring, 3-2H6]-(2R,3S)/(2S,3R)-α-phenylalanine by PAL catalysis. The source of the diagnostic fragment
ions are depicted in the reaction scheme.

50

120
137

Relative Ion Abundance

100

80

109

60

168

40
PAL 2sr3rs d6 Abund
20

0
100

120

140

160

180

200

Time (min)

Figure A.1.5. Mass spectrometry profile of cinnamic acid methyl ester derived from [ring, 2,3-2H7](2S,3S)-α-phenylalanine by PAL catalysis. The source of the diagnostic fragment ions are depicted in
the reaction scheme.

51

Relative Ion Abundance (%)

100

105

188

80

60
178

293

40
283

20
112

280

131

140
146 156

210

0
100

150

200

218

231
222 252 262
250

290
m/z

283

300

293
300

m/z

Figure A.1.6. Mass spectrometry profile of the derivatized biosynthetic β-phenylalanine product
isolated after co-incubating PaPAM with unlabeled and [U-13C9, 15N]-α-phenylalanine (+98%
enriched). Diagnostic ions at m/z 210 [M–CH2C(O)OCH3]+and 218 [{U-13C9, 3-15N}M–
CH2C(O)OCH3]+, and no m/z 211 [{3-15N}M–CH2C(O)OCH3]+ nor 217 [(U-13C9, 2-14N)M–
CH2C(O)OCH3]+ were present, confirming that the nitrogen migrates completely intramolecularly during
the isomerization. As further confirmation, the base peak ions at m/z 178 (no P + 1 present) and m/z 188
[{U-13C9, 3-15N}M – PhC(O)]+ (no P - 1 present) also indicated intramolecular nitrogen migration.

52

Relative Ion Abundance (%)

100

184

80
105
60

40

20
137

216
227

152

258

0
100

150

200

250

289
300

m/z

Figure A.1.7. Mass spectrometry profile of the derivatized biosynthetic β-phenylalanine product
isolated after co-incubating PaPAM with racemate [ring,3-2H6]-(2R,3S)/(2S,3R)-α-phenylalanine
(98+% enriched). Diagnostic ions [{2H1}M–PhC(O)]+ at m/z 184 (no m/z 183, P - 1) and [{2H1}M–
CH2C(O)OCH3]+ at m/z 216 (no m/z 215, P - 1), suggested that the deuterium and C3 of the productive
(2S,3R)-α-phenylalanine was retained at C3 of the β-isomer.

53

Relative Ion Abundance (%)

100

185

80

60

40

105

20
215
137 153

229

259

0
100

150

200

m/z

250

290
300

Figure A.1.8. Mass spectrometry profile of the derivatized biosynthetic β-phenylalanine product
isolated after incubating PaPAM with [ring, 2,3-2H7]-(2S,3S)-α-phenylalanine (90+% ee, 98+% 2Henriched). Fragment ion m/z 215 [{ring-2H5}M–CH2C(O)OCH3]+ (no P+1 enrichment) indicated that all
of the deuterium formerly at the C3 of the substrate was gone, while diagnostic ion m/z 185 [{ring, 2,22
H7}vM–PhC(O)]+ verified that it migrated to C2 (no P–1 observed).

54

Relative Ion Abundance (%)

100

180

80
107
60

40

20

120

122

149

132 133

162
192

0
100

150

m/z

200

223
250

Figure A.1.9. Mass spectrometry profile of the N-acetyl methyl ester of dideuterio-β-phenylalanine
isotopomer biosynthesized from [3,3-2H2]-(2S)-α-phenylalanine. A diagnostic fragment ion was found
at m/z 180 [PhCβD(NH)CαDHCO2CH3]+ (~99%) virtually no P-1 corresponding to
[PhCβD(NH)CαH2CO2CH3]+ was observed (<1%). The relative abundance of these fragment ions was
identical to the abundance of the counterpart ions derived from the unlabeled β-phenylalanine derivative.

55

Relative Ion Abundance (%)

100

180

80
107
60

40

20

120 122
132 133

149
162

192

223

0
100

120

140

160

180

m/z

200

220

240

Figure A.1.10. Authentic racemate [2,3-2H2]-N-acetyl-(2S,3R)/(2R,3S)-β-phenylalanine methyl ester
analyzed by GC/EI-MS. The mass spectrometry profile is shown.

56

A

B
192

100

80

Relative Ion Abundance (%)

Relative Ion Abundance (%)

145

115

100

60

40
161

176

20

80
176
60

118

40

145
20

135
161

205

265

0

0
100

120

140

160

100

180

150

200

250

300

m/z

m/z

Figure A.1.11. Mass spectrometry profiles of o-methylcinnamate methyl ester (A) and N(ethoxycarbonyl) o-methyl-β-phenylalanine methyl ester (B).

57

REFERENCES

58

REFERENCES
(1)

Cai, C. Z.; Han, L. Y.; Ji, Z. L.; Chen, Y. Z. Proteins 2004, 55, 66-76.

(2)

Wu, B.; Szymanski, W.; Heberling, M. M.; Feringa, B. L.; Janssen, D. B. Trends
Biotechnol. 2011, 29, 352-362.

(3)

Turner, N. J. Curr. Opin. Chem. Biol. 2011, 15, 234-240.

(4)

Behshad, E.; Ruzicka, F. J.; Mansoorabadi, S. O.; Chen, D.; Reed, G. H.; Frey, P. A.
Biochemistry 2006, 45, 12639-12646.

(5)

Thiruvengadam, T. K.; Gould, S. J.; Aberhart, D. J.; Lin, H. J. J. Am. Chem. Soc. 1983,
105, 5470-5476.

(6)

Morley, C. G.; Stadtman, T. C. Biochemistry 1970, 9, 4890-4900.

(7)

Tang, K. H.; Casarez, A. D.; Wu, W.; Frey, P. A. Arch. Biochem. Biophys. 2003, 418, 4954.

(8)

Wolthers, K. R.; Rigby, S. E.; Scrutton, N. S. J. Biol. Chem. 2008, 283, 34615-34625.

(9)

Christenson, S. D.; Liu, W.; Toney, M. D.; Shen, B. J. Am. Chem. Soc. 2003, 125, 60626063.

(10)

Krug, D.; Muller, R. Chembiochem 2009, 10, 741-750.

(11)

Freer, I.; Pedrocchifantoni, G.; Picken, D. J.; Overton, K. H. J. Chem. Soc. Chem. Comm.
1981, 80-82.

(12)

Grimm, B.; Bull, A.; Breu, V. Mol. Gen. Genet. 1991, 225, 1-10.

(13)

Murakami, K.; Korbsrisate, S.; Asahara, N.; Hashimoto, Y.; Murooka, Y. Appl.
Microbiol. Biotechnol. 1993, 38, 502-506.

(14)

Ruzicka, F. J.; Frey, P. A. Biochim. Biophys. Acta 2007, 1774, 286-296.

(15)

Walker, K. D.; Klettke, K.; Akiyama, T.; Croteau, R. J. Biol. Chem. 2004, 279, 5394753954.

(16)

Magarvey, N. A.; Fortin, P. D.; Thomas, P. M.; Kelleher, N. L.; Walsh, C. T. ACS Chem.
Biol. 2008, 3, 542-554.

(17)

Huang, S. X.; Lohman, J. R.; Huang, T.; Shen, B. Proc. Natl. Acad. Sci. U. S. A. 2013,
110, 8069-8074.
59

(18)

Chen, H. P.; Wu, S. H.; Lin, Y. L.; Chen, C. M.; Tsay, S. S. J. Biol. Chem. 2001, 276,
44744-44750.

(19)

Rétey, J.; Kunz, F.; Arigoni, A.; Stadtman, T. C. Helv. Chim. Acta 1978, 61, 2989−2998.

(20)

Mutatu, W.; Klettke, K. L.; Foster, C.; Walker, K. D. Biochemistry 2007, 46, 9785-9794.

(21)

Wang, S. C.; Frey, P. A. Trends Biochem. Sci. 2007, 32, 101-110.

(22)

DeMong, D. E.; Williams, R. M. J. Am. Chem. Soc. 2003, 125, 8561-8565.

(23)

Carter, J. H.; Dubus, R. H.; Dyer, J. R.; Floyd, J. C.; Rice, K. C.; Shaw, P. D.
Biochemistry 1974, 13, 1227-1233.

(24)

Robertson, D. E.; Noll, D.; Roberts, M. F. J. Biol. Chem. 1992, 267, 14893-14901.

(25)

Aberhart, D. J.; Gould, S. J.; Lin, H. J.; Thiruvengadam, T. K.; Weiller, B. H. J. Am.
Chem. Soc. 1983, 105, 5461-5470.

(26)

Chirpich, T. P.; Zappia, V.; Costilow, R. N.; Barker, H. A. J. Biol. Chem. 1970, 245,
1778-1789.

(27)

Walker, K. D.; Floss, H. G. J. Am. Chem. Soc. 1998, 120, 5333-5334.

(28)

Wu, P. C.; Kroening, T. A.; White, P. J.; Kendrick, K. E. Gene 1992, 115, 19-25.

(29)

Xiang, L.; Moore, B. S. J. Bacteriol. 2005, 187, 4286-4289.

(30)

Schwede, T. F.; Retey, J.; Schulz, G. E. Biochemistry 1999, 38, 5355-5361.

(31)

Rachid, S.; Krug, D.; Weissman, K. J.; Muller, R. J. Biol. Chem. 2007, 282, 2181021817.

(32)

Van Lanen, S. G.; Oh, T. J.; Liu, W.; Wendt-Pienkowski, E.; Shen, B. J. Am. Chem. Soc.
2007, 129, 13082-13094.

(33)

Christenson, S. D.; Wu, W.; Spies, M. A.; Shen, B.; Toney, M. D. Biochemistry 2003, 42,
12708-12718.

(34)

Wanninayake, U.; Walker, K. D. J. Am. Chem. Soc. 2013, 135, 11193-11204.

(35)

Wu, B.; Szymanski, W.; Wietzes, P.; de Wildeman, S.; Poelarends, G. J.; Feringa, B. L.;
Janssen, D. B. ChemBioChem 2009, 10, 338-344.

(36)

Poston, J. M. Science 1977, 195, 301-302.

60

(37)

Poston, J. M. J. Biol. Chem. 1976, 251, 1859-1863.

(38)

Poston, J. M. Biochem. Biophys. Res. Commun. 1980, 96, 838-843.

(39)

Stabler, S. P.; Lindenbaum, J.; Allen, R. H. J. Biol. Chem. 1988, 263, 5581-5588.

(40)

Parry, R. J.; Kurlyo-Borowska, Z. J. Am. Chem. Soc. 1980, 102, 836-837.

(41)

Prabhakaran, P. C.; Woo, N. T.; Yorgey, P. S.; Gould, S. J. J. Am. Chem. Soc. 1988, 110,
5785-5791.

(42)

Cone, M. C.; Yin, X. H.; Grochowski, L. L.; Parker, M. R.; Zabriskie, T. M.
Chembiochem 2003, 4, 821-828.

(43)

Hennig, M.; Grimm, B.; Contestabile, R.; John, R. A.; Jansonius, J. N. Proc. Natl. Acad.
Sci. U. S. A. 1997, 94, 4866-4871.

(44)

Moore, B. S.; Walker, K.; Tornus, I.; Handa, S.; Poralla, K.; Floss, H. G. J. Org. Chem.
1997, 62, 2173-2185.

(45)

Hanson, K. R.; Wightman, R. H.; Staunton, J.; Battersby, A. R. J. Chem. Soc., Chem.
Commun. 1971, 185-186.

(46)

Feng, L.; Wanninayake, U.; Strom, S.; Geiger, J.; Walker, K. D. Biochemistry 2011, 50,
2919-2930.

(47)

Szymanski, W.; Wu, B.; Weiner, B.; de Wildeman, S.; Feringa, B. L.; Janssen, D. B. J.
Org. Chem. 2009, 74, 9152-9157.

(48)

Wanninayake, U.; DePorre, Y.; Ondari, M.; Walker, K. D. Biochemistry 2011, 50,
10082-10090.

(49)

Pines, A.; Vega, S.; M., M. Phys. Rev. B 1978, 18, 112-125.

(50)

Wu, B.; Szymanski, W.; Wybenga, G. G.; Heberling, M. M.; Bartsch, S.; de Wildeman,
S.; Poelarends, G. J.; Feringa, B. L.; Dijkstra, B. W.; Janssen, D. B. Angewandte ChemieInternational Edition 2012, 51, 482-486.

(51)

Ratnayake, N. D.; Wanninayake, U.; Geiger, J. H.; Walker, K. D. J. Am. Chem. Soc.
2011, 133, 8531-8533.

61

CHAPTER 3: Structural and Mutational Insights into PaPAM Isomerization Reaction
The structural characterization investigations of PaPAM discussed in this chapter were
carried out in collaboration with Prof. James H. Geiger and Ms. Susan Strom from the
Department of Chemistry at Michigan State University.

3.1. Introduction
Phenylalanine aminomutase from Pantoea agglomerans (PaPAM) is a member of the
class I lyase-like family that includes phenylalanine1,2 and tyrosine3-5 aminomutases (PAM and
TAM, respectively), and phenylalanine,6,7 tyrosine8 and histidine9 ammonia lyases (PAL, TAL
and HAL, respectively). Ammonia lyases produce aryl acrylates from the corresponding α-amino
acid substrate by the elimination of ammonia, whereas aminomutases catalyze the isomerization
of α- to β-amino acids. Isomerization and elimination reactions catalyzed by this family of
enzymes depend on 4-methylidene-1H-imidazol-5(4H)-one (MIO) cofactor.9 Most of the early
structural and mechanistic discoveries on MIO-dependent enzymes are reported with histidine
and phenylalanine ammonia lyases.9-11

3.1.1. Discovery of the Electrophilic MIO Group
3.1.1.1. Dehydroalanine as the Electrophile
The presence of a catalytically essential electrophilic group in MIO-dependent enzymes
was first reported in 1967 by Abeles and coworkers.12 Histidine ammonia lyase (Pseudomonas
62

sp.), which catalyze the transformation of L-histidine to urocanic acid was inactivated by NaBH4,
KCN, NaHSO3, phenylhydrazine and NH2OH.12 However, tritium-labeled experiments
employing NaB3H4 did not successfully identify the exact functional group. Since the enzyme
was inactivated by reagents that reacted with carbonyl groups, it was believed that an enzyme
bound carbonyl group was involved in the catalysis.12 In the proposed mechanism, the amino
group of the histidine substrate was hypothesized to add to an enzyme bound carbonyl group to
form a protonated carbinolamine or a Schiff base. This amine activation step was envisioned to
make a better leaving group and facilitate the elimination step (Figure 3.1).12

Figure 3.1. First reported mechanism for MIO-dependent enzymes suggesting the involvement of
an enzyme bound carbonyl group as the electrophile.

Following their pioneering discovery, Abeles et. al. proposed a speculative structure of
the electrophilic cofactor based on the investigations performed using

14

C-nitromethane.13

Exposure of HAL to 14C-nitromethane resulted in loss of the enzyme activity, and nitromethane
was found to react at the active site. Catalytic reduction (H2/Pt) and acid hydrolysis (6 N HCl) of
the inactivated enzyme afforded 4-amino-2-hydroxybutyric acid (2) and 2,4-diaminobutyric acid
(3) (Figure 3.2).13 Based on the observation of these products, the electrophilic cofactor was

63

suggested to contain a dehydroalanine moiety, which is linked to the protein through an amide
bond at C1 (Figure 3.2). Furthermore, the amino group of the dehydroalanine was suggested to
form a Schiff base with a carbonyl compound. The nature of these linkages and the exact
structure remained unknown. Nevertheless, the proposed structure 1 could account for all the
reaction products identified from 14C-nitromethane inactivation (Figure 3.2).13

Figure 3.2. Formation of 4-amino-2-hydroxybutyric acid (2) and 2,4-diaminobutyric acid (3) from
the proposed electrophilic cofactor (1) in HAL. Dehydroalanine moiety in the proposed structure of the
cofactor is highlighted in red.
64

Surprisingly, the aforementioned investigation was not pursued in detail with HAL or
other ammonia lyases. But more studies on the inactivation of ammonia lyases by NaBH4 were
reported, yet without successful identification of the moiety reduced by NaB3H4.14,15 However, in
1969, Wickner identified the reduced tritiated moiety as [3H]alanine.16 When radioactively
labeled inhibitor NaB3H4 was used with HAL from Pseudomonas, total hydrolysis of the
inactivated protein afforded [3H]alanine. From these results, it was concluded that the prosthetic
electrophile was dehydroalanine. Furthermore, yeast phenylalanine ammonia-lyase treated with
Na14CN resulted [14C]aspartate following acid hydrolysis, presumably as a result of the
intermediate formation of β-cyanoalanine from cyanide addition to a dehydroalanyl moiety.17
Since then, a dehydroalanyl residue was generally accepted as the electrophilic cofactor of class I
lyase-like family of enzymes.
3.1.1.2. Identification of the Precursor of Active Site Dehydroalanine
The proposed dehydroalanyl moiety as the electrophilic group in HAL and PAL active
sites, prompted proposals of serine or cysteine as plausible precursors of the dehydroalanyl
group.16,18 Serine was identified as a precursor by mutating four serines of HAL (Pseudomonas
putida), which are conserved in all known HALs and PALs.19 The Ser143Ala mutation resulted a
loss of 99.9% catalytic activity whereas the other Ser-Ala mutants were equally active compared
to wt-PpHAL. Thus, Ser143 was identified as the most likely precursor of dehydroalanine.19
Similarly, site-directed mutagenesis investigations of PAL (Petroselinum crispum) suggested
Ser202 as the precursor of dehydroalanine.20 Ser143 in HAL was further defined by the
Ser143Cys mutation which retained similar activity to wild type enzyme.21

65

3.1.1.3. Discovery of the 4-Methylidene-1H-imidazol-5(4H)-one (MIO) Group
Thirty years after the initial proposals on the origins of the electrophilic group, Rétey and
coworkers, in 1999, solved the structure of Pseudomonas putida HAL (PpHAL) at 2.1 Å,
revealing a surprising structural moiety in the active site.9 Observed electron density around
Ser143, which was proposed as the precursor of electrophilic cofactor was inconsistent with a
dehydroalanine or any of the standard amino acids. The catalytic electrophile in the active site
was identified as a 4-methylidene-imidazol-5-one moiety, rather than a dehydroalanine residue
(Figure 3.3A).9 However, the MIO was deemed a modified dehydroalanine, and interestingly, the
structure was highly consistent with the speculative structure proposed by Abeles et. al., who
earlier investigated the electrophilic prosthesis with

14

C-nitromethane (cf. Figure 3.2).13 As

suggested before,19 the MIO cofactor arose from Ser143 with the involvement of the adjacent
Ala142
142

and

Gly144

residues.

An

autocatalytic

polypeptide

modification

of

the

Ala−143Ser−144Gly tandem was proposed to produce the catalytic MIO group by the

nucleophilic attack of the Gly144 nitrogen at the Ala142 carbonyl carbon followed by two
dehydration steps (Figure 3.3B).9

3.1.2. Characteristics and Formation of the Electrophilic MIO Group
The MIO moiety found in class I lyase-like enzymes is the only catalytically important
protein-derived cofactor with a heterocyclic imidazolone ring structure.9 A similar posttranslational modification that forms an imidazolone ring is found in the chromophore of green
fluorescent protein (GFP) from the Pacific Northwest jellyfish Aeqourea victoria identified

66

earlier.22,23 The p-hydroxybenzylidene-imidazolidinone fluorophore of GFP is generated by the
Ser-Tyr-Gly sequence, which is different from Ala-Ser-Gly triad in PpHAL.22 While there is no
evolutionary link or similarity in the tertiary structures of GFP and class I lyase-like family
enzymes, they nonetheless make their prosthetic groups through analogous pathways (Figure
3.3B, C).9,23

A

B

C
8

Figure 3.3. Discovery of the MIO group and autocatalytic peptide modifications involved in the
MIO formation. A) The structure of HAL (PDB 1B8F) from Pseudomonas putida revealing the catalytic
MIO group. Overall structure is shown in green color with the MIO group highlighted in spheres (atoms
are color-coded as carbon: yellow; oxygen: red; nitrogen: blue). Expanded view: enzyme active site
highlighting the MIO group. B) Proposed mechanism of autocatalytic formation of the MIO group from
Ala-Ser-Gly triad in PpHAL. C) Mechanism of chromophore formation in GFP from Ser-Tyr-Gly
tripeptide.

67

Posttranslational modifications of both the GFP fluorophore and MIO in PpHAL involve
three major biosynthetic steps (cf. Figure 3.3B, C): 1) backbone cyclization via covalent bond
formation between glycine nitrogen (Gly67 in GFP, Gly144 in PpHAL) and Ala142 (in
PpHAL)/Ser65 (in GFP) carbonyl-carbon to from a carbinol intermediate; 2) dehydration of the
same carbinol to form the imidazolin-5-one intermediate; and 3) oxidation of Tyr66 (GFP) or
dehydration of Ser143 (PpHAL) to generate an Cα-Cβ exocyclic methylidene (cf. Figure 3.3B,C).
The tripeptide sequence used to form the MIO prosthesis is variable, in part, among class
I lyase-like superfamily enzymes.24-26 While Ala-Ser-Gly is the most common precursor,7,9,24,27
Thr-Ser-Gly,25 Ser-Ser-Gly26 and Cys-Ser-Gly28 sequences are also observed (Figure 3.4). MIO
sequence variability is mostly observed in ammonia lyases, and PaPAM with its Thr-Ser-Gly
sequence is the only aminomutase that diverges from the Ala-Ser-Gly origin of MIO formation.25
However, the basis for using variable residues for MIO formation has not been evaluated so far.

PpHAL
PcPAL
AoTAL
SgTAM
CcTAM
SspTAM
TcPAM
PaPAM
FnHAL
SgHAL

RRVVEALLALLNRGITPQVPSQGSVG
FEILEAITKFLNQNITPCLPLRGTIT
PELIERLALYLNLGIVPAIPEQGSLG
PIILERLAQYLNEGITPAIPEIGSLG
VETVKLLAEFINRGIHPVIPQQGSLG
VEVLEQLATYLNRGITPAIPELGSLG
WEVMEALEKLLNSNVSPKVPLRGSVS
IVNFKKLIEIYNQGIVPCIPEKGSLG
RIVVEKLVELLNKEVTPWIPEKGSVG
PEVAQTMADVLNAGITPVVHEYGSLG
: :
* : * :
*::

ASG
ASG
ASG
ASG
ASG
ASG
ASG
TSG
SSG
CSG
**

YLTHMAHISIALLGVGNVS---YRGQVVSAQ
DLVPLSYIAGLLTGRPNSKAVGPTGVILSPE
DLAPLSHIATTVIGEGYVL---RDGGKVATG
DLAPLSHVASTLIGEGYVL---RDGRPVETA
DLSPLSHIALALIGEGTVS---FKGQVRKTG
DLAPLSHIASALIGEGYVL---RDGQPVPTG
DLIPLAYIAGLLIGKPSVIARIGDDVEVPAP
DLGPLAAIALVCTGQWKAR---YQGEQMSGA
DLSPLAHMSLVLIGLGKAY---YKGELLEAK
DLAPLSHCALTLMGEGEAEG--PDGTVRPAG
* :: :
*
.

176
235
179
184
170
181
208
197
177
176

Figure 3.4. Partial CustalW2 multiple sequence alignment of class I lyase-like family enzymes
showing the divergence in MIO forming residues (highlighted). PpHAL, FnHAL, and SgHAL:
Histidine ammonia lyases from Pseudomonas putida, Fusobacterium nucleatum, and Streptomyces
griseus, respectively. PcPAL: Phenylalanine ammonia lyase from Petroselinum Crispum. AoTAL:
Tyrosine ammonia lyase from Actinomadura oligospora. SgTAM, CcTAM, and SspTAM: Tyrosine
aminomutase from Streptomyces globisporus, Chondromyces crocatus, and Streptoalloteichus sp.,
respectively. TcPAM and PaPAM: Phenylalanine aminomutase from Taxus canadensis and Pantoea
agglomerans. * (asterisk) indicates positions which have a fully conserved residue. : (colon) indicates
conservation between groups of strongly similar properties. • (period) indicates conservation between
groups of weakly similar properties.

68

3.1.2.1. Mutants of PpHAL
Cyclization of the amino acids forming the MIO is difficult to detect, since MIO is not a
strong chromophore.9 Nevertheless, characteristics of autocatalytic MIO formation was
elucidated by structure determination of various PpHAL mutants.29,30 Mutation of the highly
conserved Asp145, which is strongly fixed in a hydrogen bonding network to an alanine
inactivated the enzyme presumably by preventing MIO formation (Figure 3.5A).29 MIO-forming
loop 142Ala-143Ser-144Gly was continuous without significant conformational changes outside the
loop, but the MIO group was not formed. The distance between Gly144 amide-nitrogen
(nucleophile) and Ala142 carbonyl-carbon (electrophile) was 3.7 Å, and thus, the cyclization
precluded. Furthermore, the hydroxyl group of Ser143 was clearly visible indicating that the
dehydration step was abolished.29
A similar mutant structure that lacked the MIO was observed for the Phe329Gly mutation
in PpHAL (Figure 3.5B).29 Phe329 is in short distance contacts from MIO and is highly
conserved in class I lyase-like superfamily. Asp145 retained its hydrogen bonding network and
tight turn at Ser143 resulted a Gly144-N and Ala142-C distance of 3.0 Å. In contrast to
Asp142Ala and Phe329Gly mutants, Phe329Ala mutant showed an intact MIO group.29
According to the mutant structures, the MIO was not formed if the strong interaction of
the MIO triad with the bordering residue Asp145 was loosened, or steric constraints from Phe
329-Cβ atom were removed.29 Except the MIO forming tripeptide, there were no changes in
overall structures compared to wt-PpHAL. Thus, Asp145 and Phe 329-Cβ function as anchor
points that assist the compression of the

142

Ala-144Gly loop.29 In addition to the conformational

pressure, hydrogen bonding interactions of the Gly144 amide-nitrogen and Ala142 carbonyl69

oxygen through an active site water molecule was also proposed to activate the late-stage MIO
formation.29 Furthermore, Ser143 dehydration was not affected by the environment of Ser143,
and thus, the cyclization of the tripeptide was suggested to occur before the dehydration of
Ser143.29 During the dehydration step, Tyr280 and Glu414 likely facilitate the elimination of
Ser143 Cα-proton and the hydroxyl group. Tyr280 can deprotonate Ser143-CαH while Glu414
protonates the hydroxyl as it eliminated.

B

A
Ser143

Gly144

Ala142

Ser143

Gly144

Ala145
Ser139

Phe329

Gly329

Asp145
Ala199
Ala142

Ser200

Figure 3.5. Comparison of wt-PpHAL and PpHAL mutants lacking an MIO. A) Overlay of wtPpHAL (PDB 1B8F; green) and Asp145Ala mutant structure (PDB 1GK3; yellow). Hydrogen bonding
network of Asp145 is also shown. B) Overlay of wt-PpHAL (green) and Phe329Gly mutant structure
(PDB 1GK2; pink). Asp145/Phe329 in wt-PpHAL and Ala145/Gly329 in mutants are shown in cyan and 10
gray, respectively. Atoms are color-coded as carbon: green/yellow/pink; oxygen: red; nitrogen: blue.

Additional site-directed mutations of PpHAL showed that modest substitutions in the
142

Ala-144Gly tripeptide are tolerated, and the MIO was still formed (Table 3.1).30 Ala142 and

Gly144 were replaced by an Ala without much loss of activity. By contrast, changing the
residues immediately adjacent to the MIO-forming loop prevented MIO formation, and no or
70

very low activity was seen.30 The Gly141Ala mutant was only 1.1% active while Asp145Ala was
totally inactive. These data indicate that the structure and conformation of the immediate
environment of the MIO-forming loop affect the ability of PpHAL to form the MIO group.30

Table 3.1. Mutations in 141Gly-142Ala-143Ser -144Gly-145Asp sequence in PpHAL and relative catalytic
activity of corresponding mutants.

Mutation

Relative Activity (%)

MIO Formation

Reference

None (wt-PpHAL)
Gly141Ala
Ala142Gly
Ser143Cys
Gly144Ala
Asp145Ala

100
1.1
93
100
37
0

Yes
ND*
Yes
Yes
Yes
No

31
31
31
22
31
30

* Crystal structure was not obtained. ND: Not Determined

3.1.2.2. Mutants of TcPAM
Recently, the biosynthesis of the MIO in Taxus chinensis PAM (TchPAM) was
investigated using site-directed mutagenesis.31 In a previous study, highly conserved, Tyr322 and
Asn231 were identified as essential residues for TchPAM activity; both Tyr322Ala and
Asn231Ala mutants were inactive.32 Recently, Janssen and coworkers determined Tyr322Ala
and Asn231Ala mutant crystal structures revealing the intact

175

Ala-176Ser-177Gly tripeptide.31

Enzyme containing Tyr322Ala and Asn231Ala mutants lacked a mature MIO and Ser176 was
not dehydrated (Figure 3.6). In the structure of Asn231Ala mutant, the Gly177 amide-nitrogen
and Ala175 carbonyl-carbon are in a proper position (2.8 Å) for the cyclization reaction, but
seemingly the amide-nitrogen was not activated to initiate the nucleophilic attack when the
Asn231 is removed (Figure 3.6A).31 Thus, a hydrogen bonding interaction between Ser176
71

carbonyl-oxygen and Asn231 Nδ, likely increases the nucleophilicity of the Gly177 amidenitrogen, thus MIO formation is prompted. Similarly, Arg96 in GFP was suggested to activate
the amide-nitrogen of Gly67 to increase its nucleophilicity and facilitate the cyclization of 65Ser66

Tyr-67Gly tripeptide.33 Hydrogen bonding of the carbonyl-oxygens of Arg96 and Tyr66

increases the acidity of the Gly67 amide-nitrogen, and Arg96 stabilizes the peptide bond
resonance form.33,34
In contrast to Asn231Ala mutant, the MIO-forming loop of the Tyr322Ala mutant was
deviated from the normal spatial orientation of MIO group atoms in wt-TchPAM (Figure 3.6B).31
The Gly177 amide-nitrogen atom and the Ala175 carbonyl-carbon atom were not in a favorable
orientation for MIO-formation. Tyr322 presumably positions Ala175 in a proper location for the
nucleophilic attack by Gly177.31 In addition, the carbonyl-oxygen atom of Ser176 was too far
from the Nδ atom of the Asn231 side chain for a hydrogen bonding interaction. Thus, the Gly177
nitrogen atom is likely not activated to initiate the cyclization.31
B

A

Asn231
Ser176

Ser176
5.3 Å
Tyr322

Tyr322

Asn231
Gly177

Ala175

3.3 Å
Ala175
Gly177

Figure 3.6. Comparison of wt-TchPAM and TchPAM mutants lacking an MIO. A) Overlay of wtTchPAM (PDB 2YII; green) and Asn231Ala mutant structure (PDB 4C6G; yellow). Hydrogen bonding
interaction between Asn231 in wt-TchPAM and Ser176 carbonyl-oxygen is shown. B) Overlay of wtTchPAM (green) and Tyr322Ala mutant structure (PDB 4C5U; pink). Shown are distance between
Ala175 carbonyl-carbon and the Gly177 amide-nitrogen, and hydrogen bonding interaction between
Asn231 and Ser176 carbonyl oxygen. Atoms are color-coded as carbon: green/yellow/pink; oxygen: red;
12
nitrogen: blue.
72

3.1.3. Mechanism of MIO-dependent Enzymes
3.1.3.1. Mechanistic Proposals on Ammonia Lyase Elimination Reaction
After clarifying that the electrophilic cofactor as dehydroalanine, various mechanistic
investigations were carried out on PALs and HALs.11,35 First mechanistic proposal from Hanson
and Havir (1970) described a mechanism, where the amino group of phenylalanine is added to
the β-position of dehydroalanine.18 Subsequent investigations by Hermes et. al. carried out on
PAL also proposed a Michael addition of α-amino group of the substrate (Figure 3.7, path a).35
According to this mechanism, nucleophilic addition of the α-amino group on the prosthetic
electrophile produce a covalently bound enzyme intermediate. The addition of the amino group
to the electrophile was suggested to enhance its leaving group ability.35 However, this
mechanism did not explain how the non-acidic β-proton (pKa ~ 40)36 of the substrate is
abstracted by an enzymatic base.
In 1995, Rétey and coworkers found contradictory results to this mechanism that has
been accepted widely for about 30 years.10,11 Both a PAL Ser202Ala mutant and wt-PAL
inactivated with NaBH4 reacted much faster with p-nirophenylalanine than with the natural
substrate.11 These observations were thought to disagree with a mechanism involving
nucleophilic attack of α-amino group at dehydroalanine cofactor. Thus, the function of
dehydroalanine was proposed to be similar to that of the nitro substituent on a phenyl ring to
increase the acidity of β-hydrogens of the substrate and promote the elimination reaction.11
Analogous results were obtained for HAL when 5-nitrohistidine was reacted with the mutants
lacking the dehydroalanine electrophile.10 These observations lead to the proposal of an
alternative mechanism, where the aromatic ring of the substrate reacted with the dehydroalanine
73

via Friedel-Crafts-like addition (Figure 3.7, path b).10,11 The positive charge created at Cipso
enhances the acidity of β-hydrogens and thus facilitated the abstraction by an enzymatic base.11
Further support for this mechanism was provided by the finding that m-hydroxyphenylalanine
was a slightly better substrate for PAL than phenylalanine.11 In contrast, p-hydroxyphenylalanine
was a poor substrate, because unlike m-hydroxy, p-hydroxy group does not facilitate an
electrophilic attack at ortho-position.

Figure 3.7. Proposed mechanisms of PAL reaction. Path a) Amino alkylation-mechanism proposed by
Hanson and Havir18 and modified by Hermes et.al.35 Path b) Alternative Friedel-Crafts-like addition
mechanism proposed by Rétey and coworkers.11

The Friedel-Crafts-like addition mechanism however did not explain the energy demands
for losing aromaticity of the substrate phenyl ring, and why an exocyclic double bond was
created when the ring could have been rearomatized by deprotonation of the σ-complex.37
Described earlier, as the field advanced, crystallographic analysis revealed the dehydroalanine

74

was an MIO group and this electrophile is more electrophilic than the dehydroalanyl moiety and,
thus, more opt to react with the aromatic ring.9 The planar geometry of the MIO heterocyclic
ring, which is maintained by the fold of the polypeptide chain, prevents delocalization of Gly144
nitrogen lone pairs into the α,β-unsaturated carbonyl system (cf. Figure 3.3A, B). Furthermore,
addition of a nucleophile at the methylidene carbon renders the MIO imidazole ring aromatic,
thus the transient loss of aromaticity in the phenyl ring of the substrate is compensated.9 In
addition, it was proposed that PAL and HAL prevent rearomatization of the phenyl ring by
eliminating basic side chains in the aromatic ring binding pocket, while a basic group is properly
positioned to remove a benzylic proton of the substrate.38
3.1.3.2. Mechanism of MIO-dependent Aminomutases
The dependency of aminomutases on an MIO-cofactor was identified when Streptomyces
globisporus tyrosine aminomutase (SgTAM), homologous to ammonia lyases, was
characterized.3 As with HALs14,16 and PALs15,17, SgTAM was inactivated upon NaBH4 and KCN
treatment.3 Furthermore, analogous Ser153Ala mutation on the precursor of exocyclic double
bond of the MIO group resulted a 340-fold decrease in the catalytic efficiency compared to wtSgTAM. The mutase activity of a phenylalanine aminomutase from Taxus cuspidata (TcPAM),
which is homologous to SgTAM (56% similarity) and ammonia lyases (50-70% similarity) was
also abolished upon the pretreatment with KCN or NaBH4.1 Taken together with the conserved
Ala-Ser-Gly motif, these results suggested the existence of the MIO cofactor in TAMs and
PAMs.1,3
Analysis of the product pool catalyzed by SgTAM and TcPAM indicated that the
corresponding trans-acrylic acid is produced as a minor product during catalysis, providing the
75

evidence for an initial elimination reaction during α- to β-amino acid conversion.1,3 Based on the
homology to ammonia lyases, it was envisioned that either an amino-MIO adduct (Figure 3.8A)
or a Friedel-Crafts-like (Figure 3.8B) mechanism leads to an α,β-elimination reaction resulting
trans-acrylate intermediate. Unlike ammonia lyases, which release ammonia and an α,βunsaturated propenoate, aminomutases catalyze a subsequent conjugate addition reaction of
ammonia to the β-position of trans-acrylate (Figure 3.8).1,3

A

B

Figure 3.8. Amino-MIO adduct (A) and Friedel-Crafts-like (B) isomerization mechanisms of MIOdependent aminomutases based on the proposed mechanistic proposals for MIO-based ammonia
lyases.

76

Subsequent

structure

determination

of

SgTAM27

and

Taxus

phenylalanine

aminomutases,24,32 confirmed the presence of the MIO group in their active sites. The structures
revealed that the overall architecture of aminomutases and that of their active sites are very
similar to MIO-dependent ammonia lyases.9,24,27 However, based on the mechanism,
aminomutases are expected to retain ammonia in the active site to achieve the isomerization
reaction. Bruner and co-workers compared the structures of RsTAL39 and SgTAM, to investigate
the structural basis for aminomutase activity of the latter.27,40 It was proposed that the more
solvent accessible RsTAL active site lacks the ability to maintain ammonia and p-hydroxy
cinnamate.39 In contrast, the hydrogen bonding interaction of Tyr303 and Glu71 of SgTAM
(Figure 3.9), an interaction not present in RsTAL or other ammonia lyases, blocks the solvent
accessibility of SgTAM and prevents ammonia from leaving the active site.27 Tyr303Ala and
Glu71Ala single mutants, and the double mutant probed the hypothesis of the varied solvent
accessibility between the aminomutase and the lyase activities.40 Although the mutations resulted
in open tunnel, the activity of SgTAM was not altered to a lyase. The mutants still catalyzed the
ammonia rebound albeit with lower activity compared to wt-SgTAM.40

MIO

Glu71

Tyr303

Figure 3.9. wt-SgTAM active site (PDB 2OHY) showing the closed active site due to the hydrogen
bonding between Glu71 and Tyr303. MIO is also shown. Atoms are color-coded as carbon: green;
oxygen: red; nitrogen: blue
77
26

Structural characterization of SgTAM bound to mechanism-based inhibitors provided
significant insights into the mechanism of MIO-dependent aminomutases.41,42 In their cocrystal
complexes,

product-like

inhibitors,

α,α-difluoro-β-tyrosine

(Figure

3.10A)

and

-β-p-

methoxyphenylalanine (Figure 3.10B) were covalently-bound to the MIO methylidene via the αamino group.41 The observed electron density of the active sites did not match an MIO-phenyl
ring adduct consistent with the Friedel-Crafts-like mechanism suggesting the occurrence of an
NH2-MIO adduct mechanism. Additional cinnamate epoxide inhibitors were also designed to
mimic the p-hydroxycinnamate intermediate in SgTAM reaction.42 The structure of the pfluorocinnamate epoxide bound to SgTAM confirmed the mechanism of catalysis via an NH2MIO adduct.
A

B

C

Figure 3.10. Structure of SgTAM solved with various mechanism-based inhibitors and ligands. wtSgTAM bound to α,α-difluoro-β-tyrosine (A; PDB 2QVE) and α,α-difluoro-β-p-methoxyphenylalanine
3
(B; PDB 2RJS) showing the NH2-MIO adducts. Inhibitors are shown in cyan. C) Structure of Tyr63Phe
mutant SgTAM with L-tyrosine modeled into partially occupied electron density (PDB 3KDZ). LTyrosine and MIO are shown in cyan and red respectively. Atoms are color-coded as carbon:
green/cyan/gray; oxygen: red; nitrogen: blue; fluorine: light blue.

78

Although the crystallization of mechanism-based inhibitors were successful, the attempts
to obtain an enzyme-bound L-tyrosine (substrate) were not successful with SgTAM.40 The X-ray
cocrystal structure of the SgTAM mutant of the catalytic base (Tyr63Phe) with

L-tyrosine

revealed only a partial occupancy of the substrate. The electron density of the α- and β-carbons
of L-tyrosine and their covalent bonding to the MIO via the amino group were very weak (cf.
Figure 3.10C).40 Currently, there are no reports in literature on the identification of naturally
occurring mechanism-based intermediates. The work described herein involved the identification
of naturally occurring reaction pathway intermediates using structural determination of PaPAM.
In addition, as described before (see section 3.1.2), Thr167 in the MIO forming motif
(167Thr-168Ser-169Gly) of PaPAM is different from other aminomutases,24,27 and the basis for
using variable residues for MIO formation is yet to be determined. The investigations on
determining the catalytic and structural role of Thr167 during PaPAM catalysis is described
herein.

79

3.2. Experimental

3.2.1. Mutagenesis of PaPAM cDNA
Point mutations of PaPAM were performed using PfuUltra HF DNA polymerase
provided with the QuikChange II XL site-directed mutagenesis kit or PfuTurbo polymerase
(Stratagene, La Jolla, CA). The forward oligonucleotide primers used to exchange residue F455
to alanine (A) or asparagine (N); T167 to alanine (A), serine (S), glycine (G), valine (V) or
cystine (C); and S168 to alanine (A) in the mutagenesis reactions were as follows (mutations are
bold and underlined). The corresponding reverse-complement primer was paired with each
forward primer.
F455A: 5′–GC ACC ACC GGA GAC GCT CAG GAT ATC GTC TC–3′
F455N: 5′–GC ACC ACC GGA GAC AAT CAG GAT ATC GTC TC–3′
T167A: 5′–GAA AAA GGT TCG CTG GGA GCT AGT GGC GAT–3′
T167S: 5′–GAA AAA GGT TCG CTG GGA AGC AGT GGC GAT–3′
T167G: 5′–GAA AAA GGT TCG CTG GGA GGC AGT GGC GAT–3′
T167V: 5′–GAA AAA GGT TCG CTG GGA GTG AGT GGC GAT–3′
T167C: 5′–GAA AAA GGT TCG CTG GGA TGC AGT GGC GAT–3′
S168A: 5′–GC CTG GGA ACC GCT GGC GAT CTG GGG C–3′

Expression vector pET24b(+) containing the wild-type PaPAM cDNA was used as the
DNA template in the PCR reactions. A typical PCR reaction mixture was consisited of wild-type

80

PaPAM DNA template (10 ng), corresponding forward and reverse primers (150 ng each), dNTP
mix (1 μL), 10X reaction buffer (5 μL), PfuUltra HF or PfuTurbo DNA polymerase (1 μL at 2.5
U/μL), and ddH2O to bring the volume to 50 μL. QuickSolution reagent (3 μL) was added for
PCR reactions performed with QuikChange II XL site-directed mutagenesis kit.
The PCR program conditions for PfuUltra HF DNA polymerase were as follows: initial
denaturing at 95 °C for 2 min followed by 30 cycles at 95 °C for 50 s, 55 °C for 50 s, and 68 °C
for 7 min, and finally, the reactions were held at 68 °C for 7 min. For PfuTurbo DNA
polymerase, initial denaturation with similar conditions was followed by 30 cycles at 95 °C for
30 s, 60 °C for 30 s, and 72 °C for 7 min, and the temprature was held at 72 °C for 10 min. After
temperature cycling, the reactions were place on ice, to each was added restriction enzyme DpnI
(1 μL at 10 U/μL), and the reactions were incubated at 37 °C for 1 h to digest the template DNA.
An aliquot (2 μL) of the plasmid solution from each PCR reaction was used to transform XL10Gold ultracompetent cells (provided in the QuikChange II XL site-directed mutagenesis kit) or
DH5α competent E. coli cells. The resultant plasmids encoding mutations in the papam cDNA
were confirmed by sequencing the each corresponding mutant cDNA.

3.2.2. Expression and Purification of PaPAM Mutants
Plasmids of each mutant were used separately to transform BL21(DE3) E. coli cells. The
transformants were grown on Luria-Bertani agarose medium supplemented with kanamycin (50
μg/mL) and grown overnight at 37 °C. Resistant cells harboring PaPAM mutants were grown in
liquid Luria-Bertani medium (5 mL) supplemented with kanamycin (50 μg/mL) at 37 °C for 16

81

h. The entire 5-mL inoculum of E. coli BL21(DE3) was added to Luria-Bertani medium (1 L)
and incubated at 37 °C until the optical density (A600) was ~0.6. The culture was cooled to 16 °C,
and to the cell suspension was added isopropyl-β-D-thiogalactopyranoside (100 μM) to induce
protein expression. The cells were incubated at 16 °C for 16 h. The subsequent steps were
performed at 4 °C, unless otherwise indicated. The cells were harvested by centrifugation at
6,000g (15 min) and the cell pellet was resuspended in lysis buffer (50 mM sodium phosphate
buffer containing 5% (v/v) glycerol, 300 mM NaCl, 10 mM imidazole, pH 8.0). The cells were
lysed by sonication (Misonix sonicator, Farmingdale, NY), the lysate was centrifuged at 9,700g
(45 min) to remove cell debris, and the supernatant was decanted and centrifuged at 102,000g (1
h) to remove light membranes. The resultant crude mutant proteins in the soluble fraction was
separately purified by Nickel-nitrilotriacetic acid (Ni-NTA) affinity chromatography according
to the protocol described by the manufacturer (Qiagen, Valencia, CA). Fractions containing the
mutant enzyme that eluted in 250 mM imidazole were combined and concentrated. The buffer
was exchanged with 50 mM sodium phosphate buffer containing 5% (v/v) glycerol (pH 8.0)
using a size selective centrifugal filtration unit (Centriprep centrifugal filter units, 30,000
MWCO, Millipore). The purity of the concentrated enzyme was assessed by SDS−PAGE with
Coomassie Blue staining, and the quantity was determined by the Bradford protein assay. The
overexpressed PaPAM mutants (~59 kDa) were at ~90% purity (~50 mg/L).

82

3.2.3. Assessing the Absolute Stereochemistry of the β-Phenylalanine Product Catalyzed by
PaPAM Mutants
(2S)-α-Phenylalanine (1 mM) was incubated separately with F455A and F455N (4 mg) in
phosphate buffer (10 mL, pH 8.0) at 31 °C. Each reaction was quenched after 24 h by basifying
to pH 10 (6 M NaOH). (1S)-(–)-Camphanic chloride (~2 mg) was added to each reaction
mixture. The solution was stirred (45 min), acidified to pH 2-3 (6 M HCl), and extracted with
diethyl ether (2 × 2 mL). The organic solvent was evaporated under vacuum, and the resulting
residue was dissolved in ethyl acetate/methanol (3:1, v/v) (200 μL). The solution was treated
with excess (trimethylsilyl)diazomethane until a faint yellow color persisted. The derivatized βamino acid was identified by GC/EI-MS analysis and compared against the retention time and
mass spectrometry fragmentation of authentic N-[(1'S)-camphanoyl]-(3S)-β-phenylalanine
methyl ester

3.2.4. Total Cell Protein (TCP) Fraction Analysis
BL21(DE3) E. coli cells transformed to express wt- and mutant-PaPAM were grown in
liquid Luria-Bertani medium (10 mL) supplemented with kanamycin (50 μg/mL) at 37 °C. When
the optical density (A600) reached ~0.6, a 1-mL well-mixed culture sample was obtained from
each transformant. The cultures were cooled to 16 °C, and isopropyl-β-D-thiogalactopyranoside
(100 μM) was added to induce the protein expression. The cells were incubated at 16 °C for 16 h
and a 1-mL induced cell sample was obtained from each culture. Uninduced and induced cell

83

samples were harvested by centrifugation (13,000 g for 5 min) and the TCP fractions were
analyzed by SDS-polyacrylamide gel followed by Coomassie blue staining.

3.2.5. Circular Dichroism (CD) Spectroscopic Analysis
CD spectra of each protein were recorded in a CD spectrometer (Chirascan, Applied
Photophysics, Surrey,United Kingdom) using a quartz cell of 1 mm path length. Spectra in far(180-250 nm) and near-UV (260-310 nm) regions were obtained with 0.25 mg/mL and 8 mg/mL
protein concentrations, respectively. The wavelength interval was set to 0.5 nm with 2.5 s signal
average per point and each sample was scanned for five times. All CD spectra were obtained in
phosphate buffer at pH 8. Reference spectra were obtained for all samples containing only the
buffer components, and the final spectra are the difference of the protein sample spectra after
subtraction of the reference spectra. For each sample, three separate samples were analyzed at
each wave-length region. The fractional helicity (ƒH) of wt- and mutant-PaPAM was calculated
using the following equation (θ: negative molar ellipticity).43

fH 

([ ]208  4000)
(33000  4000)

84

3.3. Results and Discussion

3.3.1. Structural Insights into Mechanism of PaPAM Isomerization
3.3.1.1. Characteristics of the Overall Structure and the Active Site
The crystal structure of PaPAM determined at 1.7 Å resolution displays structural
homology to other MIO-dependent aminomutases (TcPAM24,31 and SgTAM27) and homologous
ammonia lyases.6-9,44 The overall structure of the PaPAM monomer (541 amino acids) was
predominantly α-helical with most helices running along the long axis of the monomer structure
(Figure 3.11A). There are nine short β-strands (the longest strand consists of four amino acids)
that form three antiparallel and one parallel β-sheet. Also notable are two monomers per
asymmetric unit, and the active unit exists as a tetramer (a head-to-tail dimer of dimers related by
222 point symmetry) (Figure 3.11B, C).
Each subunit of the catalytically functional tetramer consists of an active site located
close to one end of the monomer at the termini of several α-helices (Figure 3.11A, C). Amino
acid side chains from three monomers contribute to generate a composite active site (Figure
3.11D). The carboxylate group of the bound propanoate moiety forms a mono dentate salt-bridge
interaction with the δ-guanido group of the active site Arg323. The non-polar side chains of
Leu104, Val108, Leu171, Leu216, Leu421, Met424 and Phe455, and the hydrophilic side chain
of Gln456 surround the aromatic ring of the bound phenyl propenoate moiety (Figure 3.11D).
The Arg323−carboxylate salt-bridge interaction and the hydrophobic contacts between the
aromatic ring likely anchor the aromatic amino acids in the PaPAM active site.
85

A

B

C

D
Tyr78

A

Phe455

B

Leu421

C
Arg323

Met424
Leu104
Gln456

Leu216

Tyr320

Val108

Phe428

Leu171

MIO

Figure 3.11. Overall structure and the active site characteristics of PaPAM. A) Structure of PaPAM
monomer showing the active site positioned at the end of the four helix bundle. The MIO and the phenyl
propenoate moiety are shown in spheres. The side- (B) and top- (C) view of the PaPAM tetramer. Each
monomer is distinctly colored. D) Active site of PaPAM showing residues which are in 4 Å distance from
the MIO-bound phenyl propenoate moiety (atoms are color-coded as carbon: teal; oxygen: red; nitrogen:
blue). The active site residues are color-coded (carbon atoms in green, yellow, and wheat) to differentiate
the involvement from each sub unit.

The MIO cofactor of PaPAM, formed by the self-condensation of a Thr167-Ser168-Gly169
triad, was covalently bound to a phenyl propanoate substrate in each active site. In addition to the
MIO cofactor, the PaPAM active site contains key catalytic residues that are found in other
structurally characterized enzymes6,9,24,27 of class I lyase-like family (cf. Figure 3.11D). The
enzymatic base Tyr78, required for the de- and re-protonation, is positioned ~3.5 Å above both
86

the α- and β-carbon atoms of the phenyl propanoate adducts. Further, Tyr320 is 2.6 Å from the
amino group of the α-phenylalanine intermediate, and thus believed to facilitate Cβ to Cα proton
transfer when the amino group migrates reciprocally.25
3.3.1.2. Identification of Mechanism-based Intermediates
Active sites of the PaPAM dimer revealed two different types of ligands bound to the
methylidene of the MIO cofactor. The electron density observed in monomer "A" revealed a βphenylalanine molecule covalently attached by its amino group to the methylidene of the MIO
(Figure 3.12A). The electron density of monomer "B" (Figure 3.12B, C), however, suggested
partial occupancy of two ligand types. The electron density for a β-phenylalanine-type adduct
was evident, as seen in monomer "A", but there was also electron density that is consistent with a
molecule indistinguishable from α-phenylalanine that was covalently attached by its amino group
to the methylidene carbon atom of MIO (Figure 3.12B, C).
Observation of α/β-phenylalanine-type MIO-bound ligands in the PaPAM active site was
surprising, particularly, since the crystallization buffer for PaPAM included cinnamate as a
ligand. It should be mentioned that PaPAM and homologous TcPAM were shown to function as
aminotranferases capable of transferring the amino group from an external amino source to an
arylacrylate intermediate.45-47 Both α- and β-phenylalanine (1:1) were produced by the addition
of ammonia across the double bond of cinnamic acid.46,47 With this data, the source of the α- and
β-amino acids evidenced in this crystallographic study, described herein, could be reconciled.
The LB media (pH 7.3) used to grow the bacteria, in which protein was overexpressed and used
in the crystallographic study was estimated to contain ammonia at a concentration of 2.4 mM, as
assessed in an earlier study.48 It was imagined that this source of ammonia could bind to the MIO
87

(NH2-MIO) of PaPAM overexpressed in E. coli. Therefore, the α- and β-phenylalanines were
likely formed by the reaction between cinnamate and NH2-MIO adduct remained after protein
purification.
The enzyme-bound ligands in PaPAM active sites are consistent with a mechanism that
proceeds via amino-alkylation pathway (Figure 3.12D) and not through a Friedel–Crafts-like
pathway, as recently proposed.49 This suggests that the nucleophilic amino group of the substrate
and not the π-electrons of the aryl ring attacks the MIO of PaPAM during the isomerization
reaction. Further evaluation of the α- and β-phenylalanine complexes of monomer "B" (Figure
3.12B, C) suggested that during the amino group isomerization, the phenylpropanoid carbon
backbone remains mostly stationary above NH2-MIO adduct. Previous mechanistic studies on
PaPAM supported that the amino group of the substrate is removed from Cα and reattached at Cβ
of the cinnamate intermediate on the same stereoface.25 These configurations are consistent with
the mechanism of stereoselectivity for this enzyme that proceeds with inversion-of-configuration
at each migration terminus.25
The MIO-adducts observed with PaPAM are, in part, similar to the complexes of
mechanism-based inhibitors in the SgTAM structures.41,42 The SgTAM structural data confirmed
that a covalent intermediate forms between the substrate or product and the MIO. The Cα of pfluorocinnamate epoxide and Cβ of α,α-difluoro-p-methoxy-β-phenylalanine (product mimic)
were covalently bound to the MIO methylidene carbon by the epoxide-oxygen and the β-amino
group, respectively (cf. Figure 3.10).42 Nonetheless, the structure of PaPAM was the first of the
MIO-dependent enzyme to be crystallized with naturally occurring pathway intermediates.

88

A

B

C

Figure 3.12. MIO-bound intermediates identified in PaPAM active sites. β-Phenylalanine-type (A)
and α/β-phenylalanine-type (B) molecules bound to the MIO cofactor. C) Electron density (2Fo-Fc map,
blue mesh) calculated at 1.0 σ around the α- and β-phenylpropanoid that is covalently bound to the MIO
found in monomer "B". Atoms are color-coded as carbon: green/teal; oxygen: red; nitrogen: blue. D)
Amino-alkylation mechanism of PaPAM isomerization. MIO-bond intermediates highlighted in dashedsquares resemble the enzyme-bound ligands seen in PaPAM active site.

3.3.2. Mutational Insights into Mechanism and Stereochemistry of PaPAM Isomerization
3.3.2.1 Comparison of the PaPAM Active Site with other MIO-dependent Aminomutases
PaPAM is homologous to Taxus phenylalanine aminomutase TcPAM1 (31% identity and
47% similarity),

L-tyrosine

2,3-aminomutase SgTAM from Streptomyces globisporus3 (37%

identity and 56% similarity) and CcTAM from Chondromyces crocatus4 (38% identity and 55%
similarity). Structures of SgTAM bound to different inhibitors (PDB 2QVE, 2RJS and 2RJR)42

89

and TcPAM bound to cinnamate intermediate (PDB 3KDZ and 2YII)24,31 have been solved
previously. Overall, the active sites of the MIO aminomutases exhibit very similar architecture,
and the amino acid side chains lining the active sites are largely conserved. SgTAM, however,
contains a non-conserved His93 (Val108 in PaPAM and Leu108 in TcPAM) (Figure 3.13B) and
Tyr415 (Phe428 in PaPAM and Ile431 in TcPAM) to orient to the hydroxy group of the tyrosine
substrate.27,41 In addition to Phe428 in PaPAM, two other residues of chain "B" contribute to
PaPAM active site are also not conserved among the aminomutases (cf. Figure 3.11D). Met424
in PaPAM is a charged Lys427 in TcPAM and a smaller, hydrophobic Ala411 in SgTAM.
Further, Leu421 in PaPAM is a more polar Tyr424 in TcPAM and Ser408 in SgTAM. Although
the side chain chemistry of these residues are variable among the members of aminomutase
family, given their >5 Å distance from the bound ligand, they are less likely function
catalytically in the isomerization reaction.
Two prominent differences were seen in the active sites of MIO-dependent enzymes.
First, the trajectory of ligands in PaPAM, SgTAM and TcPAM active sites was significantly
different. Compared to the bi-dentate salt bridge in TcPAM between the carboxylate of the
cinnamate intermediate and the conserved Arg325 side chain,24 ligands in PaPAM form a
presumably weaker mono-dentate salt bridge with Arg323 (Figure 3.13A). A similar alignment
with a mono-dentate salt bridge was seen in SgTAM structures containing α,α-difluoro-βtyrosine (Figure 3.13B), α,α-difluoro-p-methoxy-β-phenylalanine and p-fluoro cinnamate
epoxide intermediate.41,42 The trajectory of MIO-bound ligands was almost identical in the
PaPAM and SgTAM active sites, which catalyze an equivalent stereochemistry in the
isomerization reaction.3,25 A difference in residues near the carboxylate binding site of PaPAM
was notable when compared with those of other MIO-dependent enzymes. Phe455 in PaPAM is
90

Asn458 in TcPAM (Figure 3.13A), Asn442 in SgTAM (Figure 3.13B)] or glutamine6,9 in almost
all other enzymes in the lyase family. Presumably, steric clash from Phe455 alters the trajectory
of the covalently bound α- and β-phenylalanine ligands and forces a weaker mono-dentate salt
bridge in PaPAM. A ligand trajectory consisted with a bi-dentate salt bridge, as seen in
TcPAM,24 causes steric overlap between the carboxylate group and Phe455. Interestingly,
SgTAM which has Asn442 (similar to Asn458 in TcPAM) at a position equivalent to Phe455
also engaged in a mono-dentate salt bridge between the ligands and conserved Arg311 (Figure
3.13B). The p-hydroxy group of L-tyrosine in SgTAM, however, makes a hydrogen bond with
the non-conserved His93 (Figure 3.13B).27 The hydrogen bonding interaction, absent in PaPAM,
can therefore enforce the altered trajectory of ligands in the SgTAM active site. Thus, SgTAM
and PaPAM catalyze equivalent stereochemistries, presumably by orienting their substrates
identically in their active sites using a distinct set of enzyme–substrate interactions.

A

B
Asn442
Asn458

Phe455

Phe455
His93

Arg

Val108

Arg

Figure 3.13. Comparison of PaPAM (green) with TcPAM (yellow) and SgTAM (pink) active sites.
Atoms are color-coded as carbon: green/yellow/pink; oxygen: red; nitrogen: blue A) Overlay of the
TcPAM active site (PDB 3NZ4; showing the cinnamate, Arg 325, and Asn458) and the PaPAM active
site (showing the (S)-β-phenylalanine/MIO adduct, Arg 323, and Phe455). B) Overlay of the SgTAM
active site (PDB 2QVE; showing the α-difluoro-β-tyrosine, Arg 311, Asn442, and His93) and PaPAM
active site.
91

11

A second prominent difference in the PaPAM active site was the residues used to form
the MIO. Generally, the catalytic MIO cofactor in aminomutases forms autocatalytically by
cyclization and dehydration of the Ala-Ser-Gly triad (in TcPAM, SgTAM, and CcTAM) (cf.
Figure 3.4).9,24,27 However, the PaPAM MIO is made from a Thr-Ser-Gly tandem instead of the
common Ala-Ser-Gly sequence (Figure 3.14).

A

MIO

Figure 3.14. Overlay of the TcPAM (yellow; showing the cinnamate and MIO) and the PaPAM
B
(green;
showing the (S)-β-phenylalanine/MIO adduct) MIO groups. Atoms are color-coded as carbon:
PaPAM IVNFKKLIEIYNQGIVPCIPEKGSLGTSGDLGPLAAIALVCTGQ--WKARY-QGEQMSGA
197
green/yellow;
oxygen: red; nitrogen: blue.
TcPAM WEVMEALEKLLNSNVSPKVPLRGSVSASGDLIPLAYIAGLLIGKPSVIARIGDDVEVPAP 208
SgTAM PIILERLAQYLNEGITPAIPEIGSLGASGDLAPLSHVASTLIGE--GYVLR-DGRPVETA 184
CcTAM VETVKLLAEFINRGIHPVIPQQGSLGASGDLSPLSHIALALIGE--GTVSF-KGQVRKTG 170
AoTAL Mutational
PELIERLALYLNLGIVPAIPEQGSLGASGDLAPLSHIATTVIGE--GYVLR-DGGKVATG
179
3.3.1.2.
Analysis to Assess the Role of Phe455 in the Isomerization Reaction
SgHAL PEVAQTMADVLNAGITPVVHEYGSLGCSGDLAPLSHCALTLMGE--GEAEGPDGTVRPAG 176
PcPAL FEILEAITKFLNQNITPCLPLRGTITASGDLVPLSYIAGLLTGRPNSKAVGPTGVILSPE 235
: :
* .: * :
*:: **** **: *
*.
.
.

TcPAM is a homologue of PaPAM and makes the enantiomer of the (3S)-product made

1325
by PaPAM.50 Recent studies on these enzymes suggest that they follow a similar mechanism.

In both enzymes, the pro-(3S) proton and α-amino group exchange their positions
intramolecularly.25,50 TcPAM, however, is proposed to rotate the cinnamate intermediate on the
reaction pathway to obtain the observed product stereochemistry, while this intermediate remains
stationary in the PaPAM reaction.24,25 Therefore, point mutations within the PaPAM active site

92

were used to investigate the structural differences between these isozymes that may be
responsible for the origin of their enantioselectivities.
As described before, the sterically bulkier Phe455 side chain likely orients the
substrate/intermediates in PaPAM active site towards a mono-dentate salt bridge interaction with
active site Arg323. In contrast, Asn458 in TcPAM is void of unfavorable steric interactions with
the substrates/intermediates. Consequently, TcPAM has more flexibility that enables the
proposed bond rotation of the cinnamate intermediate.24,25 It was hypothesized that mutating
Phe455 in PaPAM to a smaller side chain would allow for the rotation of the cinnamate
intermediate in the active site, and thus change its enantioselectivity. Phe455 was mutated to
Asn455, which is analogous to Asn458 in TcPAM24 and Ala455 with a smaller side chain. The
activity and the enantioselectivity of each mutant were evaluated. At steady state, wt-PaPAM
converts (S)-α-phenylalanine to (S)-β-phenylalanine and trans-cinnamate in a 90:10 ratio.25 By
contrast, PaPAM mutants Phe455Ala and Phe455Asn form (S)-β-phenylalanine and transcinnamate in a 40:60 ratio at approximately 2% of the rate of wt-PaPAM. Therefore, in Phe455
mutants, the catalytic efficiency of the enzyme was dramatically eroded. Likely, the cinnamate
intermediate aligns in an unfavorable orientation that impairs the transfer of the amino group
from the MIO to the β-carbon. Consequently, compared to wt-PaPAM, a higher ratio of
cinnamate is produced from Phe455 mutants. These data are consistent with the hypothesis that
Phe455 is important for the proper orientation of the substrate/intermediate in the PaPAM active
site.
The stereochemistry of the biosynthetic β-phenylalanine was not however affected as
imagined. The N-(1(S)-camphanoyl) methyl esters of biosynthetic products from Phe445Ala and

93

Phe455Asn mutants (14.48 min and 14.49 min, respectively) eluted at retention times identical to
that of authentic N-[(1S)-camphanoyl]-(3S)-β-phenylalanine (14.47 min) (Figure 3.15)
suggesting that the enantioselectivity was unaffected.

Relative total ion count

1.0
14.49

0.8
14.48

F455N
F455A
(3S)--Phenylalanine

0.6
0.4

14.47

0.2
0.0

12

13

14

15

16

17

18

Time (min)

Figure 3.15. Overlay of gas chromatography profiles of N-[(1'S)-camphanoyl] methyl ester of
authentic (S)-β-phenylalanine (14.47 min) (red), N-[(1'S)-camphanoyl] methyl ester of (S)-βphenylalanine derived from Phe455Ala (14.48 min; blue), and Phe455Asn (14.49 min; green).The N[(1'S)-camphanoyl] methyl ester of (S)-α-phenylalanine used as the substrate in assays is shown at 14.06
and 14.02 min.

These results demonstrate that the difference in stereospecificity of PAMs includes as yet
unknown factors and is not exclusively based on Phe455 or Asn458 residues. Although PaPAM
and TcPAM have an almost similar active site topology, a combination of subtle structural
differences between them are likely of relevance with respect to the mechanism and
stereospecificity of catalysis.
94

3.3.2.3. Mutational Analysis of Thr167 to Assess its Dependence on MIO Construction
Thr167 in the MIO forming motif (167Thr-168Ser-169Gly) of PaPAM is different from
other aminomutases,24,27 and therefore, the catalytic or structural role of Thr167 during PaPAM
catalysis was assessed. Thr167 of PaPAM was systematically mutated to Ala, Cys, Ser, Val, and
Gly167 to prevent significant alteration of the enzyme structure, and the activity of the resulting
mutants was evaluated. The Ala167, Cys167 and Ser167 mutants were designed based on the
occurrence in homologous structures.24,26-28 As mentioned previously, in class I lyase-like family
of enzymes, MIO cofactor is commonly made from an Ala-Ser-Gly sequence. By comparison,
histidine ammonia lyase (HAL) from Streptomyces griseus28 and Fusbocterium sp.26 use CysSer-Gly and Ser-Ser-Gly MIO sequences, respectively. In addition, Thr167Ser was prepared to
assess the role of β-methyl of threonine, while maintaining the hydrogen bonding of the pendant
hydroxyl group. Thr167Val was made to introduce an isosteric residue that lacked the hydrogen
bonding capability. Thr167Gly was an exchange analogous to that of the Ala142Gly mutant of
the HAL from Pseudomonas putida. The latter did not affect on MIO formation nor the
activity.30
3.3.2.3.1. Expression of Mutant Genes
After sequence verification of the mutants, gene expression and Ni-NTA affinity
purification of mutant enzymes were performed similar to those for wt-PaPAM. While
Thr167Gly, Thr167Val and Thr167Cys mutants expressed at similar levels (~50 mg/mL) to wtPaPAM, Thr167Ala and Thr167Ser mutants did not express in E. coli. Therefore, four other
sequence verified colonies harboring Thr167Ala and Thr167Ser mutant papam genes were
analyzed for the expression of m-PaPAM. The total cell protein (TCP) fractions of uninduced
95

(U) and induced (I) cells were compared to assess the expression level of each m-PaPAM
(Figure 3.16). However, none of the Thr167Ala (Figure 3.16A) and Thr167Ser (Figure 3.16B)
mutant clones produced detectable levels of protein from SDS-PAGE. In addition, there was no
indication of truncated protein products, both uninduced and induced TCP fractions were similar
for all the clones expressed in bacteria.

MIO Dependent Aminomutases & PaPAM
A

B
#1

L U

#2
I

#3

U I U I

#4
U

#1
I

U

#2
I

U

#3
I

U

#4
I

U

I

WT L

Figure 3.16. Total cell protein (TCP) fraction analysis of Thr167Ala (A) and Thr167Ser (B)
mutants. SDS-PAGE gels of denatured proteins from uninduced (U) and induced (I) cells from each
clone (#1 − #4) are shown. L: PageRuler™ prestained protein ladder; WT: Induced TCP fraction of wtPaPAM showing expressed PaPAM (59 kDa, dashed square).

It is known that even after careful selection of the vector plasmid and host, recombinant
proteins often express poorly due to various reasons such as toxicity to the host cell,51,52
insolubility,53,54 or codon bias.55 The problem of protein toxicity arises when the recombinant
protein performs detrimental functions in the host cell. Toxic heterologous proteins cause slow
growth rates and death of host cells that ultimately affect the protein expression.51 Nevertheless,
three of the six Thr167 mutants were expressed successfully, suggesting that the recombinant
PaPAM proteins are not toxic to E. coli cells. The presence of codons in the target mRNA that
are rare in E. coli genes is one of the most common reasons for poor heterologous protein
96

expression.55,56 A high frequency of rare codons leads to amino acid misincorporation and/or
truncation of the polypeptide, thus affecting the heterologous protein expression levels.
However, the reading frame of the Thr167 mutants do not contain any rare codons that could
stall the translation machinery in E. coli. In addition, analysis of TCP fractions of mutant clones
indicated that the expressed proteins are not associated with the insoluble fraction, and none of
the clones evaluated indicated that the protein was truncated (cf. Figure 3.16).
Properties of the expression vector can also influence the level of recombinant protein
expression; For example, promoters, regulatory sequences and transcriptional terminators can
affect the expression levels.57 Three of the six mutants were expressed successfully in pET24b(+) vector. Therefore, Thr167Ala mutant papam gene was sub-cloned into a fresh pET24b(+) vector, which also gave no observable expression as assessed by TCP analysis (data not
shown).
The presence of recombinant plasmid DNA and the expression of a recombinant protein
generally impose a metabolic consumption of cellular energy resources.58,59 This added
metabolic burden often causes metabolic, genetic and physiological changes in the host cell. As a
result, conditions such as temperature, growth medium, inducer concentration, and duration of
the induction phase may affect expression levels of target proteins.60-63 To assess the effect of
duration of the induction-phase, protein production of three sequence verified clones from
Thr167Ala and Thr167Ser mutants were analyzed at different times after induction of gene
expression with IPTG. Similar to the aforementioned colony screening, TCP fractions of cells
collected at 1, 2, 3, 5, 7 and 20 h after induction of expression of each clone with IPTG were
compared with those of uninduced cells. Cells harboring wt-PaPAM were used as a positive

97

control for this investigation. Analysis of TCP fractions containing wt-PaPAM showed, as
expected, that the PaPAM production increased with longer induction periods (Figure 3.17A).
The highest PaPAM expression was observed 20 h after induction. However, there was no
protein production seen either with Thr167Ser (Figure 3.17B) or Thr167Ala (Figure 3.17C) over
the duration of the time course, suggesting that no mutant protein was produced or prematurely
degraded. In addition, there was no effect on mutant gene expression with varying temperature.
Thr167Ala and Thr167Ser mutants of PaPAM were not expressed at 20 or 25 C which are
different from the original induction temperature (16 C) for wt-PaPAM.

98

1. Production of T167A Enzyme over Time

A
Clone #1

wt-PaPAM

Clone #2

L U 1 2 3 5 7 20 U 1 2 3 5 7 20

U 1 2 3 5 7 20 L

n of T167S Enzyme
over Time of T167A Enzyme over Time
1. Production

B

C
Clone #1

Clone #2
wt-PaPAM

Clone #3

Clone #1

U L 1 2 3 5 7 U20 1U 21 32 53 75 20
7 20
L L U 1 2 3 5 7 20

L U 1 2 3

Clone #2

Clone #3

U : Uninduced cell pellet
5 7 20 U 1 2 3 5 7 20 U 1 2 3 5 7 20 L
1, 2, 3, 5, 7, 20 hr : Induced cell pellets
L : PageRuler™ prestained protein ladder

59 kDa

Figure 3.17. Protein production levels of wt-PaPAM (A), Thr167Ser (B), and Thr167Ala (C)
mutants. Expressed PaPAM (59 kDa) of induced wt-PaPAM TCP fractions are shown in a dashed
U : Uninduced square.
cell pellet
U : Uninduced
cell pellet
SDS-PAGE gels of TCP fractions
from uninduced
(U) and induced cells from each clone (#1 −
1, 2, 3, 5, 7, 20 hr
:
Induced
cell
pellets
1,
2,
3,
5,
7,
20
hr
:
Induced
cell pellets
#3) at 1, 2, 3, 5, 7 and 20 h induction periods are shown for mutants.
L: PageRuler™ prestained protein
L : PageRuler™ladder.
prestained protein ladder
L : PageRuler™ prestained protein ladder

3.3.2.3.2.Activity and Stability of Expressible Thr167 Mutants
The purified PaPAM mutant enzymes Thr167Val, Thr167Cys, and Thr167Gly were
operationally expressed in soluble form but were shown to be functionally inactive after
incubating each protein separately with (S)-α-phenylalanine. β-Phenylalanine and trans-cinnamic
acid were not detected. The inactivity of the three mutants was likely caused by a non-functional
MIO resulting from mutation of the key Thr167 residue that participate in the formation of this
prosthesis. In addition, the loss of activity could also be attributed to conformational changes in

99

C

U 1

the protein structure upon Thr167 mutation. Since single point-mutations can significantly
change the stability and conformation of a protein,64,65 circular dichroism (CD) spectroscopy66-68
was used to assess the effect of Thr167 mutations on protein conformation and stability. Both
far- (180-250 nm) and near- (260-310 nm) UV CD spectra of mutant proteins were compared
with those of wt-PaPAM (Figure 3.18A, B). The far-UV spectra showed a significant decrease in
negative molar ellipticity at 208 and 222 nm upon replacement of Thr167 by Val and Cys
(Figure 3.18A). According to the molar ellipticity, the fractional helicity of wt-PaPAM,
Thr167Val, and Thr167Cys were 44%, 34%, and 17%, respectively. Similarly, near-UV spectra
indicated a significant change in tertiary structure of PaPAM due to the Thr167 to Val and Cys
mutations (Figure 3.18B). The shape and magnitude of the near-UV CD spectrum of a protein is
a fingerprint of the number of aromatic amino acids and their mobility, interactions with other
residues and the local environment.67 Thus, the Thr167 to Val and Cys mutations likely altered
both secondary and tertiary structural features of PaPAM. However, the Thr167Gly mutation
displayed a fractional helicity (41%) similar to that of wt-PaPAM (44%). Furthermore, the nearUV CD spectrum for this mutant was of similar shape and magnitude as the spectrum for wtPaPAM (Figure 3.18B).
Melting temperature (Tm) curves of all three mutants were also compared with that of wtPaPAM to investigate the relative stability of the mutants (Figure 3.18C). Compared to wtPaPAM (Tm = 57 C), both Thr167Val and Thr167Cys mutants (Tm = 35 C) displayed a 22 C
decrease in their Tm that supported a change in their tertiary structures. The Tm change for
Thr167Gly (Tm = 52 C) was only slight (5 C) suggesting a retention of the native tertiary
conformation. Taken together, these results indicate that the Thr167Val and Thr167Cys

100

mutations significantly affect the secondary and tertiary conformations of PaPAM, while
Thr167Gly mutation had less of an effect.
B

A

150

2 dmol
[[]] (deg
) -1)
cm2 -1
x 10-3cm
dmol
(deg

-4

T167V
T167G
T167C
WT-PaPAM

-6

60000

-8
40000
-10

-12
20000

-14
WT-PaPAM
T167V
T167G
T167C

-16 0
-18
-20000

10

20
200

30

40

50

60

240

-6
100
-8
50

-10
-120
-14
WT-PaPAM
T167V
T167G
T167C

-50

-16
-100
-18

10

70

210 Temperature
220
(oC)230

WT-PaPAM
T167V
T167C
T167G

-4

2 dmol
[[]] (deg
) -1)
cm2 -1
x 10-3cm
dmol
(deg

80000

250

250

260

20

30
260

40

60

70
290

300

310

Wavelength
Wave Length(nm)
(nm)

Wavelength (nm)

C

50

270
Temperature
(o280
C)

[] x 10-3 (deg cm2 dmol-1)

-4
-6
-8

gth20

-10
-12
-14
WT-PaPAM
T167V
T167G
T167C

-16
-18
10

20

30

40

50

60

70

Temperature (oC)

Figure 3.18. Far- (A) and near- (B) UV CD spectra of wt-PaPAM and Thr167 mutants. Each
spectrum represents the average of three plots, which were scanned five times in corresponding
wavelength window. C) Thermal denaturation profiles of wt-PaPAM and Thr167 mutants acquired in far
UV region.

The local environment around of Thr167 was examined to identify the local interactions
of its β-hydroxy and β-methyl groups. The β-hydroxy group is strongly fixed in an extended
hydrogen-bonding network with Gln456, Tyr320, Leu165, Leu171, Gly172, Gly169 and active

101

site water molecules (Figure 3.19). The β-methyl group makes hydrophobic interactions with
Val459 side chain (Figure 3.19). It was therefore hypothesized that removal of the Thr167 side
chain may locally alter the binding of water molecules and disrupt the hydrogen-bonding
network. In addition, the hydrophobic contacts will also be disordered. In turn, this is imagined
to possibly affect the conformation and stability of the protein. Thr167Val mutation, which
significantly altered the protein conformation and stability, would likely retain the hydrophobic
contacts, but disrupt the hydrogen bonding network in PaPAM. Although Thr167Cys mutant
could still make hydrogen-bonding, interactions of the Cys167 are weaker than that of the
Thr167. Sulfhydryl group of Cys167 is a moderately good hydrogen-bond donor and a very poor
hydrogen-bond acceptor,69,70 and, thus, likely alter the conformation and stability of the protein.
Theoretically, Thr167Gly mutation should also disrupt the local interactions similar to
aforementioned mutants. Surprisingly, according to CD spectroscopic investigations, the latter
mutant did not affect the conformation and the stability of PaPAM as much as Thr167Val and
Thr167Cys mutations. Conceivably, binding of water molecules in the void left by Thr167Gly
mutation could maintain the hydrogen-bonding network and maintain the conformational and
stability of PaPAM mutant.
In addition to the effects on the stability and conformation, Thr167 mutants likely affect
the MIO formation. As mentioned before, Thr167 is highly involved in hydrogen-bonding in
PaPAM (Figure 3.19). β-Hydroxy group of Thr167 is hydrogen-bonded to Gly169 of the MIOforming loop via a bound water molecule (HOH242; PDB: 3UNV) in the local vicinity. Thus,
the β-hydroxy group of Thr167 likely fixes the MIO-forming tandem

167

Thr-168Ser-169Gly in a

proper orientation for cyclization and dehydration. It was suggested for GFP,33 PpHAL29,30 and
TcPAM,31 that the MIO-forming loop needs to orient favorably for the intramolecular
102

rearrangement to happen. Furthermore, Gln456, which is hydrogen-bonded to Thr167, makes
hydrogen bonding interactions with the ζ-hydroxy group of Tyr320 (Figure 3.19). Analogous
residue Tyr322 in TcPAM was suggested to position Ala175 (Thr167 in PaPAM) in a proper
location for the nucleophilic attack from Gly177 (cf. Figure 3.6).31 Similarly, in PaPAM, an
extended hydrogen bonding interaction between the side chain of Thr167 and the hydroxyl group
Tyr320 likely positions the latter residue to facilitate the development of the MIO. Presumably,
mutating Thr167 affected the favorable interactions required to from the MIO in PaPAM.

A

B
Tyr320
Leu171
Gln456
MIO

MIO
Val459

Leu165

Val459
Leu165
Gly172

Asn462

Val173

Figure 3.19. Local environment of PaPAM MIO group. Extended hydrogen-bonding network of
Thr167 side chain. MIO is shown in cyan. Atoms are color-coded as carbon: green/yellow; oxygen: red;
nitrogen: blue

103

3.4. Conclusion
The structure of PaPAM was solved in complex with α- and β-phenylpropanoid adducts
with (S)-α- and (S)-β-phenylalanine bound to the active site MIO. These intermediates provide
strong evidence that PaPAM reacts by an alkylamine elimination pathway, which involves
covalent attachment between the α-amino group of the substrate and the MIO cofactor. These
results also indicate that the carbon skeleton of the (S)-α-phenylalanine substrate remains in one
rotameric conformation, while the exocyclic C‒N bond of the NH2–MIO adduct initially
positioned below the α-carbon rotates to a position below the β-carbon to complete the
isomerization reaction. The PaPAM structure also confirms an inversion-of-configuration
mechanism to account for the inverted stereochemistry at each migration terminus.
Residue Phe455 in PaPAM is required for catalysis and likely positions the substrate and
cinnamate intermediate in the active site. However, Phe455 alone does not account for the
stereochemical outcome, since mutation of this residue to the analogously positioned Asn in
TcPAM (that makes the antipode β-isomer) did not change the product stereochemistry of
PaPAM. Future investigations are planned to substitute residues that putatively bind the
substrate and reaction intermediate in PaPAM with those found at the corresponding positions in
TcPAM. This would potentially unravel the phenomenon that causes these similar enzymes to
catalyze reactions of opposite enantioselectivity.
The construction of the MIO prosthesis in PaPAM is highly dependent on the local
hydrophobic and hydrogen bonding interactions, and Thr167 of PaPAM likely orients MIOforming loop in a proper orientation for the autocatalytic post translational modification.

104

REFERENCES

105

REFERENCES
(1)

Walker, K. D.; Klettke, K.; Akiyama, T.; Croteau, R. J. Biol. Chem. 2004, 279, 5394753954.

(2)

Magarvey, N. A.; Fortin, P. D.; Thomas, P. M.; Kelleher, N. L.; Walsh, C. T. ACS Chem.
Biol. 2008, 3, 542-554.

(3)

Christenson, S. D.; Liu, W.; Toney, M. D.; Shen, B. J. Am. Chem. Soc. 2003, 125, 60626063.

(4)

Krug, D.; Muller, R. Chembiochem 2009, 10, 741-750.

(5)

Huang, S. X.; Lohman, J. R.; Huang, T.; Shen, B. Proc. Natl. Acad. Sci. U. S. A. 2013,
110, 8069-8074.

(6)

Moffitt, M. C.; Louie, G. V.; Bowman, M. E.; Pence, J.; Noel, J. P.; Moore, B. S.
Biochemistry 2007, 46, 1004-1012.

(7)

Calabrese, J. C.; Jordan, D. B.; Boodhoo, A.; Sariaslani, S.; Vannelli, T. Biochemistry
2004, 43, 11403-11416.

(8)

Louie, G. V.; Bowman, M. E.; Moffitt, M. C.; Baiga, T. J.; Moore, B. S.; Noel, J. P.
Chem. Biol. 2006, 13, 1327-1338.

(9)

Schwede, T. F.; Retey, J.; Schulz, G. E. Biochemistry 1999, 38, 5355-5361.

(10)

Langer, M.; Pauling, A.; Rétey, J. Angew. Chem. Int. Ed. 1995, 34, 1464–1465.

(11)

Schuster, B.; Retey, J. Proc. Natl. Acad. Sci. U. S. A. 1995, 92, 8433-8437.

(12)

Smith, T. A.; Cordelle, F. H.; Abeles, R. H. Arch. Biochem. Biophys. 1967, 120, 724–
725.

(13)

Givot, I. L.; Smith, T. A.; Abeles, R. H. J. Biol. Chem. 1969, 244, 6341-6353.

(14)

Rechler, M. M. J. Biol. Chem. 1969, 244, 551-559.

(15)

Havir, E. A.; Hanson, K. R. Biochemistry 1968, 7, 1904-1914.

(16)

Wickner, R. B. J. Biol. Chem. 1969, 244, 6550-6552.

(17)

Hodgins, D. S. J. Biol. Chem. 1971, 246, 2977-2985.

(18)

Hanson, K. R.; Havir, E. A. Arch. Biochem. Biophys. 1970, 141, 1-17.
106

(19)

Langer, M.; Reck, G.; Reed, J.; Retey, J. Biochemistry 1994, 33, 6462-6467.

(20)

Schuster, B.; Retey, J. FEBS Lett. 1994, 349, 252-254.

(21)

Langer, M.; Lieber, A.; Retey, J. Biochemistry 1994, 33, 14034-14038.

(22)

Cody, C. W.; Prasher, D. C.; Westler, W. M.; Prendergast, F. G.; Ward, W. W.
Biochemistry 1993, 32, 1212-1218.

(23)

Barondeau, D. P.; Kassmann, C. J.; Tainer, J. A.; Getzoff, E. D. J. Am. Chem. Soc. 2006,
128, 4685-4693.

(24)

Feng, L.; Wanninayake, U.; Strom, S.; Geiger, J.; Walker, K. D. Biochemistry 2011, 50,
2919-2930.

(25)

Ratnayake, N. D.; Wanninayake, U.; Geiger, J. H.; Walker, K. D. J. Am. Chem. Soc.
2011, 133, 8531-8533.

(26)

Kapatral, V.; Anderson, I.; Ivanova, N.; Reznik, G.; Los, T.; Lykidis, A.; Bhattacharyya,
A.; Bartman, A.; Gardner, W.; Grechkin, G.; Zhu, L. H.; Vasieva, O.; Chu, L.; Kogan,
Y.; Chaga, O.; Goltsman, E.; Bernal, A.; Larsen, N.; D'Souza, M.; Walunas, T.; Pusch,
G.; Haselkorn, R.; Fonstein, M.; Kyrpides, N.; Overbeek, R. J. Bacteriol. 2002, 184,
2005-2018.

(27)

Christianson, C. V.; Montavon, T. J.; Van Lanen, S. G.; Shen, B.; Bruner, S. D.
Biochemistry 2007, 46, 7205-7214.

(28)

Wu, P. C.; Kroening, T. A.; White, P. J.; Kendrick, K. E. Gene 1992, 115, 19-25.

(29)

Baedeker, M.; Schulz, G. E. Structure 2002, 10, 61-67.

(30)

Baedeker, M.; Schulz, G. E. Eur. J. Biochem. 2002, 269, 1790-1797.

(31)

Wybenga, G. G.; Szymanski, W.; Wu, B.; Feringa, B. L.; Janssen, D. B.; Dijkstra, B. W.
Biochemistry 2014, 53, 3187-3198.

(32)

Wu, B.; Szymanski, W.; Wybenga, G. G.; Heberling, M. M.; Bartsch, S.; de Wildeman,
S.; Poelarends, G. J.; Feringa, B. L.; Dijkstra, B. W.; Janssen, D. B. Angewandte ChemieInternational Edition 2012, 51, 482-486.

(33)

Barondeau, D. P.; Putnam, C. D.; Kassmann, C. J.; Tainer, J. A.; Getzoff, E. D. Proc.
Natl. Acad. Sci. U. S. A. 2003, 100, 12111-12116.

(34)

Sniegowski, J. A.; Lappe, J. W.; Patel, H. N.; Huffman, H. A.; Wachter, R. M. J. Biol.
Chem. 2005, 280, 26248-26255.

107

(35)

Hermes, J. D.; Weiss, P. M.; Cleland, W. W. Biochemistry 1985, 24, 2959-2967.

(36)

Bordwell, F. G.; Zhao, Y. Y. J. Org. Chem. 1995, 60, 6348-6352.

(37)

Hess, B. A.; Schaad, L. J. J. Am. Chem. Soc. 1983, 105, 7500-7505.

(38)

Gloge, A.; Zon, J.; Kovari, A.; Poppe, L.; Retey, J. Chem-Eur. J. 2000, 6, 3386-3390.

(39)

Louie, G. V.; Bowman, M. E.; Moffitt, M. C.; Baiga, T. J.; Moore, B. S.; Noel, J. P.
Chem. Biol. 2006, 13, 1327-1338.

(40)

Cooke, H. A.; Bruner, S. D. Biopolymers 2010, 93, 802-810.

(41)

Christianson, C. V.; Montavon, T. J.; Festin, G. M.; Cooke, H. A.; Shen, B.; Bruner, S.
D. J. Am. Chem. Soc. 2007, 129, 15744-15745.

(42)

Montavon, T. J.; Christianson, C. V.; Festin, G. M.; Shen, B.; Bruner, S. D. Bioorg. Med.
Chem. Lett. 2008, 18, 3099-3102.

(43)

Greenfield, N. J.; Fasman, G. D. Biopolymers 1969, 7, 595-610.

(44)

Ritter, H.; Schulz, G. E. Plant Cell 2004, 16, 3426-3436.

(45)

Strom, S.; Wanninayake, U.; Ratnayake, N. D.; Walker, K. D.; Geiger, J. H. Angew.
Chem. Int. Ed. 2012, 51, 2898-2902.

(46)

Szymanski, W.; Wu, B.; Weiner, B.; de Wildeman, S.; Feringa, B. L.; Janssen, D. B. J.
Org. Chem. 2009, 74, 9152-9157.

(47)

Wu, B.; Szymanski, W.; Wietzes, P.; de Wildeman, S.; Poelarends, G. J.; Feringa, B. L.;
Janssen, D. B. ChemBioChem 2009, 10, 338-344.

(48)

Dunkley, K. D.; Callaway, T. R.; Chalova, V. I.; Anderson, R. C.; Kundinger, M. M.;
Dunkley, C. S.; Nisbet, D. J.; Ricke, S. C. Anaerobe 2008, 14, 35-42.

(49)

Bartsch, S.; Bornscheuer, U. T. Angew. Chem. Int. Ed. 2009, 48, 3362-3365.

(50)

Mutatu, W.; Klettke, K. L.; Foster, C.; Walker, K. D. Biochemistry 2007, 46, 9785-9794.

(51)

Dong, H. J.; Nilsson, L.; Kurland, C. G. J. Bacteriol. 1995, 177, 1497-1504.

(52)

Doherty, A. J.; Connolly, B. A.; Worrall, A. F. Gene 1993, 136, 337-340.

(53)

Hartley, D. L.; Kane, J. F. Biochem. Soc. Trans. 1988, 16, 101-102.

(54)

Carrio, M. M.; Villaverde, A. J. Biotechnol. 2002, 96, 3-12.
108

(55)

Robinson, M.; Lilley, R.; Little, S.; Emtage, J. S.; Yarranton, G.; Stephens, P.; Millican,
A.; Eaton, M.; Humphreys, G. Nucleic Acids Res. 1984, 12, 6663-6671.

(56)

Gustafsson, C.; Govindarajan, S.; Minshull, J. Trends Biotechnol. 2004, 22, 346-353.

(57)

Francis, D. M.; Page, R. Curr. Prot. Protein Sci. 2010, Chapter 5, 1-29.

(58)

Glick, B. R. Biotechnol. Adv. 1995, 13, 247-261.

(59)

Bentley, W. E.; Mirjalili, N.; Andersen, D. C.; Davis, R. H.; Kompala, D. S. Biotechnol.
Bioeng. 1990, 35, 668-681.

(60)

Wood, T. K.; Peretti, S. W. Biotechnol. Bioeng. 1991, 38, 397-412.

(61)

Chalmers, J. J.; Kim, E.; Telford, J. N.; Wong, E. Y.; Tacon, W. C.; Shuler, M. L.;
Wilson, D. B. Appl. Environ. Microbiol. 1990, 56, 104-111.

(62)

Whitney, G. K.; Glick, B. R.; Robinson, C. W. Biotechnol. Bioeng. 1989, 33, 991-998.

(63)

Neubauer, P.; Hofmann, K.; Holst, O.; Mattiasson, B.; Kruschke, P. Appl. Microbiol.
Biotechnol. 1992, 36, 739-744.

(64)

Cheng, J.; Randall, A.; Baldi, P. Proteins 2006, 62, 1125-1132.

(65)

Parthiban, V.; Gromiha, M. M.; Schomburg, D. Nucleic Acids Res. 2006, 34, 239-242.

(66)

Greenfield, N. J. Nat. Protoc. 2006, 1, 2876-2890.

(67)

Kelly, S. M.; Price, N. C. Curr. Protein Pept. Sc. 2000, 1, 349-384.

(68)

Greenfield, N. J. Nat. Protoc. 2006, 1, 2527-2535.

(69)

Pal, D.; Chakrabarti, P. J. Biomol. Struct. Dyn. 1998, 15, 1059-1072.

(70)

Gregoret, L. M.; Rader, S. D.; Fletterick, R. J.; Cohen, F. E. Proteins 1991, 9, 99-107.

109

CHAPTER 4: Substrate Scope of PaPAM and Effect of Ring Substituents on the
Isomerization Mechanism
The computational modeling investigations related to this study were carried out in
collaboration with Prof. Leslie Kuhn and Ms. Nan Liu from the Departments of Biochemistry
and Molecular Biology, and Computer Science and Engineering at Michigan State University.

4.1. Introduction
4.1.1. Biocatalytic Production of Novel β-Aryl-β-Amino Acids
β-Amino acids, including β-aryl-β-amino acids are widely used in numerous
applications.1,2 Novel thiazolone compounds containing p-fluoro, p-bromo, p-methyl, and o,pdichloro-β-phenylalanine as key components were recently developed as highly potent hepatitis
C virus (HCV) NS5B polymerase inhibitors (Figure 4.1A).3 Inhibition of NS5B controls the
replication of HCV and, thus represents a valid target for antiviral therapy. Additionally, omethyl-(S)-β-phenylalanine derivative 2 was identified as a potent inhibitor for Cathepsin A
(Figure 4.1B), which is a serine carboxypeptidase with potential beneficial applications in
cardiovascular diseases.4
Strategic placement of fluorines in drug molecules is known to enhance
pharmacodynamic and pharmacokinetic properties because of the strong electronegativity, small
size, lipophilicity, and electrostatic interactions of the fluorine. The adventitious effects
dramatically influence the chemical reactivity of pharmaceuticals.5 In addition, metabolic
stability of drug candidates can be increased by replacing aromatic hydrogens with fluorine.6 m110

Fluoro-β-phenylalanine containing dipeptidyl boronic acid 3 (Figure 4.1C) inhibits the 20S
human proteasome at <2 nM level and showed comparable activity to the commercial
counterpart bortezomib.7 m-Fluoro-β-phenylalanine has also been used as an intermediate in the
synthesis of potent chemokine receptor CCR5 antagonist.8
A

B

C

Figure 4.1. β-Aryl-β-amino acids as key components of pharmaceutically important molecules.

Class I lyase-like family aminomutases produce β-aryl-β-amino acids by isomerizing
readily available natural and non-natural α-amino acids.9-11 Phenylalanine aminomutases from
Pantoea agglomerans (PaPAM) and Taxus sp. (TcPAM) isomerize (S)-α-phenylalanine to (S)and (R)-β- phenylalanine, respectively.9,11 TcPAM was shown convert various aryl-substituted
and heteroaromatic α-alanines to corresponding β-amino acids.10 However, tyrosine
aminomutases SgTAM (from Streptomyces globisporus)12 and CmdF (from Chondromyces
crocatus)13 that isomerize (S)-α- to β-tyrosine has a limited substrate scope. Substrate specificity
of SgTAM has been limited only to m-chloro- and m-hydroxytyrosine in addition to its natural
substrate (S)-α-tyrosine.14 A recently discovered aminomutase on the biosynthetic pathway to the
enediyne kedarcidin increased the scope of aminomutases and now includes a catalyst for
making of (R)-2-aza-β-tyrosine from 2-aza-α-tyrosine.15
111

4.1.2. MIO-dependent Isomerization Mechanism of PaPAM
Aminomutases of the class I lyase-like family require the 4-methylidene-1H-imidazol5(4H)-one (MIO) cofactor to carry out the chemically challenging exchange of NH2/H pair.16
The role of MIO cofactor has been long-debated and the highly electrophilic MIO was postulated
to react either with the aromatic ring or the amino group of the substrate.17 However, recent
structural characterization of PaPAM supports the formation of an NH2-MIO adduct, where the
amino group of the substrate is covalently attached to the enzyme during the α/β-isomerization
(Figure 4.2).18 The pro-(3S) proton and the NH2-MIO group are eliminated from the substrate to
form a cinnamate intermediate (released occasionally as a minor by-product), followed by the
hydroamination of the intermediate from NH2-MIO to form the β-amino acid (Figure 4.2).

Figure 4.2. Mechanism of the MIO-dependent isomerization catalyzed by PaPAM.
: the rate at
which the cinnamate by-product is released;
: the rate at which the β-amino acid product is released.

112

4.1.2.1. Elimination of NH2/H from α-Arylalanine Substrate
Phenylalanine aminomutases are mechanistically and structurally closely related to MIOdependent ammonia lyases, which include tyrosine,19 phenylalanine,20,21 and histidine16 ammonia
lyases (TAL, PAL and HAL, respectively). Ammonia lyases catalyze the ammonia elimination
from L-arylalanines (L-tyrosne, L-phenylalanine, and L-histidine) to produce corresponding (E)aryl acrylates (coumaric acid from TAL, cinnamic acid from PAL and urocanic acid from HAL).
Similar to the reactions catalyzed by homologous ammonia lyases, the PaPAM reaction
produces a cinnamate intermediate after elimination of the amino group and pro-(3S) benzylic
hydrogen from the α-amino acid substrate.22
The nucleophilic addition of the α-amine group of the substrate to the MIO is proposed to
make a good alkyl ammonium leaving group.23 Thereafter, α,β-elimination of the pro-(3S) βhydrogen and α-alkyl ammonium can occur via a stepwise or concerted processes. The
concerted, one-step E2 (bimolecular elimination) mechanism leads to the cinnamate intermediate
without involving high-energy charged intermediates. In the two-step E1cB (unimolecular
conjugate-base elimination) mechanism, abstraction of the pro-(3S) proton leads to a benzylic
carbanion intermediate, whereas for the E1 (unimolecular elimination) mechanism, cleavage of
the Cα-N bond results a Cα-carbocation (Figure 4.3). However, the two-step E1 (unimolecular
elimination) reaction is not likely for MIO-dependent reactions. The attached electronwithdrawing carboxylate of the substrate would destabilize the Cα-carbocation formed after
displacement of the NH2-MIO adduct (Figure 4.3).

113

Figure 4.3. Proposed elimination mechanisms for displacement of the NH2-MIO adduct. E1:
unimolecular, E2: bimolecular and E1cB: conjugate-base eliminations

MIO-dependent aminomutase reactions likely follow an E2 or E1cB mechanism, which
require the abstraction of the non-acidic (pKa = 40)24 β-proton by an enzymatic base. The pKa of
the Cβ methylene protons would need to drop ~30 pH units to match the pKa of an unactivated
general base such as a tyrosine residue. The E1cB mechanism has also been accepted for decades
for ammonia lyases.23 However, there are no experimental results that explain how the nonacidic benzylic proton is extracted by an enzymatic base. Molecular modeling studies in
combination with structural analysis of PAL from Rhodosporidium toruloides (RtPAL)
suggested that the influence of α-helix dipole moments favor the development of a carbanion
intermediate and assist the abstraction of the β-protons.25 Most of the active site residues of
RtPAL are associated with α-helices and the positive poles of six of the seven α-helices of the
PAL structure are directed toward the active site.25 Therefore, the helix dipoles of the PAL active

114

site were proposed to likely support the formation of carbanion intermediate produced in the
E1cB mechanism. In addition, electron-withdrawing active site residues that interact with the
phenyl ring of the substrate were suggested to direct electron density away from the phenyl ring
and thus lower the pKa of the β-protons.25 MIO cofactor also resides atop the positive poles of
three helices, and therefore, its electrophilicity was also suggested to increase.
Earlier deuterium isotope studies (kH/kD > 2) on TcPAM suggested that the deprotonation
step of the elimination reaction is rate-determining.26 Electron-withdrawing substituents on the
aryl ring of the substrate that stabilize a δ– charge on Cβ should therefore increase the rate of the
elimination step. Klee and coworkers reported that in a HAL from Pseudomonas sp., 4-nitro
histidine with a KM similar to that of the natural substrate deaminates eight times faster than the
natural substrate L-histidine.27 Furthermore, no deuterium isotope effect was found for the βdideuterated 4-nitro histidine; the β-proton abstraction is no longer the rate determining step. It
was suggested that the nitro group decreased the electron density of the imadazole ring and
therefore increased the acidity of β-hydrogens. By contrast, the β-dideuterated histidine showed a
kinetic isotope effect of 1.5-2.0 with HAL.28 In other ammonia lyase studies, para-nitro
phenylalanine was found to react faster than the un-substituted substrate with PAL from
Petroselinum crispum (PcPAL).29 The wt-PcPAL was inactivated with NaBH4 and, PcPAL
mutant (Ser202Ala that ablated MIO formation) were reported to react 70-fold faster with paranitro phenylalanine than the natural (2S)-α-phenylalanine substrate.29 The highly electron
withdrawing nitro group was posited to increase the acidity of the β-hydrogens and make the
electrophilic MIO cofactor unnecessary for the elimination reaction.
A recent molecular mechanics calculations on TAL suggested that the ligand alignment
in the active site caused a strain in Cα−N and Cβ−pro-(3S)−H bonds and therefore an E2
115

elimination is favored over other mechanisms.30 According to this study, both E1 and E1cB
intermediates or transition states were not feasible in the TAL deamination reaction.
4.1.2.2. Hydroamination of the Cinnamate Intermediate
The final reaction sequence of the MIO-dependent aminomutases involves an α,βaddition reaction, where the NH2-MIO and a proton (H+) add across the double bond of the
acrylate intermediate. To obtain the β-amino acid from the hydroamination, the polarity of the Cβ
(δ+) needs to be opposite of that in the earlier elimination sequence. Here, the nucleophilic NH2MIO binds to Cβ and the electrophilic H+ attaches to Cα (Figure 4.4).

Figure 4.4. Hydroamination of the acrylate intermediate. Shown is a transition state intermediate
highlighting the polarization of the π-bond in which the nucleophilic NH2-MIO and the electrophilic H+
approach Cβ and Cα, respectively.

Similar to the ammonia elimination, hydroamination reaction of PaPAM can follow
either concerted or stepwise addition mechanisms (Figure 4.5). In the concerted mechanism, the
NH2-MIO and the pro-(3S) proton abstracted from the catalytic base add simultaneously onto the
β- and α-carbons, respectively (Figure 4.5, path a). Alternatively, PaPAM could use a stepwise
addition sequence where the nucleophile (NH2−MIO) or the electrophile (TyrO∙∙∙∙H) first add at

116

the respective carbons of the arylacrylate double bond. In the nucleophilic addition mechanism,
NH2−MIO couples to form a 1,4-Michael adduct (Figure 4.5, path b). However, a presumed
resonance structure has two repelling oxyanions on the carboxylate of the reactant that normally
forms a monodentate salt bridge (Figure 4.5, path b), as evidenced in the PaPAM crystal
structure.18 This conjugate addition route benefits from an electropositive (δ+) Cβ by delocalizing
the π-electrons towards the carboxylate of the substrate. Theoretically, a substituent that places
negative charge inductively within the ring or mesomerically on Cipso of the β-arylacrylate
intermediate should also strengthen the formation of a δ+ on Cβ.

Figure 4.5. Concerted (A) and stepwise (B) hydroamination mechanisms of PaPAM. Route a)
Concerted hydroamination of the acrylate π-bond. Shown is an intermediate with maximal charge
separation between repelling negative charges in the carboxylate group and the cation and anion. Route b)
A stepwise Michael-addition pathway. Shown is an intermediate adduct with the π-electrons delocalized
into the carboxylate group forming a repelling dianion prior to Cα-protonation. Route c) A stepwise
hydroamination sequence. Shown is a proposed intermediate resulting from Cα-protonation as the first
step, which places a positive charge at Cβ, and Cβ is now primed for nucleophilic attack by the NH2-MIO
adduct.

117

An electrophilic addition pathway is envisioned to first add the pro-(3S) proton at Cα of
the acrylate intermediate (cf. Figure 4.5, path c). The resulting intermediate has a positive charge
(δ+) on the benzylic Cβ, which is resonance stabilized by the aryl ring and further stabilized by
electron-releasing substituents (cf. Figure 4.5, path c). Rapid, nucleophilic attack by the NH2MIO on the carbocation would ensue to complete the β-amino acid catalysis.
The second half of the aminomutase reaction has been employed to synthesize α- and βamino acids from arylacrylates.31,32 Phenylalanine aminomutase from Taxus chinensis (TchPAM)
catalyzes the highly enantioselective amination of variously substituted cinnamic acids from
NH4OH.31 It has been reported that the regioselectivity of ammonia addition on the acrylate
double bond is dependent on the substituents on the phenyl ring.32 Electron rich aromatic rings
with the electron donating substituents showed higher β-selectivity due to the lowered ability of
the aromatic ring to accept electrons.32 The carboxylate group is suggested to act as the electron
sink in this reaction, whereas, electron deficient aromatic rings favor the α-addition and accepts
electron delocalization into the aromatic ring.32
While this study above provided information about the partitioning of cinnamate to αand β-phenylalanine it did not assess the substituent effects on the forward reaction of PAMs.
The current study evaluated the substituent effects on α- to β-phenylalanine conversion by
PaPAM that involves both elimination and hydroamination steps.

118

4.2. Experimental

4.2.1. Substrates, Authentic Standards and Reagents
(S)-α-, p-Methoxy-(S)-α-, p-nitro-(S)-α-, and p-chloro-(R/S)-β-phenylalanine and (E)-omethyl-, (E)-p-methyl-, (E)-p-methoxy- and (E)-p-nitro-cinnamic acid, (E)-o-furyl-acrylate and
(trimethylsilyl)diazomethane (2.0 M in diethyl ether) were purchased from Sigma-Aldrich-Fluka
(St. Louis, MO). Racemic p-nitro-β-phenylalanine was purchased from Oakwood Products, Inc.
(West Columbia, SC), and o-methoxy-(S)-α-, m-methoxy-(S)-α-, o-nitro-(S)-α-, m-nitro-(S)-α-, omethoxy-(S)-β-, m-methoxy-(S)-β-, o-nitro-(S)-β-, and m-nitro-(S)-β-phenylalanine were
purchased from Chem-Impex International, Inc. (Wood Dale, IL). 2-Amino-5-phenylpentanoic
acid was purchased from Acros Organics (New Jersey). All other (S)-α- and β-amino acids were
purchased from PepTech Corporation (Burlington, MA) and the other (E)-cinnamic acids were
purchased from Alfa Aesar (Ward, Hill, MA). All chemicals were used without further
purification, unless noted.

4.2.2. General Instrumentation
GC-MS analysis was performed with an Agilent 6890N gas chromatograph equipped with a
capillary GC column (30 m × 0.25 mm × 0.25 μM; HP-5MS; J&W Scientific) with helium as the
carrier gas (flow rate, 1 mL/min). The injector port (at 250 °C) was set to splitless injection
mode. A 1-μL aliquot of each sample was injected using an Agilent 7683 auto-sampler (Agilent,

119

Atlanta; GA). The column temperature was increased from 50 – 110 °C at 30 °C/min, then
increased by 10 °C/min to 250 °C (total run time of 16 min), and returned to 50 °C over 5 min,
with a 5 min hold. The gas chromatograph was coupled to a mass selective detector (Agilent,
5973 inert) operated in electron impact mode (70 eV ionization voltage). All spectra were
recorded in the mass range of 50 – 400 m/z.

4.2.3. Expression and Purification of PaPAM
Luria-Bertani medium (1 L) supplemented with kanamycin (50 μg/mL) was inoculated with 5
mL of an overnight culture of E. coli BL21(DE3) cells engineered to express the papam cDNA
from the pET-24b(+) vector as a C-terminal His6-tagged PaPAM. These cultures were grown at
37 °C to an optical density of A600 ~ 0.6. PaPAM expression was induced with isopropyl-β-Dthiogalactopyranoside (100 μM) at 16 °C and the cultures were grown for 16 h. The subsequent
steps were performed at 4 °C, unless indicated otherwise. Cells were harvested by centrifugation
at 6,000g (15 min) and the cell pellet was resuspended in lysis buffer (50 mM sodium phosphate
buffer containing 5% (v/v) glycerol, 300 mM NaCl and 10 mM imidazole, pH 8.0). The cells
were lysed by sonication (Misonix sonicator, Farmingdale, NY), and the lysate was centrifuged
at 9,700g (45 min) and then at 102,000g (1 h) to remove cell debris and light membranes. The
resultant crude, C-terminal His6-tagged aminomutase in the soluble fraction was purified by
Nickel-nitrilotriacetic acid (Ni-NTA) affinity chromatography according to the protocol
described by the manufacturer (Qiagen, Valencia, CA). PaPAM fractions, eluting in 250 mM
imidazole, were concentrated by size-selective centrifugal filtration (Centriprep centrifugal filter
units, 30,000 MWCO; Millipore), the buffer was exchanged with 50 mM sodium phosphate
120

buffer containing 5% (v/v) glycerol (pH 8.0). The purity of the concentrated enzyme was
assessed by SDS−PAGE with Coomassie Blue staining, and the quantity was determined by the
Bradford protein assay. The overexpressed PaPAM (~59 kDa) was obtained at 95% purity (~25
mg/L).

4.2.4. Assessing the Substrate Specificity of PaPAM for (2S)-α-Phenylalanine Analogues
(S)-α-Phenylalanine and each of its analogues (1 mM) were incubated for 2 h with PaPAM
(50 μg) in 1-mL assays of 50 mM phosphate buffer (pH 8.0) containing 5% glycerol. Control
assays contained all ingredients except either the substrate or enzyme was omitted. Each reaction
was quenched by acidifying to pH 2-3 (6 M HCl). Three internal standards (m-fluoro-βphenylalanine, p-methyl-β-phenylalanine and β-phenylalanine at 20 μM) were used, respectively,
to quantify three sets of biosynthetic β-amino acids products–Set 1: β-phenylalanine; o-, m-, and
p-methyl-; o-, m-, and p-methoxy-; m- and p-nitro-; m- and p-chloro-β-phenylalanine; and (2furyl)-β-alanine; Set 2: o- and p-fluoro; m-, and p-bromo-β-phenylalanine; and (2-thienyl)- and
(3-thienyl)-β-alanine; and Set 3: m-fluoro-β-phenylalanine. Two internal standards (pmethylcinnamic acid and cinnamic acid at 20 μM) were used, respectively, to quantify two sets
of biosynthetic aryl acrylic acid products: Set 1, cinnamic acid, o-, m-, and p-fluorocinnamic
acid, and (2-thienyl)- and (3-thienyl)-acrylic acid; Set 2, o-, m-, and p-methyl-; o-, m-, and pmethoxy-; m- and p-nitro-; m- and p-chloro-; m- and p-bromo-cinnamic acid and (2-furyl)-acrylic
acid. After acidifying the reactions solution, the aryl acrylates were extracted with diethyl ether
(2 × 2 mL). The remaining aqueous fractions were basified to pH 10 (6 M NaOH) and treated
with ethylchloroformate (50 μL) for 10 min. Each reaction was basified again to pH 10, a second
121

batch of ethylchloroformate (50 μL) was added, and each was stirred for 10 min. The solutions
were separately acidified to pH 2-3 (6 M HCl), extracted with diethyl ether (2 × 2 mL). For each
sample, the diethyl ether fractions were separately combined. The organic fraction was removed
under vacuum, and the resulting residue was dissolved in ethyl acetate:methanol (3:1, v/v) (200
μL). The solution was treated with excess (trimethylsilyl)diazomethane until the yellow color
persisted. The derivatized aromatic amino acids and aryl acrylates were quantified by GC/EIMS. The peak area was converted to concentration by solving the linear equation obtained from
the standard curves constructed with the corresponding authentic standards quantified by GC/EIMS (Figure A.2.1-A.2.19).

4.2.5. Kinetic Parameters of PaPAM for (2S)-α-Phenylalanine Analogues
PaPAM (10, 25, 50 or 100 µg/mL) was incubated with each productive substrate (1000, 2000
or 2250 μM) in 12-mL assays to establish linearity with respect to time at a fixed protein
concentration at 31 °C. Aliquots (1 mL) were withdrawn from each assay at 0.5 h intervals over
5 h and the reactions were quenched by adding 6 M HCl (100 μL). The products were
derivatized and quantified as described above, and steady state conditions for each substrate were
determined. To calculate the kinetic constants, each substrate was varied (10 – 2250 μM) in
separate assays under the predetermined steady state conditions. Resultant products were
quantified after terminating the reaction as described previously. Kinetic parameters (KM and kcat)
were determined from Hanes-Woolf plots by plotting [S]/v against [S] (R2 = 0.97 – 0.99) (Figure
A.2.20-A.2.38).

122

4.2.6. Inhibition Assays for Non-productive Substrates
(2S)-α-Phenylalanine (at 10, 20, 40, 80, 100, 200, 300, 500, 750, 1000 μM) and PaPAM (10
μg, 0.17 nmol) were mixed and incubated separately for 40 min with non-productive substrates
o-chloro-, o-bromo-, or o-nitro-(S)-α-phenylalanine (at 50, 100 and 200 μM). The products were
derivatized and quantified as described earlier. Inhibition constants (KI) were calculated by nonlinear regression analysis using GraphPad Prism 6 Software (La Jolla, CA).

123

4.3. Results and Discussion

4.3.1. Substrate Scope of PaPAM and Kinetic Parameters of α-Phenylalanine Analogues
To gain further insights into the mechanism of PaPAM, the substrate specificity was
queried with 18 α-phenylalanine analogues and three heteroaromatic compounds. The
substituents on the phenyl ring varied in position, size, inductive and mesomeric effects,
polarizability, hydrophobicity, and the ability to form hydrogen- and halogen-bonds. PaPAM
was productive with a broad range of α-arylalanine analogues including fluoro-, chloro-, bromo-,
methyl-, methoxy-, and nitro-substituents on the phenyl ring. Additionally furyl- and thienylalanines with heteroaromatic rings were also isomerized to their respective β-arylalanine
products (Figure 4.6). Therefore, in contrast to tyrosine aminomutases with a narrow substrate
scope,14 both PaPAM and TcPAM10 have broader steric and electronic flexibility for its
isomerization reaction. Interestingly, ortho-chloro, -bromo, and -nitro substituted analogues were
non-productive substrates for PaPAM while same substituents at the meta- and para-positions
were productive.

Figure 4.6. Broad substrate scope of PaPAM. Various aryl-substituted and heteroaromatic α-amino
acids were productive substrates for PaPAM.
124

The kinetic constants kcat and KM of all the productive substrates were calculated from the
Hanes-Wolf plots (Figure A.2.20-A.2.38) constructed by measuring the product formation of
each substrate at steady state conditions. The kinetic parameters of PaPAM for the natural
substrate α-phenylalanine (1) were used to compare against the values for each analogue (2 –
22). In general, natural substrate 1 had the highest catalytic efficiency (kcat/KM) of all the
substrates tested. The relative catalytic efficiency for each analogue was negatively affected by a
decrease in kcat total and/or increase in KM (Table 4.1). Except the para-fluoro substrate 5, all
other para-substituted analogues showed a lower catalytic efficiency as a series. ortho-Methoxy
substrate 19, the only productive bulkier group at ortho-position had the lowest catalytic
efficiency due to its exceptionally low catalytic rate (0.003 s−1).
Fluoro-substituted analogue series displayed interesting kinetic parameters. All three
ortho- (10), meta- (3), and para- (5) substituted analogues had similar catalytic rates (0.022,
0.031, and 0.023 s−1, respectively) regardless of the substituent position. Furthermore, they
bound to PaPAM better than the natural substrate 1 and all the other substrate analogues. Of all
the substrates tested, ortho-fluoro analogue 10 showed the lowest binding constant of 0.027 ×
10−3 M and para-fluoro substrate had a comparative binding affinity (0.029 × 10−3 M).
Fluorinated molecules are often exploited in medicinal chemistry to enhance the ligand binding
to their target proteins.33,34 It has been suggested that the fluorine substitution leads to an
enhancement of binding affinity due a combination of its smaller size (~1.5 Å), higher
electronegativity, and increased lipophilicity of the molecule.33 Thus, fluoro-α-phenylalanines
could also bind well to PaPAM likely due to these reasons.

125

4.3.2. Kinetic Parameters of ortho-Substituted Analogues
Of the three productive ortho-substrates, the relative catalytic efficiency was highest for
o-methyl substrate 6 (0.73 compared to 1.93 s−1•M−1 × 103 for 1). o-Methoxy substituent 19
showed the lowest catalytic efficiency of all the substrates (0.02 s−1•M−1 × 103). The KM values of
PaPAM for each of the three productive ortho-substrates (6, 10, and 19) varied only between 1and 2-fold compared to that of 1. The ortho-substituents, regardless of size, including the bulkier
o-methoxy of 19, did not affect the substrate binding. Of the three, PaPAM turned over o-methyl
substrate (6) faster (0.064 s−1) than the o-fluoro (10, 0.022 s−1) and o-methoxy (19, 0.003 s−1)
compounds (Table 4.1). However, each was isomerized substantially slower (5-, 14-, and 108fold, respectively) than 1. The relatively satisfactory binding (i.e., low KM values) yet poor
turnover for 6, 10, and 19 suggests that PaPAM binds these substrates in a catalytically
ineffective orientation. It should be noted that the ortho-substituents on the arylalanine substrates
are positioned vicinally to the alanine side chain. The proximity of these groups to the alanyl side
chain of the substrates likely creates a steric barrier that skews the aryl ring plane. A canted aryl
ring would relax the sterics yet reduce potentially beneficial resonance effects of the substituents
on Cβ in a charged transition state that could influence substrate turnover. Steric shielding of the
β-position of the cinnamate intermediate by o-fluoro, -chloro, -bromo and -methyl substituents
was also seen during the ammonia addition reactions with TcPAM.31 ortho-substituted
cinnamates added NH2 exclusively at Cα likely due to the unfavorable steric effects caused at Cβ
from proximal substituents.
Although the o-bromo, o-chloro, and o-nitro substrates 20 – 22 exert similar electronic
effects as the corresponding para-isomers (Figure 4.7D, J), interestingly, 20 – 22 did not yield

126

any detectable product in the enzyme reaction. However, their competitive inhibition constants
(KI) of 15.9 (±1.67), 17.7 (±2.11), and 16.9 (±3.35) µM, respectively, indicate that they bind well
to PaPAM. The lack of turnover of 20 – 22 by PaPAM was therefore likely caused by poor
access of the substrates to a catalytically competent conformation. This may have resulted from
the bulkier substituents at the ortho-position of the phenyl ring. In contrast to 20 – 22, o-methoxy
substrate 19 turned over with PaPAM, yet with the lowest catalytic rate. The productivity of 19
was likely caused by the favorable binding orientation resulted from the H-bonding interactions
of o-methoxy substituent with active site residue Tyr320 (Figure A.2.39).
4.3.3. Kinetic Parameters of meta-Substituted Analogues
The relative catalytic efficiencies were highest for meta-halogenated substrates (2 – 4)
(Table 4.1). The KM values of PaPAM for m-bromo (2) and m-chloro (4) substrates were only
negatively affected ~2-fold, and the

values remained essentially unchanged compared to

the parameters for 1 (Table 4.1). Interestingly, the relative

for the m-fluoro substrate 3 was

~10-fold lower (0.031 s−1) than that for 1, 2 and 4, yet the 5-fold lower KM of PaPAM for 3 made
the

similar to those for 1, 2 and 4. The latter suggests that 3 binds tighter than 2 and 4,

which carry halogens (Br and Cl) with larger atomic radii of 185 pm and 175 pm, respectively,
compared to the smaller F (147 pm) of 3. In addition, the fluoro-substituted substrates binds
better than the natural substrate containing a smaller H-atom.
Analysis of other meta-substituted substrates showed the catalytic efficiencies for m-nitro
(9), m-methoxy (11), and m-methyl (13) analogues were 6- to 10-fold lower than that for 1. The
m-nitro of 9 only reduced the relative
and 2.6-fold negative effects on

of PaPAM by 5.7-fold due to the modest 2.2and KM, respectively, compared with 1 (Table 4.1). m127

Methoxy substrate 11, showed the second highest KM (990 µM) of all the substrates while the
was only ~1.6-fold lower than that of 1. The increased KM suggested that the sterics of the
m-methoxy affected substrate binding. In contrast,

of m-methyl substrate 13 was ~6-fold

lower (compared to 1) with a KM negatively affected only by 1.2-fold. Consequently, ~6-fold
negative effects of KM and

, reduced the catalytic efficiency of 11 and 13, respectively.

4.3.4. Kinetic Parameters of para-Substituted Analogues
With the exception of p-fluoro substrate 5, catalytic efficiencies of para-substituent
analogues (14 – 18) were the lowest as a series. Each substrate containing a para-substituent (5,
14 – 18), significantly reduced the
(

of PaPAM by 6 – 25-fold compared to the value for 1

= 0.323 s−1). As seen for the trend with the meta-substituent series, the p-bromo and p-

chloro substituents were turned over fastest; the chloro substrate was turned over slightly faster.
The substrates turned over slowest by PaPAM in this series contained a p-nitro, p-methyl, or pmethoxy substituent (Table 4.1). In general, para-substituted analogues bound poorly to PaPAM
with p-methoxy analogue being the substrate with the lowest affinity (KM = 1187 µM; a 10-fold
increase compared to 1). However, similar to ortho- and meta-substituted analogues, p-fluoro
substrate showed the highest catalytic efficiency of the series due to its unusually low KM value.
In addition to 5, p-methyl substituent also had a lower KM value (163 µM) compared to other
analogues in the series (Table 4.1).

128

Table 4.1. Kinetic parameters of PaPAM for various substituted aryl and heteroaromatic
substrates

R
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19

/KM
(s−1•M−1 × 103)

KM (µM)
168
(± 7)
339
(± 15)
27
(± 5)
432
(± 26)
29
(± 1)
88
(± 6)
415
(± 79)
337
(± 27)
430
(± 15)
73
(± 6)
990
(± 124)
132
(± 5)
204
(± 4)
491
(± 82)
525
(± 44)
163
(± 9)
752
(± 39)
1187
(± 76)
164
(± 7)

20

0.301
92.8%
0.396
93.9%
0.027
85.2%
0.462
95.2%
0.020
85.7%
0.055
83.6%
0.143
34.8%
0.139
97.2%
0.136
92.6%
0.021
95.5%
0.201
99.0%
0.024
90.9%
0.048
78.3%
0.050
94.1%
0.043
95.6%
0.010
63.6%
0.025
48.0%
0.022
97.7%
0.002
70.0%

0.022
7.2%
0.024
6.1%
0.004
14.8%
0.022
4.8%
0.003
14.3%
0.009
16.4%
0.093
65.2%
0.004
2.8%
0.01
7.4%
0.001
4.5%
0.002
1.0%
0.002
9.1%
0.010
21.7%
0.003
5.9%
0.002
4.4%
0.003
36.4%
0.013
52.0%
0.0005
2.3%
0.0007
30.0%

21

0.323
(± 0.013)
0.420
(± 0.014)
0.031
(± 0.002)
0.484
(± 0.02)
0.023
(± 0.001)
0.064
(± 0.002)
0.236
(± 0.01)
0.143
(± 0.004)
0.146
± 0.003)
0.022
(± 0.001)
0.203
(± 0.012)
0.026
(± 0.001)
0.058
(± 0.001)
0.053
(± 0.003)
0.045
(± 0.001)
0.013
(± 0.001)
0.038
(± <10–3)
0.022
(± <10–3)
0.003
(± <10–3)
22

20-22: Non-productive substrates

129

1.93
(± 0.20)
1.24
(± 0.12)
1.2
(± 0.4)
1.12
(± 0.14)
0.79
(± 0.06)
0.73
(± 0.09)
0.588
(± 0.066)
0.428
(± 0.063)
0.340
(± 0.025)
0.31
(± 0.04)
0.209
(± 0.050)
0.19
(± 0.02)
0.19
(± 0.01)
0.11
(± 0.03)
0.09
(± 0.01)
0.082
(± 0.010)
0.050
(± 0.005)
0.019
(± 0.002)
0.02
(± <10–2)

4.3.5. Substituent Effects on the PaPAM Isomerization Reaction
4.3.5.1. Electronic Considerations on PaPAM Isomerization Reaction
The PaPAM isomerization mechanism is a sequence of two main reaction steps that
involves an elimination followed by a hydroamination step (cf. Figure 4.2). In the first reaction
sequence, the amino group and the pro-(3S) benzylic hydrogen from the α-amino acid eliminate
to produce the cinnamate intermediate.18 MIO-dependent aminomutase reactions likely follow an
E2 or E1cB elimination mechanism (Figure 4.7A), where both depend on the rate of
deprotonation of Cβ, as proposed in an earlier work.35 The electron-withdrawing substituents on
the aryl ring of the substrate that stabilize a δ– charge on Cβ should therefore increase the rate of
the elimination step (Figure 4.7C, D).
The initial elimination sequence of MIO-containing aminomutases requires the
abstraction of a proton from a site of low C-H acidity (pKa = 40).24 The alkylation of the αamino group from the MIO cofactor, however leads to the acidification of the β-hydrogens and
favors the elimination.23 Although this is sufficient to allow the abstraction of benzylic
hydrogens, no MIO-based aminomutases reported to date can isomerize aliphatic amino
acids.10,14

PaPAM

was

unable

to

convert

(S)-2-amino-4-phenylbutyric

acid

(L-

homophenylalanine) and (S)-2-amino-5-phenylpentanoic acid to their corresponding β-amino
acids. Presumably, additional methylene group/s between the α-amino group and the phenyl ring
of these substrates disrupt the potentially beneficial effects of the aromatic ring for proton
abstraction. Thus, difficulty of methylene β-hydrogen (pKa = 60) abstraction likely resulted the
non-productivity of these substrates. (S)-styrylalanine was however, a productive substrate of
PaPAM although the β-styrylalanine ratio was much lower (10% β-styrylalanine; 90%
130

styrylacrylate). This suggests that an extended conjugated allyl π-system next to the α-carbon, is
required for the α- to β-phenylalanine isomerization.
The final reaction sequence of the MIO-dependent aminomutases involves an α,βaddition reaction, where the NH2-MIO and a proton (H+) add across the double bond of the
acrylate intermediate (Figure 4.7B). To obtain the β-amino acid in a concerted hydroamination,
the polarity of the Cβ (δ+) needs to be opposite of that in the earlier elimination sequence. Here,
the nucleophilic NH2-MIO binds to Cβ and the electrophilic H+ attaches to Cα (Figure 4.7B).
Theoretically, a substituent that places negative charge inductively within the ring or
mesomerically the on Cipso of the β-arylacrylate intermediate should strengthen the formation of
a δ+ on Cβ. Therefore, the electron-donating substituents on the aryl ring of the substrate that
stabilize a δ+ charge on Cβ should increase the rate of the hydroamination step (Figure 4.7G, H).
To further evaluate the mechanistic basis of the differences in turnover by PaPAM for various
meta- and para-substituted substrates, the dependence of the relative turnover rate on the
substituent of the substrate was analyzed using Hammett plots.

131

A

B

Favorable Resonance Hybrids
C

D

G

H

Unfavorable Resonance Hybrids
E

F

I

J

Figure 4.7. Elimination (A) and Hydroamination (B) reaction sequences of PaPAM isomerization.
C-J) Resonance hybrids formed from electron-donating (D) or -withdrawing (E) substituents on the
phenyl ring. An increase or decrease in electron density, caused by the substituent, within the ring or at
Cipso is predicted to support a transient δ+ or δ– respectively at Cβ.

132

4.3.5.2. Substituent effects on Michaelis Parameters: meta-Substituents
The Hammett plot between the calculated

of PaPAM and substituent

constants (σ)36 for the meta-substituted (mX) arylalanines (m-bromo (2), m-chloro (4), m-nitro
(9), m-methoxy (11), and m-methyl (13)) follow a concave-down, parabolic regression curve37
(Figure 4.8A). The fastest reactions at the apex of the curve occurred with the m-bromo and mchloro substrates, and the slowest with m-methyl and m-nitro, at the extremes. The m-methoxy
substituent reacted at an intermediate rate. The right side of the correlation plot tends towards a
slope (ρ) < 0 and suggested the rate of the PaPAM reaction was slowed by electron-withdrawing
substituents. m-Bromo (2) and m–chloro (4) substrates however occupy an ambiguous position at
the apex of the Hammett plot (Figure 4.8A). Electron-donating substituents appear on the
Hammett correlation plot where the slope (ρ) is ≈ +2.9 suggesting that the PaPAM rate was
markedly slowed by stronger electron-donating meta-substituents.
4.3.5.2.1. Electron-withdrawing Substituents
The

for m-nitro substrate 9 fits on the negative slope ( ≈ –1.4) of the

correlation curve (Figure 4.8A). The resonance hybrid of the m-nitro places δ+ in the aromatic
ring favoring a buildup of δ− on Cβ (Figure 4.7C). Thus, the strong electron-withdrawing m-nitro
group is foreseen to accelerate deprotonation of Cβ that produces a transient δ– on the elimination
step (Figure 4.7A). In turn, the nitro group was anticipated to negatively affect the β-amination
step, which forms a transient δ+ on Cβ (Figure 4.7B, J). The m-nitro was therefore expected to
slow the amination rate of the PaPAM reaction involving 9 compared to that for 1. The mnitrocinnamate (7.4%)/m-nitro-β-amino acid (92.6%) product apportioned similar to that of
analogous products made from 1 (cinnamate (7.2%) and β-amino acid (92.8%)). This result
133

suggested that the m-nitro of 9 likely had a less than imagined affect on the hydroamination of
the m-nitrocinnamate intermediate.
A

B
0.4

-2.6
Cl

0.2
0.0

F

H

-0.4

log(kcat/KM)

OCH3
NO2

-0.6

-1.0

0.0

-3.4
NO2

-3.8

F

-1.2
-0.2

-3.2

-3.6

CH3

-0.8

Br
Cl

-3.0

-0.2

log(kx/kH)

H

-2.8

Br

0.2

0.4

0.6

-4.0
-0.2

0.8

 (meta-Substituents)

CH3

0.0

OCH3

0.2

0.4

0.6

0.8

 (meta-Substituents)

Figure 4.8. Dependence of the observed
[designated as log(kx/kH)] (A) and
[designated as log(kcat/KM)] (B) on the Hammett substituent constant for the PaPAM-catalyzed
isomerization of meta-substituted α-arylalanines. Here,
is
for entries 2 – 4, 9, 11, and 13;
is
for entry 1. Open circle: outlier m-fluoro substrate 3. Open diamond: outlier m-methoxy substrate
11.

In contrast, m-bromo (2) and m-chloro (4) substrates turned over ~3-fold faster than 9
(Table 4.1). Halogens are a group of substituents of the ‘push-pull’ type. They withdraw electron
density by induction and donate electrons by resonance, depending on the type of reaction. The
overall effect of the halogens is considered electron-withdrawing as estimated by their Hammett
substituent constants. The ‘push-pull’ effect of 2 and 4 likely tells that electron-release by mbromo and m-chloro reduces the electron-withdrawing magnitude that negatively affects the rate,
as did the m-nitro of 9. The electron-withdrawing effect of bromo and chloro likely support the
transient δ– on Cβ and increases the rate of the elimination step (Figure 4.7C); the electrondonating effect would improve stabilization of a transient δ+ formed during the hydroamination
134

across the double bond of the intermediate (Figure 4.7G). These effects likely caused an increase
in

of substrates 2 (0.420 s–1) and 4 (0.484 s–1) compared to that for natural substrate 1

(0.323 s–1).
It is worth noting, the proportions of m-halo-β-amino acid (93.9% m-bromo-β-amino acid
and 95.2% m-chloro-β-amino acid) and m-halo-cinnamate (6.1% m-bromo-cinnamate and 4.8%
m-chloro-cinnamate) made by PaPAM from 2 and 4, respectively, were similar to that of
analogous products made from 1 (Table 4.1). Thus, the amination of the m-halocinnamate
intermediates was not significantly affected by the substituents. This observation supports a
mechanism where release of the intermediate as a by-product is slower than hydroamination.
Interestingly, based on Hammett constants, the inductive effects of fluoro (σ = 0.34) on
an aryl ring are in principle similar to those of the chloro- (σ = 0.37) and bromo- (σ = 0.39)
substituents.36 Therefore, it was surprising that the m-fluoro substrate 3 had a significantly lower
value and did not fit the Hammett correlation for the meta-substituent series
(Figure 4.8A). The significant decrease (~10-fold) in

and

of PaPAM for 3 (compared

with the same parameters for 1) suggested that the m-fluoro substituent affected the chemistry at
Cβ during the elimination and the hydroamination steps, respectively. The higher proportion of
m-fluorocinnamate (14.8%) relative to m-fluoro β-amino acid (85.2%) made by PaPAM from 3,
(compared with the cinnamate (7.2%) and β-amino acid (92.8%) products made from 1)
suggested that the electronic effects of the m-fluoro affected the amination step more than the
elimination step.

135

4.3.5.2.2. Electron-donating Substituents
m-Methoxy and m-methyl substrates 11 and 13, respectively, appear on the Hammett
correlation plot where the slope (ρ) is ≈ +2.9 (Figure 4.8A). This suggested that the PaPAM rate
was markedly slowed by stronger electron-donating meta-substituents.
The electron-donation by m-methoxy substituent 11 likely destabilizes the δ– on Cβ upon
removal of Hβ (Figure 4.7E) and decreases the elimination rate. Reciprocally, the m-methoxy
substituent would promote the formation of an electrophilic (δ+) Cβ formed during the amination
step (Figure 4.8G). Here, the electronic effects of the m-methoxy that deterred the elimination
rate were likely offset by the rate-enhancing effects on the amination step. The PaPAMcatalyzed product pool from 11 contained the m-methoxy-β-amino acid (99.0%) and the mmethoxy acrylate at 1.0%. PaPAM converted 1 with less selectivity (β-isomer at 92.8%;
cinnamate at 7.2%). This supported that the amination efficiency to make the β-isomer of 11 was
most likely facilitated by the substituent. In addition, based on the Hammett constant (+0.12),36
m-methoxy has an electron-withdrawing component that slightly reduces the significant metasubstituent effect of its electron-donation into the ring by resonance. m-Methoxy substrate 11
showed a moderate

with only a 1.6-fold reduction compared to 1. This supported the

balancing effects of electron-donating and -withdrawing ability of the m-methoxy substituent on
.
A methyl-substituent contributes electron density through hyperconjugation (quasimesomeric)38 to the attached aryl ring and exerts resonance effects, to a lesser extent, but similar
to those of methoxy.36 The

for 13 with an electron-“releasing” m-methyl (σ = -

0.07) fits on the parabolic Hammett correlation curve (Figure 4.8A). The steep slope in this
136

region suggested that the rate of the PaPAM reaction is strongly affected by the electronreleasing meta-substituent. Despite the smaller meta-substituent constant for methyl than for
methoxy, the mesomeric m-methoxy releases more electron density to the ring than the methyl
does through hyperconjugation. Thus, it was postulated that the rate enhancement of the addition
step, through a favorable transition state (Figure 4.7G), with 13 was not as significant as with 11.
This likely accounted for the >3.5-fold faster

for m-methoxy 11 than for m-methyl

substrate 13.
The product pool catalyzed by PaPAM from 13 contained more cinnamate analogue
(21.7%) compared to that made from other m-substituted substrates 2, 4, 9, and 11 that contained
between 1.0% and 7.4%. Therefore, the amination of the m-methyl aryl acrylate is likely more
sensitive to the effects of the substituent. m-Fluoro substrate 2 was converted to the cinnamate
analogue (14.8%) at a similar proportion as was 13. Compared with 1, it is intriguing that
substrates 2 and 13, with opposing electronic and steric properties, similarly affect the kcat of
PaPAM and the ratio of the cinnamate:β-amino acid analogues.
4.3.5.2.3. meta-Substituent Effects on Catalytic Efficiency
The plot between

and σ for the meta-substituted (mX) arylalanines (Figure

4.8B) showed that the substituent effects on the catalytic efficiency
the nonlinear relationship between

mostly paralleled

and σ (Figure 4.8A). That is, the catalytic

efficiency decreased paradoxically with substituents of higher electron-withdrawing or -donating
strength. Thus, the substituent effects on the kcat value of the catalytic efficiency were not
masked by the KM. Interestingly, the m-fluoro (2) substrate, which was an outlier on the
parabolic regression plot of

and σ, fits the linear regression plot between
137

and σ ( = –1.05). The effects of the electron-withdrawing m-fluoro substituent on
the catalytic efficiency correlated well with those of m-chloro and m-bromo (Figure 4.8B).
Reciprocally, the m-methoxy (11) substrate fit the parabolic regression plot between
and σ (Figure 4.8A), and was an outlier on the

correlation plot (Figure 4.8B).

This result suggested that the catalytic efficiency of PaPAM for substrates 2 and 11 was
influenced more by their affinity for PaPAM than by electronic substituent effects. The relatively
low KM (27 µM) for 2 likely revealed that the acrylate intermediate and β-amino acid products
were also released poorly and affected the turnover. Contrastingly, the high KM (990 µM) for 11
showed poor substrate binding, which masked the correlation between the electronic effects of
the m-methoxy and catalytic efficiency.
4.3.5.3. Substituent effects on Michaelis Parameters: para-Substituents

The calculated

of PaPAM and substituent constants (σ) for the para-

substituted (pX) arylalanines (p-fluoro (5), p-chloro (14), p-bromo (15), p-methyl (16), p-nitro
(17), and p-methoxy (18)) do not follow a single Hammett plot (Figure 4.9A). Electron-donating
substrates 16 and 18, and the natural substrate 1 follow a positive-slope (ρ = +4.74) linear
regression curve with a moderate correlation coefficient (R = 0.87). The linear regression curve
between electron-donating substituents 14 and 15, and 1 showed a negative-slope (ρ = -0.93)
with a low correlation coefficient (R = 0.71). The  value (+4.74), which is much greater than
unity for the electron donating substrates, suggests that catalysis is very dependent on the nature
of these substituents. By comparison, the  ≈ –1.0 for substrates 14, 15, and 17 suggests a
moderate yet significant dependence on the electron-withdrawing strength of the substituent
(Figure 4.9A). The intersecting linear regressions of opposite slope (ρ) suggest the substituent
138

effects transition from affecting the elimination step to affecting the amination step. 37 Unlike the
meta-substituted substrates with a higher catalytic rate for m-chloro and m-bromo analogues, all
para-substituents slowed the isomerization reaction compared to 1.
A

B
-2.5
H

H

0.0

-3.0

-0.2

log(kcat/KM)

log(kx/KH)

-0.4
-0.6
Cl

-0.8

Br

-1.0

NO2

F

-3.5
Cl

-4.0

CH3

Br

OCH3

-1.2

F

-1.4

CH3

-1.6
-0.4

-0.2

NO2

-4.5
OCH3

0.0

0.2

0.4

0.6

0.8

-5.0
-0.4

1.0

-0.2

0.0

0.2

0.4

0.6

0.8

1.0

(para-Substituents)

(para-Substituents)

Figure 4.9. Dependence of the observed
[designated as log(kx/kH)] (A) and
[designated as log(kcat/KM)] (B) on the Hammett substituent constant for the PaPAM-catalyzed
isomerization of para-substituted α-arylalanines. Here,
is
for entries 14, 15, 16, 17, and 18,
is
for entry 1; kcat is
for entries 5, 14, 15, 16, 17, and 18. Open circle: outlier p-fluoro
substrate 5.

4.3.5.3.1. Electron-withdrawing Substituents
The resonance hybrid of the p-nitro substrate 17 has a δ+ directly on Cipso attached to Cβ
(Figure 4.7D). While this was imagined to strongly increase the elimination rate (i.e., facilitates
Hβ removal), the 8.5-fold slower

of PaPAM for 17 (0.031 s-1) than for 1 (Table 4.1) likely

resulted because the p-nitro slowed the hydroamination rate (i.e., deterred nucleophilic attack at
Cβ) (Figure 4.7J) more than it improved the elimination rate. The higher ratio of pnitrocinnamate (52%) compared to cinnamate (7.2%) made from 1 further supports an affected
hydroamination step.
139

The inductive electron-withdrawing effect of p-chloro and p-bromo of substrates 14 and
15 on Cipso are lower than thatfor the corresponding meta-isomers. The lone-pair electrons of the
former however can delocalize by resonance and place a δ– directly on Cipso attached to Cβ in the
resonance hybrid. The δ– will promote the amination step (Figure 4.7H), yet dramatically retard
the deprotonation of the elimination step of the PaPAM reaction (Figure 4.7F). Almost similar
ratio of p-chloro- (5.9%) and -bromocinnamate (4.4%) compared to cinnamate (7.2%) made
from the natural substrate suggested that the hydroamination step is not affected from chloro and
bromo substituents. Thus, the lower

of 14 and 15, would likely have caused by the

unfavorable effects on the elimination step.
The p-fluoro substrate 5 was turned over by PaPAM at about the same rate as the mfluoro substrate 3 and o-fluoro substrate 10, but coincidentally at the same rate as the other parasubstituted substrates. It seems that regardless of regiochemistry, the overarching electronic
effect(s) of the fluoro substituent stalls the elimination and hydroamination steps. In addition,
based on the β-amino acid:aryl acrylate (85.7:14.3) distribution catalyzed by PaPAM from 5, it
seems that the fluoro affects the efficiency of the β-amination step, compared to the reaction
involving 1. A similar product distribution was also seen for the m-fluoro substrate 3.
4.3.5.3.2. Electron-donating Substituents
The electron-releasing p-methyl of 16 and p-methoxy of 18 place a δ– on Cipso of the
substrate via hyperconjugation and resonance, respectively. Each theoretically causes the pKa of
Hβ to increase and discourages the deprotonation of the presumed rate-limiting elimination step
(Figure 4.7F). The Hammett substituent constants predicted the electron-releasing p-methyl
= 0.013 s−1) more than the methoxy (

substituent would affect PaPAM turnover (
140

=

0.022 s−1), as observed (Table 4.1). The electronic effects of 16 and 18 facilitate the
hydroamination of the corresponding cinnamate intermediates and the product pool from 18
contained a higher ratio of p-methoxy-β-amino acid (97.7%) as expected. Surprisingly, p-Methyl
substrate had a higher proportion of p-methylcinnamate (36.4%) compared to 7.2% cinnamate
for 1.
4.3.5.3.3. para-Substituent Effects on Catalytic Efficiency

The relationship between

and σ for the para-substituted (pX) arylalanines

(Figure 4.9B) showed a similar trend in substituent effects on the catalytic efficiency
as seen between

for PaPAM and Hammett substituent constants (Figure 4.9A). As

with the meta-substituents, the catalytic efficiency of PaPAM was also sensitive to the parasubstituents. Electron-releasing substituents (16 and 18) fit into a linear regression curve with 1
(correlation coefficient (R) = 1.0; ρ = 7.50) while electron-withdrawing substituents (5, 14, 15,
and 17) follow a decay curve (R = 0.99). In contrast to the poor correlation between
and σ, para-substituents showed a superior correlation in the Hammett plot of catalytic
efficiencies. This informed that a reduction in catalytic efficiency was principally dictated by
large KM and not the electronic effects of the para-substituent that separately affected kcat.

141

4.3.6. Structure Activity Relationships of PaPAM Substrate Analogues
4.3.6.1. Structural Characteristics of PaPAM Active Site
The overall structure of PaPAM is composed of multiple parallel α-helices that form the
catalytically functional tetramer arranged as a dimer of head-to-tail dimers.18 Each monomer of
the tetramer contains an active site which is located at the interfaces between three subunits
(Figure 4.10A). Two of the PaPAM active sites revealed β-phenylalanine covalently attached by
the amino group to the methylidene carbon atom of MIO (Figure 4.10A), while the other active
sites showed an MIO covalent adduct with partial occupancy of both β- and α-phenylalanines.18
The phenylpropanoid carbon backbone of the substrate resides above the MIO facilitating the αto β-amino group transfer during the isomerization reaction. The mono-dentate salt bridge
interaction between the substrate carboxylate group and highly conserved Arg323 anchors the
substrate in the PaPAM active site.18 In addition, the aromatic binding pocket of PaPAM
provides an optimum environment to orient the aromatic ring of arylalanines. The aromatic
binding pocket is mostly lined by the side chains of the hydrophobic residues (Leu104, Val108,
Leu171, Leu216, Leu421, Met424 and Phe428), which likely contribute to the binding of an aryl
ring via van der Waals interactions.
Although the structure of PaPAM bound to the natural substrate/product provides
information on key interactions of substrate binding, favorable binding orientations of ringsubstituted substrate analogues are not predictable. Additionally, ortho- and meta-substituted
analogues, can orient their substituent groups in two different conformations, where the
substituent is cis or trans to the amino group of the substrate (Figure 4.10B). In "NH2-cis"
orientation, the ring substituent is on the same side as the amino group of α-phenylalanine. In the
142

"NH2-trans" conformation obtained by a 180 rotation of the Cβ-Cipso bond, the ring substituent is
oriented on the opposite side of the amino group. These two substrate binding orientations
interact differently with PaPAM due to the differences in the active site residues in either side of
the aromatic ring. Therefore, molecular docking of substrate analogues in PaPAM active site
was performed to analyze the preferred binding orientations and interactions of substrate
analogues.

A

B
Tyr78

Phe455
Leu421
Arg323
Met424

NH2-trans

NH2-cis

Leu104
Gln456

Leu216

Tyr320

Val108

NH2

Phe428

Leu171

MIO

Figure 4.10. Structural characteristics of the PaPAM active site, and NH2-cis and NH2-trans
configurations of substrate analogues. A) Active site of PaPAM showing side-chains which are in 4 Å
distance from the MIO-bound ligand. (ligand atoms are color-coded as: carbon: teal; nitrogen: blue;
oxygen: red). The active site residues are color-coded (carbon atoms in green, yellow, and wheat) to
differentiate the involvement from each sub unit. B) An overlay of the NH2-cis and NH2-trans
configurations is illustrated using meta-methyl-(S)-α-phenylalanine substrate. The methyl group can be
positioned on the same side (NH2-cis) or the opposite side (NH2-trans) as the reactive amino group of the
chiral substrate.

4.3.6.2. Correlation between Substrate-PaPAM Interaction Energies and KM
To gain insight into the differences of KM values of substrate analogues and their
preferred binding orientations, all the substrates (1 – 22) were docked in PaPAM active site
143

using SLIDE (Screening for Ligands by Induced-fit Docking, Efficiently) docking tool (version
3.4)39,40 and Szybki41-43 1.7.0 (OpenEye Scientific Software). Bio-active conformation of αphenylalanine found in the structure of PaPAM was used to generate the substrate conformations
employed for molecular docking. The aryl rings of the substrate analogues 2 – 22 were overlaid
on to the active conformation of α-phenylalanine structure by using molecular editing in PyMOL
1.5.0.4 (Schrӧdinger, Inc., New York, NY) and reference coordinates were fixed in OMEGA
2.4.6 (OpenEye Scientific Software, Santa Fe, NM; http://www.eyesopen.com).44,45
SLIDE and other docking tools handle cofactors as rigid parts of the protein and
therefore, cofactors can have strong constraints on ligand docking.39,40 In addition, covalently
bound ligands are generally interpreted as unfavorable van der Waals collisions during docking
processes. Methylidene carbon of the MIO cofactor, which lies in bonding distance from the
amino group of the substrate caused unfavorable steric collisions and forced the substrate to dock
in a catalytically incompetent orientation. Thus, the alkene carbon atoms of the MIO were
removed to facilitate the ligand docking in PaPAM active site.
For SLIDE docking of substrates 2 – 22, a ligand-based template was created using
known hydrophobic and hydrogen-bonding interactions of α-phenylalanine (1) with the active
site residues. This template was used as a basis for docking other substrates in PaPAM active
site. Interactions of the substrate carboxylate- and amino-groups were used as key points to
anchor the substrates in both SLIDE and Szybki docking.
The sum of protein-ligand interaction energy [E(p-l)] and ligand internal energy [E(l)]
values for docked ligand candidates were calculated using Szybki.41-43 The E(p-l) term was
consisted of electrostatic Coulombic [EC(p-l)] and steric van der Waals interaction energy [EV(p-l)]
144

terms for each docked conformer. Two approaches were employed to calculate the
aforementioned energy terms of the ligand-PaPAM complexes of 1 – 22. In the first model
(static model), the PaPAM binding site was kept in its crystallographic orientation and the
energy terms were calculated without any energy minimization. The second model, which is a
flexible model, however, allowed the rotation of protein side chains within 4 Å and ligand aryl
ring to relieve unfavorable van der Waals interactions.
The correlation coefficient between each energy term and experimental KM was
calculated to understand the energy terms that dictate the substrate binding in PaPAM. For the
static ligand modeling, the linear correlation coefficient between [E(p-l) + E(l)] and KM was 0.48.
EV(p-l) and EC(p-l) which are components of E(p-l), had correlation coefficients of 0.54 and 0.33,
respectively. The lower correlation between the Coulombic energy and KM compared to that of
van der Waals interaction energy, reflects that the steric effects in the protein-ligand complex
and within the ligand are dominant over electrostatic interactions upon substrate binding.
Interestingly, energy minimization lowered the correlation coefficient between [E(p-l) + E(l)] and
KM to 0.35. This suggests that the crystallographic orientation of PaPAM active site and the
substrates provide the most favorable conformation for the analogues used in this study. The
energy minimization protocol likely cause side chain rotations that may reflect catalytically
unproductive conformations. For example, nitro substituent was rotated out-of-plane relative to
the phenyl ring during energy minimization. However, the analysis of 200 nitrophenyl groups in
small-molecule

crystal

structures

in

the

Cambridge

Structural

Database

1.1.1

(http://webcsd.ccdc.cam.ac.uk) indicated that 87.5% of the nitrophenyl groups are entirely coplanar, regardless of other features in the structure.46

145

4.3.6.2.1. Preferred Conformations of Ring-substituted Substrates
To identify the preferred binding orientations of ortho- and meta-substituents, the sum of
protein-ligand interaction energy [E(p-l)] and ligand internal energy [E(l)] values were calculated
for docked NH2-cis and NH2-trans conformers. Van der Waals interaction energy [EV(p-l)] of
preferred orientations were also extracted for each docked substrate. If the difference in total
energy [E(p-l) + E(l)] between two conformations was ≤25 kcal/mol, the substrates were
considered to have no preference for NH2-cis or NH2-trans orientation.
Several active site residues caused unfavorable van der Waals overlaps with the ring
substituted substrates binding to PaPAM active site (Figure 4.11A). Compared to NH2-cis
conformers, NH2-trans conformers of ortho-substituted substrates have less steric barriers
created from Leu104 and Leu216. NH2-cis conformers were sterically hindered from Tyr320,
Phe428 and Gln456 (Figure 4.11B). Consequently, o-methyl (6), o-bromo (20), o-chloro (21),
and o-nitro (22) energetically favored the NH2-trans orientation (Table A.2.1). However, ofluoro (10) substituent with a smaller van der Waals radius had no preference for binding in
either orientation. Interestingly, bulkier o-methoxy (19) substituent showed a lower [E(p-l) + E(l)]
energy for the NH2-cis conformer likely due to the favorable hydrogen-bonding interactions of
the methoxy group with active site Tyr320 (Figure A.2.39).
For meta-substituted substrates, NH2-cis conformers encounter lower steric barriers from
Gly85, Tyr320, Phe428 and Gln456, while NH2-trans conformers have higher steric overlaps
from Leu104, Val108, and Leu421 (Figure 4.11C). Thus, m-bromo (2), m-chloro (4), m-nitro (9),
and m-methyl (13) preferred to bind in NH2-cis orientation (Table A.2.1). Similar to ortho-

146

substituted analogue, m-fluoro (3) showed no significant preference for NH2-cis or NH2-trans
conformation.
Except p-methoxy substituent (18), para-substrates with a symmetrical aromatic ring
showed no significant preference for NH2-cis or NH2-trans orientation. However, parasubstituents encounter van der Waals overlaps with active site residues Val108, Leu421 and
Phe428 (Figure 4.11D). Bulkier tBu group was not productive at para-position likely due to the
unfavorable steric constrains for binding. However, bulky, methoxy substituent of 18 avoids
steric repulsion due to its rotational flexibility and preferred to orient towards the NH 2 group in
the active site.
The binding affinity order for all substrates approximately corresponded to the van der
Waals (vdW) radii of the substituents. PaPAM bound substrates with a fluoro group (~1.5 Å) the
best, followed by methyl (~1.9 Å), then bromo and chloro groups (~1.8 Å). All three fluoro
substituted substrates (3, 5, and 10) bound to PaPAM better than the natural substrate. The least
favorable substrate binding to PaPAM contained the bulkiest substituents: nitro (~3.1 Å; from
the vdW radii of the Car–N bond length and the terminal O–N=O) and methoxy (~3.4 Å; from
the vdW radii of the Car–O bond and the methyl C–H bonds of the methoxy).47,48 p-Methoxy
substrate 18 showed the least favorable binding with the highest KM of all the substrates.
Surprisingly, o-methoxy (11) (KM = 164 µM) group bound PaPAM better than expected from its
calculated EV(p-l) (108 kcal/mol).

147

A

Leu421

NH2cis

NH2trans

Phe428
Val108

NH2

Gln456

Leu104

Tyr320

Leu216

Partial MIO

B

C

D
Leu421

Phe428

Leu421
Val108

Val108

Gln456

Phe428

Leu104
Tyr320

Figure 4.11. PaPAM active site residues that cause van der Waals overlap with the ring-substituted
substrates (A) and, the active site residues that sterically hinder the ortho- (B), meta- (C), and para(C) substituted substrates. Partial MIO group is shown. Atoms are color-coded as: carbon: teal (active
site residues)/green (ligand); nitrogen: blue; oxygen: red.

148

4.3.7. Kinetic Parameters and Interaction Energies of Heteroaromatic Substrates
Of the three heteroaromatic substrates, 2-furylalanine (7), showed the highest
(0.236 s−1) and the catalytic efficiency. 3-Thienylalanine (8) turned over with a 1.7-fold slower
rate compared to 7. The 1.2-fold reduction in KM, however increased the catalytic efficiency of
3-thienylalanine analogue 8. In contrast to 7 and 8, KM of 2-thienylalanine (12, 132 µM) was
almost similar to the natural substrate 1. However, 12 was among the substrates with a lower
(0.026 s−1) for the isomerization reaction.

values of PaPAM for all heteroaromatic

substrates (7, 8, and 12) are about 1.4- to 12-fold lower than that for 1.
Evaluation of resonance structures of 3-thienylalanine (8) showed that a δ– charge resides
on Cipso of the thienyl ring (Figure 4.12A, resonance path 'a') analogous to δ– charge on Cipso of
substrates 5, 14 – 18 containing an electron-donating para-substituent on the phenyl ring (see
Figure 4.7F or H). For substrate 8, however, two vicinal δ– charges induce a "δ+" on the Cipso,
which likely reduce the magnitude of the δ– at Cipso (Figure 4.12A, resonance path 'b'). Thus, the
lower magnitude δ– at Cipso of 8, compared to the δ– at Cipso for , 5, 14 – 18 (0.013 – 0.053 s-1)
likely affected the rate-determining deprotonation step lesser, as evidenced by its 3- to 10-fold
higher

of PaPAM for 8 (0.143 s-1).
Similar to the resonance hybrid of 8, 2-thienylalanine substrate has a resonance structure

with a δ– charge on Cipso (Figure 4.12B, resonance path 'a'). However, 12 has only one δ– charge
vicinal to Cipso (Figure 4.12B, resonance path 'b'), and therefore the induced charge on C ipso
(illustrated as "ƒδ+") is less than the induced "δ+" in 8 flanked by two vicinal δ– charges (Figure
4.12B, resonance path 'b'). Consequently, Cipso of 12 is more δ– charged compared to 8. The
greater δ– charge on Cipso of 12 than on 8 likely conflicts with the δ– formed on Cβ during the
149

transition state of the deprotonation step. Thus, this effect likely caused the ~6-fold slower
PaPAM reaction for 12 compared to 8 (0.026 s-1 for 12 and 0.143 s-1 for 8).
It was interesting that the 2-furylalanine (7, 0.236 s−1) was turned over by PaPAM ~9
fold faster than the analogous 2-thienylalanine (12, 0.026 s-1), particularly, since these two
heteroaromatic substrates have similar resonance hybrids (Figure 4.12B). However, more
electronegative oxygen compared to sulfur of 12 likely induced a larger δ+ charge on the vicinal
Cipso of 7. Moreover, more electronegative oxygen of 7 distributes its lone pair electrons less than
sulfur, and thus likely reduced the magnitude of the negative charge (δ –) delocalized on Cipso
(Figure 4.12B, resonance path 'a'). As a result, the relative magnitude of δ+ on Cipso of 7 is larger
than that of 12. Increased magnitude of δ+ at Cipso promotes the removal of Hβ in the PaPAM
reaction. The higher proportion of (2-furyl)acrylate (65.2%) from 7 compared to only 9.1% (2thienyl)acrylate from 12, suggests that the amination step during the isomerization of 7 is
negatively affected by the comparatively larger δ+ on Cipso.

Figure 4.12. Resonance hybrids of 3-thienylalanine (8) (A) and, composite resonance hybrids of 2furylalanine (7) and 2-thienylalanine (12) (B). The induced charges on Cipso are designated in quotation
marks ("").
150

Heteroaromatic substrates 7, 8, and 12 were also analyzed for their preferred binding
orientations in PaPAM active site. Analogous to NH2-cis and -trans conformations of aryl
substituted substrates, heteroatom of the aryl rings were also oriented in the same- or oppositeside of the substrate amino group. According to energy calculations, both conformations of 7, 8,
and 12 were equally preferred in PaPAM active site. Calculated van der Waals energy [EV(p-l)] (7
and 8: 21 kcal/mol; and 12: 20 kcal/mol) predicted that all three substrates bind equally well as
the natural substrate 1 (19 kcal/mol, 168 µM). However, only 2-thienylalanine (12) showed a
higher affinity with a lower KM value (132 µM). Surprisingly, both 2-furylalanine (7) and 3thienylalanine (8) bound less tightly to PaPAM (KM = 415 and 337 µM respectively).

151

4.4. Conclusion
PaPAM-catalyzed α/β-isomerization reaction has broad substrate scope for 19 arylalanine
analogues. α-Arylalanine analogues including fluoro-, chloro-, bromo-, methyl-, methoxy-, and
nitro-substituents on the phenyl ring as well as furyl- and thienyl-alanines with heteroaromatic
rings were isomerized to their respective β-arylalanine products by PaPAM catalysis. The
influence of the substituents on the kcat of PaPAM revealed a concave-down correlations with
Hammett substituent constants (σ). The trend of these correlations37 suggests that the ratedetermining step changes from the elimination to the hydroamination step based on the direction
and magnitude of the electronic properties of the substituent.
The computational analyses provided a means to predict the docking conformation of
substituted 22 arylalanine substrates. The information from computational modeling was useful
for designing future targeted amino acid mutagenesis of PaPAM to increase the catalytic
efficiency by improving the binding affinity for various other non-natural substrates. The KM of
PaPAM was higher for several substrates with meta- and para-substituents (except fluoro and
methyl) than for 1. The presumed lower binding affinity was likely due to steric interactions
between the substituents and the active site residues of PaPAM. meta-Substituted substrates
were shown to prefer the NH2-cis configuration to avoid steric clashes with branched
hydrophobic residues. Mutation of Leu104, Val108, and Leu421 to alanines may improve the
binding of meta-substituted substrates by providing flexibility to bind in the NH2-cis or NH2trans configuration. Furthermore, computational docking predicted that para-substituents
sterically clash with Phe428, Val108, and Leu421. Therefore, exchange of these residues for
alanine may facilitate the binding of para-substituted substrates. Surprisingly, the computational

152

analysis predicted that all ortho-substituted α-arylalanines bound well to PaPAM; however,
relief of the active site sterics may enable these ortho-substituted α-arylalanines to better access a
catalytically competent conformation and improve the turnover number for these substrates.

153

APPENDIX

154

178

100

Biosynthetic Product
GC Rt = 12.11 min

80

Relative Ion Abundance (%)

60
40

104

20

121

191

162
134 146

251

0
178

100

Authentic Standard
GC Rt = 12.15 min

80
60
40

104

20

121

191
134 146

162

251

0
100

150

200

250

m/z

Figure A.2.1. EI-MS spectra of the N-(ethoxycarbonyl) methyl ester
derivatives of biosynthetic β-phenylalanine made from PaPAM catalysis (top)
and authentic β-phenylalanine (bottom). GC retention times (GC Rt) are shown.

155

256

100

Biosynthetic Product
GC Rt = 14.69 min

80

Relative Ion Abundance (%)

60

184

40
20

270
77

103
117

329

212

157

242

0
256

100

Authentic Standard
GC Rt = 14.68 min

80
60

184

40
20

77

270
103
117

157

212

329

242

0
100

150

200

250

300

350

m/z

Figure A.2.2. EI-MS spectra of the N-(ethoxycarbonyl) methyl ester
derivatives of biosynthetic m-bromo-β-phenylalanine made from PaPAM
catalysis (top) and authentic m-bromo-β-phenylalanine (bottom). GC retention
times (GC Rt) are shown.

156

196

100

Biosynthetic Product
GC Rt = 11.97 min

80

Relative Ion Abundance (%)

60
124

40
209

20

139 152

109

164 180

269

230

0
196

100

Authentic Standard
GC Rt = 11.94 min

80
60
124

40

209

20

139
109

152 164 180

269

0
100

150

200

250

m/z

Figure A.2.3. EI-MS spectra of the N-(ethoxycarbonyl) methyl ester
derivatives of biosynthetic m-fluoro-β-phenylalanine made from PaPAM
catalysis (top) and authentic m-fluoro-β-phenylalanine (bottom). GC retention
times (GC Rt) are shown.

157

212

100

Biosynthetic Product
GC Rt = 13.80 min

80

Relative Ion Abundance (%)

60
140

40

225

20
77

102 113

155 168

196

285

180

0
212

100

Authentic Standard
GC Rt = 13.98 min

80
60
40

140
225

20
77

102 113

155

168

196

285

180

0
100

150

200

250

300

m/z

Figure A.2.4. EI-MS spectra of the N-(ethoxycarbonyl) methyl ester
derivatives of biosynthetic m-chloro-β-phenylalanine made from PaPAM
catalysis (top) and authentic m-chloro-β-phenylalanine (bottom). GC retention
times (GC Rt) are shown.

158

196

100

Biosynthetic Product
GC Rt = 9.80 min

80

Relative Ion Abundance (%)

60
122

40

180

20

109

139

209
152 164

269
238

0
196

100

Authentic Standard
GC Rt = 9.79 min

80
60
40

122

20

139
109

209

152 164 180

100

150

269

238

0
200

250

m/z

Figure A.2.5. EI-MS spectra of the N-(ethoxycarbonyl) methyl ester
derivatives of biosynthetic p-fluoro-β-phenylalanine made from PaPAM
catalysis (top) and authentic p-fluoro-β-phenylalanine (bottom). GC retention
times (GC Rt) are shown.

159

192

100

Biosynthetic Product
GC Rt = 10.60 min

80

Relative Ion Abundance (%)

60
40

118

20

135 148

176

205

160

265
234

0
192

100

Authentic Standard
GC Rt = 10.61 min

80
60
40

118

20
135 148

205

176

265

160

234

0
100

150

200

250

m/z

Figure A.2.6. EI-MS spectra of the N-(ethoxycarbonyl) methyl ester
derivatives of biosynthetic o-methyl-β-phenylalanine made from PaPAM
catalysis (top) and authentic o-methyl-β-phenylalanine (bottom). GC retention
times (GC Rt) are shown.

160

168

100

Biosynthetic Product
GC Rt = 10.10 min

80

Relative Ion Abundance (%)

60
40

181
96

20
111 124

136 152

241

0
168

100

Authentic Standard
GC Rt = 10.11 min

80
60
40

96

181

20
111

136

152
241

124

0
50

100

150

200

250

m/z

Figure A.2.7. EI-MS spectra of the N-(ethoxycarbonyl) methyl ester
derivatives of biosynthetic 2-furyl-β-alanine made from PaPAM catalysis (top)
and authentic 2-furyl-β-alanine (bottom). GC retention times (GC Rt) are shown.

161

184

100

Biosynthetic Product
GC Rt = 12.31 min

80

Relative Ion Abundance (%)

60
40
20

110

257
149

85
127

97

197

168
138

226

0
184

100

Authentic Standard
GC Rt = 12.30 min

80
60
40
20

110
257

85
97

127

138

149 168

197
226

0
100

150

200

250

m/z

Figure A.2.8. EI-MS spectra of the N-(ethoxycarbonyl) methyl ester
derivatives of biosynthetic 3-thienyl-β-alanine made from PaPAM catalysis
(top) and authentic 3-thienyl-β-alanine (bottom). GC retention times (GC Rt) are
shown.

162

223

100

Biosynthetic Product
GC Rt = 15.95 min

80
151

Relative Ion Abundance (%)

60
40
20

279
105
77
91

166
177

133

117

207
191

236

247
267

296

0
223

100

Authentic Standard
GC Rt = 15.96 min

80
151

60
40
20

279
105
77
91

166
207
177191

133
117

236

247
267

296

0
100

150

200

250

300

m/z

Figure A.2.9. EI-MS spectra of the N-(ethoxycarbonyl) methyl ester
derivatives of biosynthetic m-nitro-β-phenylalanine made from PaPAM
catalysis (top) and authentic m-nitro-β-phenylalanine (bottom). GC retention
times (GC Rt) are shown.

163

196

100

Biosynthetic Product
GC Rt = 11.82 min

80

Relative Ion Abundance (%)

60

124

40
152
139

109

20

97

180
209
269

0
196

100

Authentic Standard
GC Rt = 11.80 min

80
60
124

40
20
97 109

139 152

209

180

269

0
50

100

150

200

250

300

m/z

Figure A.2.10. EI-MS spectra of the N-(ethoxycarbonyl) methyl ester
derivatives of biosynthetic o-fluoro-β-phenylalanine made from PaPAM
catalysis (top) and authentic o-fluoro-β-phenylalanine (bottom). GC retention
times (GC Rt) are shown.

164

208

100

Biosynthetic Product
GC Rt = 14.28 min

80

Relative Ion Abundance (%)

60
40
134

20
77 109

281

221

121

151 164

176 192

250

0
208

100

Authentic Standard
GC Rt = 14.29 min

80
60
40
134

20

281

221
77 109 121

151

164 176

192
250

0
100

150

200

250

300

m/z

Figure A.2.11. EI-MS spectra of the N-(ethoxycarbonyl) methyl ester
derivatives of biosynthetic m-methoxy-β-phenylalanine made from PaPAM
catalysis (top) and authentic m-methoxy-β-phenylalanine (bottom). GC
retention times (GC Rt) are shown.

165

184

100

Biosynthetic Product
GC Rt = 12.11 min

80

Relative Ion Abundance (%)

60
40
20

110
197

85

168
127

97

257

152
138

226

0
184

100

Authentic Standard
GC Rt = 12.11 min

80
60
40
20

110
197

85

168
97

127

257

152
138

226

0
100

150

200

250

m/z

Figure A.2.12. EI-MS spectra of the N-(ethoxycarbonyl) methyl ester
derivatives of biosynthetic 2-thienyl-β-alanine made from PaPAM catalysis
(top) and authentic 2-thienyl-β-alanine (bottom). GC retention times (GC Rt) are
shown.

166

192

100

Biosynthetic Product
GC Rt = 10.60 min

80

Relative Ion Abundance (%)

60
40

118

20

135 148

176

205

160

265
234

0
192

100

Authentic Standard
GC Rt = 10.61 min

80
60
40

118

20
135 148

205

176

265

160

234

0
100

150

200

250

m/z

Figure A.2.13. EI-MS spectra of the N-(ethoxycarbonyl) methyl ester
derivatives of biosynthetic m-methyl-β-phenylalanine made from PaPAM
catalysis (top) and authentic m-methyl-β-phenylalanine (bottom). GC retention
times (GC Rt) are shown.

167

212

100

Biosynthetic Product
GC Rt = 13.99 min

80

Relative Ion Abundance (%)

60
40

140
225

20
77

102 113

155

168

196

285

180

0
212

100

Authentic Standard
GC Rt = 13.98 min

80
60
40

140
225

20
77

102 113

155

168

196

285

180

0
100

150

200

250

300

m/z

Figure A.2.14. EI-MS spectra of the N-(ethoxycarbonyl) methyl ester
derivatives of biosynthetic p-chloro-β-phenylalanine made from PaPAM
catalysis (top) and authentic p-chloro-β-phenylalanine (bottom). GC retention
times (GC Rt) are shown.

168

256

100

Biosynthetic Product
GC Rt = 14.92 min

80

Relative Ion Abundance (%)

60

184

40
20

77 103
117

271
157

199 212

242

329

0
256

100

Authentic Standard
GC Rt = 14.92 min

80
60
184

40
20

271

77 103
117

157

329

199 212 242

0
100

150

200

250

300

350

m/z

Figure A.2.15. EI-MS spectra of the N-(ethoxycarbonyl) methyl ester
derivatives of biosynthetic p-bromo-β-phenylalanine made from PaPAM
catalysis (top) and authentic p-bromo-β-phenylalanine (bottom). GC retention
times (GC Rt) are shown.

169

192

100

Biosynthetic Product
GC Rt = 10.83 min

80

Relative Ion Abundance (%)

60
176

40

118
148

20

135

105

206

160

265

0
192

100

Authentic Standard
GC Rt = 10.81 min

80
60
40

118

20
105

135 148

206

176

265

160

0
100

150

200

250

m/z

Figure A.2.16. EI-MS spectra of the N-(ethoxycarbonyl) methyl ester
derivatives of biosynthetic p-methyl-β-phenylalanine made from PaPAM
catalysis (top) and authentic p-methyl-β-phenylalanine (bottom). GC retention
times (GC Rt) are shown.

170

223

100

Biosynthetic Product
GC Rt = 16.33 min

80

Relative Ion Abundance (%)

60
40
236

151

20

105

133

166 179

119

207
191

265

296

0
223

100

Authentic Standard
GC Rt = 16.33 min

80
60
40

236

151

20

105

133

166 179

119

207
191

265

296

0
100

150

200

250

300

m/z

Figure A.2.17. EI-MS spectra of the N-(ethoxycarbonyl) methyl ester
derivatives of biosynthetic p-nitro-β-phenylalanine made from PaPAM
catalysis (top) and authentic p-nitro-β-phenylalanine (bottom). GC retention
times (GC Rt) are shown.

171

208

100

Biosynthetic Product
GC Rt = 12.25 min

80

Relative Ion Abundance (%)

60
40
134

20
109

121

151
162

180

192

281

221

0
208

100

Authentic Standard
GC Rt = 12.29 min

80
60
40
134

20
109
121

151 162

180 192

281

221

0
100

150

200

250

300

m/z

Figure A.2.18. EI-MS spectra of the N-(ethoxycarbonyl) methyl ester
derivatives of biosynthetic p-methoxy-β-phenylalanine made from PaPAM
catalysis (top) and authentic p-methoxy-β-phenylalanine (bottom). GC retention
times (GC Rt) are shown.

172

208

100

Biosynthetic Product
GC Rt = 13.84 min

80

Relative Ion Abundance (%)

60
121

40

192

91
136

20

161

107

176
281

222

0
208

100

Authentic Standard
GC Rt = 13.86 min

80
60
40
136

20
91 107 121

161

176

192

281

222

0
100

150

200

250

300

m/z

Figure A.2.19. EI-MS spectra of the N-(ethoxycarbonyl) methyl ester
derivatives of biosynthetic o-methoxy-β-phenylalanine made from PaPAM
catalysis (top) and authentic o-methoxy-β-phenylalanine (bottom). GC retention
times (GC Rt) are shown.

173

-1

[S]/(M/nmol·s ) × 10-3

25

20

15

10

5

0
0

200

400

600

800

1000

[-Phenylalanine] (M)
Figure A.2.20. Hanes-Woolf plot of biosynthetic β-phenylalanine (designated as velocity, v)
catalyzed by PaPAM from α-phenylalanine (S).

15

-1

[S]/(M/nmol·s ) × 10-3

20

10

5

0
0

200

400

600

800

1000

[meta-Bromo--Phenylalanine] (M)
Figure A.2.21. Hanes-Woolf plot of biosynthetic m-bromo-β-phenylalanine (designated as
velocity, v) catalyzed by PaPAM from m-bromo-α-phenylalanine (S).

174

-1

[S]/(M/nmol·s ) × 10-3

150
125
100
75
50
25
0
0

200

400

600

800

[meta-Fluoro--Phenylalanine] (M)

Figure A.2.22. Hanes-Woolf plots of biosynthetic m-fluoro-β-phenylalanine (designated as
velocity, v) catalyzed by PaPAM from m-fluoro-α-phenylalanine (S).

15

-1

[S]/(M/nmol·s ) × 10-3

20

10

5

0
0

200

400

600

800

1000

[meta-Chloro--Phenylalanine] (M)

Figure A.2.23. Hanes-Woolf plots of biosynthetic m-chloro-β-phenylalanine (designated as
velocity, v) catalyzed by PaPAM from m-chloro-α-phenylalanine (S).

175

200

-1

[S]/(M/nmol·s ) × 10-3

250

150

100

50

0
0

200

400

600

800

1000

[para-Fluoro--Phenylalanine] (M)

Figure A.2.24. Hanes-Woolf plots of biosynthetic p-fluoro-β-phenylalanine (designated as
velocity, v) catalyzed by PaPAM from p-fluoro-α-phenylalanine (S).

75

-1

[S]/(M/nmol·s ) × 10-3

100

50

25

0
0

200

400

600

800

1000

[ortho-Methyl--Phenylalanine] (M)
Figure A.2.25. Hanes-Woolf plots of biosynthetic o-methyl-β-phenylalanine (designated as
velocity, v) catalyzed by PaPAM from o-methyl-α-phenylalanine (S).
176

-1

[S]/(M/nmol·s ) × 10-3

50

40

30

20

10

0
0

500

1000

1500

2000

[2-Furylalanine] (M)
Figure A.2.26. Hanes-Woolf plots of biosynthetic 2-furyl-β-alanine (designated as velocity, v)
catalyzed by PaPAM from 2-furyl-α-alanine (S).

-1

[S]/(M/nmol·s ) × 10-3

25

20

15

10

5

0
0

250

500

750

1000

[3-Thiophene-alanine] (M)
Figure A.2.27. Hanes-Woolf plots of biosynthetic 3-thiophenyl-β-alanine (designated as velocity, v)
catalyzed by PaPAM from 3-thiophenyl-α-alanine (S).
177

-1

[S]/(M/nmol·s ) × 10-3

25

20

15

10

5

0
0

500

1000

1500

2000

[meta-Nitro--Phenylalanine] (M)
Figure A.2.28. Hanes-Woolf plots of biosynthetic m-nitro-β-phenylalanine (designated as velocity, v)
catalyzed by PaPAM from m-nitro-α-phenylalanine (S).

-1

[S]/(M/nmol·s ) × 10-3

300

250

200

150

100

50

0
0

200

400

600

800

1000

[ortho-Fluoro--Phenylalanine] (M)
Figure A.2.29. Hanes-Woolf plots of biosynthetic o-fluoro-β-phenylalanine (designated as
velocity, v) catalyzed by PaPAM from o-fluoro-α-phenylalanine (S).
178

15

-1

[S]/(M/nmol·s ) × 10-3

20

10

5

0
0

500

1000

1500

2000

[meta-Methoxy--Phenylalanine] (M)
Figure A.2.30. Hanes-Woolf plots of biosynthetic m-methoxy-β-phenylalanine (designated as
velocity, v) catalyzed by PaPAM from m-methoxy-α-phenylalanine (S).

225

-1

[S]/(M/nmol·s ) × 10-3

300

150

75

0
0

250

500

750

1000

[2-Thiophene-alanine] (M)
Figure A.2.31. Hanes-Woolf plots of biosynthetic 2-thiophenyl-β-alanine (designated as velocity, v)
catalyzed by PaPAM from 2-thiophenyl-α-alanine (S).
179

100

-1

[S]/(M/nmol·s ) × 10-3

125

75

50

25

0
0

200

400

600

800

1000

[meta-Methyl--Phenylalanine] (M)
Figure A.2.32. Hanes-Woolf plots of biosynthetic m-methyl-β-phenylalanine (designated as
velocity, v) catalyzed by PaPAM from m-methyl-α-phenylalanine (S).

45

-1

[S]/(M/nmol·s ) × 10

-3

60

30

15

0
0

500

1000

1500

2000

[para-Chloro--Phenylalanine] (M)
Figure A.2.33. Hanes-Woolf plots of biosynthetic p-chloro-β-phenylalanine (designated as
velocity, v) catalyzed by PaPAM from p-chloro-α-phenylalanine (S).
180

-1

[S]/(M/nmol·s ) × 10-3

75

60

45

30

15

0
0

500

1000

1500

2000

[para-Bromo--Phenylalanine] (M)
Figure A.2.34. Hanes-Woolf plots of biosynthetic p-bromo-β-phenylalanine (designated as
velocity, v) catalyzed by PaPAM from p-bromo-α-phenylalanine (S).

400

-1

[S]/(M/nmol·s ) × 10-3

500

300

200

100

0
0

200

400

600

800

1000

[para-Methyl--Phenylalanine] (M)
Figure A.2.35. Hanes-Woolf plots of biosynthetic p-methyl-β-phenylalanine (designated as
velocity, v) catalyzed by PaPAM from p-methyl-α-phenylalanine (S).
181

75

-1

[S]/(M/nmol·s ) × 10-3

100

50

25

0
0

500

1000

1500

2000

[para-Nitro--Phenylalanine] (M)
Figure A.2.36. Hanes-Woolf plots of biosynthetic p-nitro-β-phenylalanine (designated as velocity, v)
catalyzed by PaPAM from p-nitro-α-phenylalanine (S).

150

-1

[S]/(M/nmol·s ) × 10-3

200

100

50

0
0

500

1000

1500

2000

[para-Methoxy--Phenylalanine] (M)
Figure A.2.37. Hanes-Woolf plots of biosynthetic p-methoxy-β-phenylalanine (designated as
velocity, v) catalyzed by PaPAM from p-methoxy-α-phenylalanine (S).
182

-1

[S]/(M/nmol·s ) × 10-3

450
375
300
225
150
75
0
0

500

1000

1500

2000

2500

[ortho-Methoxy--Phenylalanine] (M)
Figure A.2.38. Hanes-Woolf plots of biosynthetic o-methoxy-β-phenylalanine (designated as
velocity, v) catalyzed by PaPAM from o-methoxy-α-phenylalanine (S).

Figure A.2.39. H-bonding interaction of ortho-methoxy-α-phenylalanine (19) and active site Tyr320.
ortho-Methoxy-α-phenylalanine atoms are colored as: carbon: green; nitrogen: blue; oxygen: red and
Tyr320 atoms are colored as: carbon: teal; nitrogen: blue; oxygen: red.

183

Table A.2.1. Calculated E(p-l) and E(l) values, and preference for NH2-cis versus -trans orientation.
(E(p-l) + E(l)) (kcal/mol)
NH2-trans

NH2-cis

EV(p-l)
(kcal/mol)

1

149

149

19

168

Symmetrical

2

429

188

55

339

NH2-cis

3

153

148

19

27

NSDa

4

273

166

33

432

NH2-cis

9

1640

236

48

430

NH2-cis

11

265

240

86

990

NSD

13

245

174

40

204

NH2-cis

10

149

165

21

73

NSD

6

190

489

55

88

NH2-trans

19

409

292

108

164

NH2-cis

20

338

525

204

-b

NH2-trans

21

226

401

93

-

NH2-trans

22

393

2065

205

-

NH2-trans

5

150

150

19

29

Symmetrical

14

170

170

37

491

Symmetrical

15

193

193

60

525

Symmetrical

16

179

179

46

163

Symmetrical

17

360

360

186

752

Symmetrical

18

219

947

81

1187

Pseudo NH2trans

R

a

KM
(µM)

Preferred
Orientation

NSD: no significant difference in energy for the NH2-cis or NH2-trans conformation. bNon-productive substrates
are indicated by hyphens. Note, all energies reported should be considered relative rather than absolute.

184

REFERENCES

185

REFERENCES
(1)

Grayson, J. I.; Roos, J.; Osswald, S. Org. Process Res. Dev. 2011, 15, 1201-1206.

(2)

Lelais, G.; Seebach, D. Biopolymers 2004, 76, 206-243.

(3)

Yan, S.; Larson, G.; Wu, J. Z.; Appleby, T.; Ding, Y.; Hamatake, R.; Hong, Z.; Yao, N.
Bioorg. Med. Chem. Lett. 2007, 17, 63-67.

(4)

Ruf, S.; Buning, C.; Schreuder, H.; Horstick, G.; Linz, W.; Olpp, T.; Pernerstorfer, J.;
Hiss, K.; Kroll, K.; Kannt, A.; Kohlmann, M.; Linz, D.; Hubschle, T.; Rutten, H.; Wirth,
K.; Schmidt, T.; Sadowski, T. J. Med. Chem. 2012, 55, 7636-7649.

(5)

Ismail, F. M. D. J. Fluorine Chem. 2002, 118, 27–33.

(6)

Wang, J.; Sanchez-Rosello, M.; Acena, J. L.; del Pozo, C.; Sorochinsky, A. E.; Fustero,
S.; Soloshonok, V. A.; Liu, H. Chem. Rev. 2014, 114, 2432-2506.

(7)

Zhu, Y.; Wu, G.; Zhu, X.; Ma, Y.; Zhao, X.; Li, Y.; Yuan, Y.; Yang, J.; Yu, S.; Shao, F.;
Lei, M. J. Med. Chem. 2010, 53, 8619-8626.

(8)

Huang, X.; O’Brien, E.; Thai, F.; Cooper, G. Org. Process Res. Dev. 2010, 14, 592-599.

(9)

Walker, K. D.; Klettke, K.; Akiyama, T.; Croteau, R. J. Biol. Chem. 2004, 279, 5394753954.

(10)

Klettke, K. L.; Sanyal, S.; Mutatu, W.; Walker, K. D. J. Am. Chem. Soc. 2007, 129, 69886989.

(11)

Magarvey, N. A.; Fortin, P. D.; Thomas, P. M.; Kelleher, N. L.; Walsh, C. T. ACS Chem.
Biol. 2008, 3, 542-554.

(12)

Christenson, S. D.; Liu, W.; Toney, M. D.; Shen, B. J. Am. Chem. Soc. 2003, 125, 60626063.

(13)

Krug, D.; Muller, R. Chembiochem 2009, 10, 741-750.

(14)

Christenson, S. D.; Wu, W.; Spies, M. A.; Shen, B.; Toney, M. D. Biochemistry 2003, 42,
12708-12718.

(15)

Huang, S. X.; Lohman, J. R.; Huang, T.; Shen, B. Proc. Natl. Acad. Sci. U. S. A. 2013,
110, 8069-8074.

(16)

Schwede, T. F.; Retey, J.; Schulz, G. E. Biochemistry 1999, 38, 5355-5361.

186

(17)

Christianson, C. V.; Montavon, T. J.; Festin, G. M.; Cooke, H. A.; Shen, B.; Bruner, S.
D. J. Am. Chem. Soc. 2007, 129, 15744-15745.

(18)

Strom, S.; Wanninayake, U.; Ratnayake, N. D.; Walker, K. D.; Geiger, J. H. Angew.
Chem. Int. Ed. 2012, 51, 2898-2902.

(19)

Kyndt, J. A.; Meyer, T. E.; Cusanovich, M. A.; Van Beeumen, J. J. FEBS Lett. 2002, 512,
240-244.

(20)

Camm, E. L.; Towers, G. H. N. Phytochemistry 1973, 12, 961-973.

(21)

MacDonald, M. J.; D'Cunha, G. B. Biochem. Cell Biol. 2007, 85, 273-282.

(22)

Ratnayake, N. D.; Wanninayake, U.; Geiger, J. H.; Walker, K. D. J. Am. Chem. Soc.
2011, 133, 8531-8533.

(23)

Hermes, J. D.; Weiss, P. M.; Cleland, W. W. Biochemistry 1985, 24, 2959-2967.

(24)

Bordwell, F. G.; Zhao, Y. Y. J. Org. Chem. 1995, 60, 6348-6352.

(25)

Calabrese, J. C.; Jordan, D. B.; Boodhoo, A.; Sariaslani, S.; Vannelli, T. Biochemistry
2004, 43, 11403-11416.

(26)

Mutatu, W.; Klettke, K. L.; Foster, C.; Walker, K. D. Biochemistry 2007, 46, 9785-9794.

(27)

Klee, C. B.; Kirk, K. L.; Cohen, L. A. Biochem. Biophys. Res. Commun. 1979, 87, 343348.

(28)

Klee, C. B.; Kirk, K. L.; Cohen, L. A.; McPhie, P. J. Biol. Chem. 1975, 250, 5033-5040.

(29)

Schuster, B.; Retey, J. Proc. Natl. Acad. Sci. U. S. A. 1995, 92, 8433-8437.

(30)

Pilbak, S.; Farkas, O.; Poppe, L. Chem-Eur. J. 2012, 18, 7793-7802.

(31)

Szymanski, W.; Wu, B.; Weiner, B.; de Wildeman, S.; Feringa, B. L.; Janssen, D. B. J.
Org. Chem. 2009, 74, 9152-9157.

(32)

Wu, B.; Szymanski, W.; Wietzes, P.; de Wildeman, S.; Poelarends, G. J.; Feringa, B. L.;
Janssen, D. B. ChemBioChem 2009, 10, 338-344.

(33)

Bohm, H. J.; Banner, D.; Bendels, S.; Kansy, M.; Kuhn, B.; Muller, K.; Obst-Sander, U.;
Stahl, M. Chembiochem 2004, 5, 637-643.

(34)

Kim, C. Y.; Chang, J. S.; Doyon, J. B.; Baird, T. T.; Fierke, C. A.; Jain, A.; Christianson,
D. W. J. Am. Chem. Soc. 2000, 122, 12125-12134.

187

(35)

Schuster, B.; Rétey, J. Proc. Natl. Acad. Sci. U.S.A. 1995, 92, 8433-8437.

(36)

Hammett, L. P. J. Am. Chem. Soc. 1937, 59, 96-103.

(37)

Hoffmann, J.; Klicnar, J.; Štěrba, V.; Večeřa, M. Collect. Czech. Chem. Commun. 1970,
35, 1387-1398.

(38)

Fernández, I.; Wu, J. I.; von Ragué Schleyer, P. Org. Lett. 2013, 15, 2990-2993.

(39)

Zavodszky, M. I.; Rohatgi, A.; Van Voorst, J. R.; Yan, H.; Kuhn, L. A. J. Mol. Recognit.
2009, 22, 280-292.

(40)

Zavodszky, M. I.; Sanschagrin, P. C.; Korde, R. S.; Kuhn, L. A. J. Comput. Aided Mol.
Des. 2002, 16, 883-902.

(41)

Nicholls, A.; Wlodek, S.; Grant, J. A. J. Comput. Aided Mol. Des. 2010, 24, 293-306.

(42)

Wlodek, S.; Skillman, A. G.; Nicholls, A. J. Chem. Theory Comput. 2010, 6, 2140-2152.

(43)

Halgren, T. A. J. Comput. Chem. 1996, 17, 490-519.

(44)

Hawkins, P. C.; Skillman, A. G.; Warren, G. L.; Ellingson, B. A.; Stahl, M. T. J. Chem.
Inf. Model. 2010, 50, 572-584.

(45)

Hawkins, P. C.; Nicholls, A. J. Chem. Inf. Model. 2012, 52, 2919-2936.

(46)

Carpy, A. J. M.; Haasbroek, P. P.; Ouhabi, J.; Oliver, D. W. J. Mol. Struct. 2000, 520,
191-198.

(47)

Batsanov, S. S. Russ. Chem. Bull. 1995, 44, 18-23.

(48)

Li, A. J.; Nussinov, R. Proteins 1998, 32, 111-127.

188

Chapter 5: Biocatalytic Production of β-Aryl-β-Amino Acids using PaPAM

5.1. Introduction
5.1.1. Structural diversity and Significance of β-Amino Acids
β-Amino acids are isomers of α-amino acids where the amino group is attached to the βcarbon instead of the α-carbon. In contrast to two side chain substitution positions in α-amino
acids, β-amino acids contain four substitution positions (Figure 5.1A). Consequently, β-amino
acids with specific side chains can exist as various stereoisomers either at the α- or β-carbon. The
flexibility to generate a vast range of regio- and stereo- isomers (Figure 5.1B),1,2 significantly
expands the structural diversity of β-amino acids thereby providing a diverse array of structural
elements for molecular design.

Figure 5.1 Comparison between the substitution patterns of α- and β-amino acids (A) and
fundamental constitutional-isomers of β-amino acids (B). * Indicates a chiral carbon where two
stereoisomers are possibly formed increasing the structural diversity of β-amino acids.

Non-proteinogenic β-amino acids are not as abundant in nature as α-amino acids.
However, they occur in nature as key components of pharmacologically important natural
189

products (Table 5.1). Aromatic β-amino acids are found in antibiotic agent andrimid3 (1) from
Pantoea agglomerans, potent antifungal and insecticidal agent jasplakinolide4 (2) from marine
sponges, and potent aminopeptidase inhibitor bestatin,5 (3) from Streptomyces olivoreticuli.
Furthermore, pharmaceuticals such as sitagliptin6 (Merck) (4) and Taxol7 (Bristol-Myers Squibb)
(5) contain moieties derived from β-amino acid precursors. A CCR-5 receptor antagonist
maraviroc (Pfizer) (6), which is a treatment for HIV infection is built from (S)-β-phenylalanine
as a synthetic intermediate (Table 5.1).8
β-Amino acids have drawn the greatest attention as building blocks for synthetic
peptides. Unlike α-peptides, oligomers composed of β-amino acids with proteinogenic side
chains are stable against proteolytic, hydrolytic and metabolizing enzymes. 9,10 The stability of βpeptides occur likely due to the lack of substrate recognition by peptidases or proteases and the
change of electronic environment of peptide bond upon homologation of α-amino acids. Another
important aspect of β-peptide oligomers is their ability to fold into well-defined and stable
conformations in solution as well as in the solid state.11 In contrast to their α-peptidic
counterparts, β-peptides with chain lenghts as short as four residues form helical-, turn- and
pleated-sheet conformations even in protic solutions (MeOH, H2O).11,12 The in vivo stability and
the diversity of secondary structures enhance the potential application of β-peptides as
peptidomimetics in medicinal chemistry. A cyclic-β-tetrapeptide (7) was shown to mimic the
natural peptide hormone somatostatin (regulator of endocrine and nervous system function), and
display biological activity and affinity for human somatostatin receptors.13 In addition to
peptidomimetic ability, poly-β-peptides display other bioactivity properties as well. More
recently, a new family of nylon-3 polymers (poly-β-peptides) (8) with significant and selective
toxicity towards the human fungal pathogen Candida albicans was reported (Table 5.1).14
190

Table 5.1 Bioactive natural products, pharmaceuticals, and β-peptides based on β-amino acids
β-Amino Acid
Component

Therapeutic
Potential

1

(3S)-βphenylalanine

Antibiotic

2

(3R)-β-tyrosine

Antifungal and
insecticidal

3

(2S,3R)--phenylβ-homoisoserine

Aminopeptidase
inhibitor

4

2',4',5'-trifluoroβ-homoalanine

Antidiabetic

5

(2R,3S)-βphenylserine

Anticancer

6

(3S)-βphenylalanine

CCR-5 receptor
antagonist

7

Cyclo-βtetrapeptide

Human hormone
somatostatin mimic

8

Poly-β-peptide

Antifungal

Compound

y = 40, 50, 60, 70, 80 or 90; x + y = 100

191

5.1.2. Chemical Approaches for Asymmetric Synthesis of β-Amino Acids
Given the importance of β-amino acids as precursors for pharmaceuticals and
peptidomimetics, a plethora of methods are known for the stereoselective chemical synthesis of
β-amino acids.15-17 Arndt-Eistert homologation of α-amino acids, diastereoselective and
enantioselective conjugate addition reactions, and catalytic asymmetric hydrogenation reactions
are highlighted among the various advantageous methods developed.
5.1.2.1. Arndt-Eistert Homologation
Preparation of β-amino acids from α-amino acids via Arndt-Eistert homologation (Figure
5.2) is particularly attractive, since the stereogenic center of α-amino acids is retained during the
process without significant racemization. Additionally, the α-amino acids are readily available
with low cost and high enantiopurity. Yuan and coworkers have successfully employed the
Arndt-Eistert homologation to synthesize the β-homoarginine subunit (58% yield) of TAN-1057A, and -B dipeptides with antibacterial activity.18

Figure 5.2. Synthesis of β-amino acids from α-amino acids via Arndt-Eistert homologation.

192

However, costly silver catalyst used in the Wolff rearrangement limits the use of this
method for large scale syntheses. In contrast, photochemically induced rearrangement is more
suitable for the development of a cost-effective, environmentally friendly process. In the
synthesis of the tripeptide Boc-β-HVal-β-HAla-β-HLeu-OMe, the photochemical Wolff
rearrangement was demonstrated to be more effective over the silver catalysis.9
In addition to the use of silver catalysis, safety concerns and difficult handling associated
with carcinogenic, highly volatile, and hazardous diazomethane (CH2N2) is a major drawback in
Arndt-Eistert homologation method. Nevertheless, in a recent example, risk of handling CH2N2
was reduced by generating, extracting, and using diazomethane in a continuous flow system
(Figure 5.3).19 Here in this system, all four successive steps of the Arndt-Eistert homologation
synthesis; 1) activation of α-amino acid 9 to the mixed anhydride 10, 2) synthesis of αdiazoketone 11 by reacting the mixed anhydride with CH2N2, 3) photochemical Wolff
rearrangement of 11, and 4) reaction of intermediate ketene 12 with water to form N-protected βamino acid were carried out in a flow system. CH2N2 was generated in a microreactor
environment and directly extracted from the aqueous feed using a micro-porous gas-permeable
membrane. The anhydrous CH2N2 extracted was then reacted with the activated α-amino acid.
Excess CH2N2 was removed from the reaction stream employing a second gas-selective
membrane and destroyed in a quench solution thus avoiding the exposure of CH2N2. This
method was used to synthesize a variety of N-protected β-amino acids with modest yields (8
examples, 34-54% yield).

193

Figure 5.3. Continuous four-step flow system for the production of N-protected β-amino acids. RT:
residence loops; TiT: tube-in-tube reactor.

5.1.2.2. Conjugate Addition Reactions
Conjugate addition of nitrogen nucleophiles to an α,β-unsaturated carboxylic acid
derivatives is among the most important and atom-economic asymmetric synthetic strategies for
the synthesis of β-amino acids. Diastereoselective methods are composed of addition of chiral
nucleophiles across the double bond (Figure 5.4A) or addition of an achiral nucleophile to a
chiral acceptor (Figure 5.4B). Although very challenging, some enantioselective methods (Figure
5.4C) were developed in last few decades.

Figure 5.4. Various Conjugate addition approaches for the synthesis of β-amino acids. A and B)
Diastereoselective methods involving a chiral nucleophile and a chiral template. C) Enantioselective
approaches.
194

5.1.2.2.1. Diastereoselective Methods
Several chiral amines readily available in both enantiomeric forms are used as synthetic
ammonia equivalents in diastereoselective Michael addition reactions.20 Davies and co-workers
demonstrated the highly diastereoselective (>99% de) addition of lithium R-N-benzylphenylethylamide to E-crotonate esters and methyl p-benzyloxycinnamate (Figure 5.5).21
Subsequent debenzylation with Pd(OH)2 and acid hydrolysis produced R-β-amino butanoic acid
and S-β-tyrosine in excellent enantioselectivity. More recently this diastereoselective strategy
was employed in the asymmetric synthesis of α-deuterio-β3-phenylalanine derivatives.22

Figure 5.5. Synthesis of chiral β-amino acids using diastereoselective Michael addition of a chiral
amine.

Based on computational modeling analyses, a model has been proposed for the origin of
highly stereoselective conjugate addition.23 According to theoretical calculations, dibenzylamide
nucleophile adopted a stable conformation where the phenyl rings are parallel to each other. In
the lowest energy transition state (18), the α,β-unsaturated acceptor reacts from its s-cis
conformation, lithium is chelated between the carbonyl oxygen and the nitrogen lone pair, and
the (R)-amide favors the 3-si face of the unsaturated ester (Figure 5.5).

195

In addition to the strategies involving the chiral nucleophiles, diastereoselective α,βadditions are carried out using chiral acceptors and achiral nucleophiles as well. For example
chiral chrotonates24 and chiral p-tolylsulfinyl cinnamates25 were successfully used as the chiral
templates for stereoselective conjugate addition reactions.
5.1.2.2.2. Enantioselective Methods
In contrast to the diastereoselective methods, enantioselective conjugate amine addition
reactions are generally more challenging. The first enantioselective conjugate addition reaction
was developed by JØrgensen and coworkers in 1996 using a BINOL catalyst with only 42%
enantioselectivity.26 Following this initial report, various Lewis acid catalyzed amine additions
with good isolated yields and excellent levels of enantioselectivity were reported. Although a
variety of alkyl substituents were successfully used in chiral Lewis acid methodologies, β-aryl
groups are considerably less reactive than the corresponding β-alkyl substrates. Sibi and
coworkers reported the first example of highly enantioselective conjugate addition to cinnamates
using catalytic amounts of a chiral Lewis acid.27 Various β-aryl-β-amino acid derivatives were
synthesized with moderate to excellent ee using the addition of a highly reactive nitrogen
nucleophile N-benzylhydroxylamine to the pyrrolidinone-derived enoate 19 (Figure 5.6).

Figure 5.6. Enantioselective addition of N-benzylhydroxylamine to pyrrolidinone-derived enoates in
presence of a chiral Lewis acid.
196

5.1.2.3. Catalytic Asymmetric Hydrogenation Reactions
Asymmetric hydrogenation of acrylic acid or nitrile derivatives is one of the most
promising chemical approaches for the large-scale synthesis of β-amino acids. Since the initial
report of asymmetric hydrogenation of N-acyl-β-(amino)acrylates by Noyori,28 Ru- and Rhcatalyzed homogeneous hydrogenations using chiral phosphorous ligands became a well
established procedure. However, both (E)- and (Z)-isomers of β-acylamido acrylates are
produced simultaneously in common synthetic protocols, and their individual hydrogenation
requires prior separation. When racemic β-(amino)acrylates are employed, (E)-isomers generally
lead to higher enantioselectivities, and (Z)-isomers frequently react faster, with low
enantioselectivity.29 Nonetheless, a collaborative effort of Bӧrner et. al. and Evonik Degussa
GmbH R&D discovered the chiral bisphospholane ligand MalPHOS (marketed as catASium,
Solvias AG, Switzerland) bearing a maleic anhydride backbone for the Rh(I)-catalyzed
enantioselective hydrogenation of isomeric β-acylamido acrylates (Figure 5.7).30 With this new
catalyst, comparable selectivities were obtained for both (E)- and (Z)-isomers and the
enantioselectivity was higher particularly for the (Z)-configured substrates bearing bulkier
substituents at the β-position (Et, i-Pr, Ph).

Figure 5.7. Hydrogenation of enamides using chiral bisphospholane ligand MalPHOS.
197

The enantioselectivity of asymmetric hydrogenation was further improved using a series
of diphosphine ligands based on camphor.31 The hydrogenation of methyl-3-acetylamino-but-2enoates proceeded with >99% and 94% ee, respectively, for the E- and Z-isomers.
Although ligand design and tuning have improved the enantioselectivity (>95%) of the
catalytic asymmetric hydrogenation of unsaturated double bonds, this route is still complex for a
large-scale production partly due to the more stages and reaction steps. The need for the largescale manufacture of the required complex ligands and starting materials will produce more
waste and impede the development of a commercial process for the production of β-amino
acids.32 Consequently, more applicable biocatalytic methods have been developed for the
synthesis of β-amino acids and their analogues.

5.1.3. Biochemical Approaches for Synthesis of β-Amino Acids
Biocatalysis, use of enzymes as catalysts in synthetic organic chemistry offer some
unique advantages over conventional enantioselective chemical catalysis.33,34 The most important
advantage of a biocatalyst is the excellent stereo-, regio- and chemo-selectivity. Additionally, no
protection/deprotection steps are required, and therefore, the reactions are generally atom- and
reaction-step economical. Other advantages, such as mild operational conditions and reduced
hazardous waste generation are also very attractive in commercial applications.34
Narrow substrate scope, substrate or product inhibition, lower stability and difficulty of
enzyme production in large enough quantities for practical applications were often considered as
the most serious drawbacks of biocatalysts.34 Nonetheless, advent of recombinant DNA
technology, developments in protein design tools such as rational design and directed evolution,
198

and advances in understanding protein structure–function relationships are enabling scientists to
rapidly tailor the properties of biocatalysts for particular chemical processes.33 Substrate
specificity, stability, activity, selectivity, and large scale production of enzymes are routinely
engineered in the laboratory. Presently, several biocatalytic processes are implemented in
pharmaceutical, chemical, agricultural, and food industries.35
Three main biocatalysts 1) lipases,32,36 2) transaminases,37 and 3) aminomutases38,39 are
explored for the synthesis of β-amino acids. Compared to the enantioselective syntheses using
isomerases and transaminases, enzymatic resolution of lipases is widely investigated in the
context of industrial scale synthesis of β-amino acids.32

5.1.3.1. Enzymatic Resolution of β-amino acids
One of the earliest, most efficient method for enzymatic resolution of β-amino acids is
the use of penicillin G acylase from E. coli. Soloshonok and co-workers prepared (R)-β-aryl-βamino acids in good yield and high enantiomeric purity (>95% ee) via penicillin G acylase
catalyzed hydrolysis of corresponding N-phenylacetyl derivatives.40 Synthesis of the (S)-isomer
requires separation of the unreacted amide and subsequent acid hydrolysis (Figure 5.8). In
addition to the complex procedure of (S)-isomer production, the need to separate the products
from corresponding aryl acetic acids, and lower atom economy arose from the high mass of
phenylacetyl protecting group makes this process less attractive for an industrial scale
manufacture of β-amino acids.

199

Figure 5.8. Penicillin G acylase route to β-aryl-β-amino acids

Recently, a lipase catalyzed commercial manufacturing route for β-aryl-β-amino acids
was developed by modifying the initial conditions reported by Faulconbridge et.al. 36 With the
modified conditions, the commercially available Amano lipase PS was capable of hydrolyzing a
range of aromatic and heteroaromatic propyl esters with excellent enantioselectivity (98-99% ee)
and low to high yields (16-50%) (Figure 5.9).32

Figure 5.9. Amano lipase PS catalyzed production of β-aryl-β-amino acids.

Enantioselective production of (S)-β-phenylalanine using this lipase route was carried out
by Evonik Degussa GmbH (Germany) at variuos scales: kilo-lab (3.7-7.4 kg), pilot plant (100
200

kg), and commercial plant scale (430 kg) without affecting the yield of the reaction (41-43%).
The chemical purity of the isolated (S)-β-phenylalanine was 98.5-99.5% and the optical purity
was >99.5% in each case. The reaction was carried out in a two phase liquid mixture where the
lipase and the racemic-propyl ester is dissolved in water and MTBE, respectively. After the
reaction, (S)-acid crystallizes out from the two phase liquid mixture, and thus the process
eliminates additional purification steps.
Although the lipase route was developed for the commercial production of β-aryl-βamino acids, the enzymatic resolution strategies are limited to a maximum 50% yield.
Additionally, corrosive reagents such as SOCl2 is involved in the synthesis of the racemic propyl
esters. Furthermore, distillation and extraction steps added before the two-phase lipase resolution
reaction. The Amano lipase PS, which contributes most to the process cost, was partially
inactivated after the first cycle. A fresh enzyme supplement was required for the recycling of the
lipase solution for further catalytic cycles.32
Use of aminomutases and transaminases37 are promising over the enzymatic resolution
efforts since the former reactions can lead theoretically to a 100% conversion of the substrate.
Furthermore, implementing a whole-cell biocatalytic route instead of in vitro enzymes would
likely enhance catalyst stability and promote sustainability. For multiple rounds of catalysis, this
approach would contrast the Amano lipase PS route where in vitro biocatalytic activity was
reduced over time.

201

5.1.3.2. Aminomutases for the Production of β-Amino Acids
In nature, aminomutases are involved in the catabolism of amino acids and biosynthesis
of various biologically active natural products.38 These enzymes catalyze the vicinal exchange of
a proton and an amine group present in the substrate. Members of this family, lysine 2,3-,
glutamate 2,3-, phenylalanine- and tyrosine-aminomutases are capable of isomerizing α-amino
acids into β-amino acids that are part of pharmaceuticals and bioactive natural products.3,38,41-44
However, lysine- and glutamate-2,3-aminomutases belong to the SAM-dependent family, and
anaerobic conditions are required to purify these enzymes.42,45 In addition, multiple expensive
cofactors are needed if these enzymes are used in an in vitro biocatalytic processes of lysine- and
glutamate-2,3-aminomutases. By contrast, MIO-dependent phenylalanine- and tyrosineaminomutases function under aerobic conditions and are not dependent on external cofactors.46,47
Phenylalanine aminomutases (PAM) are the most extensively studied enzymes of this
class of aminomutases. In Taxus plants, TcPAM/TchPAM catalyzes the conversion of (2S)-αphenylalanine to (3R)-β-phenylalanine, which is the proposed biosynthetic precursor of the
phenylisoserine side chain of Taxol (Figure 5.10).38 A homologous enzyme from Pantoea
agglomerans (PaPAM) produces the enantiomeric (3S)-β-phenylalanine in the biosynthetic
pathway to antibiotic andrimid (Figure 5.10).39 Accessibility to the both enantiomeric forms of βphenylalanines using homologous enzymes significantly improve the synthetic value of PAM as
a biocatalyst.

202

Figure 5.10. Reactions catalyzed by phenylalanine aminomutases PaPAM and TcPAM in their
corresponding biosynthetic routes for andrimid and Taxol.

The enzymatic conversion catalyzed by PAMs proceed with excellent enantioselectivity
(>99.9%) and these enzymes have a broad substrate scope. A range of α-phenylalanine analogues
consisting of various substituents at the phenyl ring (fluoro, chloro, bromo, methyl, methoxy and
nitro) as well as other heteroaromatic analogues were successfully converted to corresponding βamino acids by both PAMs.48,49
Recently, various other routes toward enantiomerically pure β-amino acids have been
investigated with MIO-dependent aminomutase catalysis. TchPAM from Taxus chinensis
catalyses the highly enantioselective addition of ammonia to t-cinnamic acid, which is an
intermediate in the isomerization reaction. Addition of ammonia to cinnamic acid produces a 1:1
mixture of (S)-α- and (R)-β-phenylalanines with >99.9% ee.50 This amino addition reaction has a
broad substrate scope and produces variously substituted α- and β-phenylalanines. While
enantioselectivity was not affected, the regioselectivity of this process was significantly affected
by the substituents at the aromatic ring. For example, ortho-substituted cinnamic acids resulted
203

almost exclusive production of (S)-α-phenylalanine analogues with >98:2 ratio.51 Furthermore,
the regioselectivity of amino addition to cinnamic acid by TchPAM was tailored towards the (R)β-phenylalanines by engineering the active site residues.52
Additionally, the synthetic utility of PAMs was recently exploited as an
aminotransferase.53 TcPAM from Taxus canadensis was employed as an amino acid:arylacrylate
transaminase. The amino group from an α-amino acid ((S)-styryl-α-alanine and some other nonnatural amino acids) was successfully transferred to an arylacrylate to produce a mixture of αand β-amino acids. Hence, in an organic synthesis context, the isomerization reaction catalyzed
by PAMs is a highly enantioselective, atom economical, cost-effective, and a relatively
environmentally benign biocatalytic method for the production of various β-aryl-β-amino acids.
However, PAMs have not yet evaluated as whole-cell biocatalysts, which could be applicable in
large scale production of β-aryl-β-amino acids.

5.1.4. Whole-cell Biocatalysis
Biocatalysis may be carried out using whole cells or isolated enzymes.33 Compared to
purified enzymes, whole-cell biocatalysts or so called 'designer cells' have several advantages
that are particularly attractive for large-scale processes. The use of microbial whole-cells
eliminates costly, laborious enzyme purification steps.54 In whole-cells, enzymes enclosed in
their typical environments protected by the cell envelopes are more stable than purified enzymes.
Thus, whole-cell biocatalytic processes which use inexpensive biomass such as bacteria and
yeast are cost-effective, and the risk of denaturing the enzymes is reduced.54

204

Chemical transformations requiring multiple enzymes are quite complicated with isolated
enzymes. Optimizing conditions such as temperature, pH, and reaction buffer for multiple
enzymes in one process is often challenging.55 Tailor-made whole cells that contain all the
desired enzymes for a specific transformation can overexpress multiple enzymes and perform
multi-step transformations under similar conditions. For example, May and co-workers have
developed a production route for L-methionine using a tailor-made recombinant whole-cell
biocatalyst.55 E.coli cells were designed to express a modified

L-hydantoinase

and

L-N-

carbamoylase from Arthrobacter sp. DSM 9771, and a hydontoin racemase from Arthrobacter
sp. DSM3747 (Figure 5.11). This three-enzyme system synthesized 91 mM L-methionine from
D,L-5-(2-methylthioethyl)hydantoin (D,L-MTEH)

substrate. Furthermore,

L-tert-leucine

in less than 2 h with >90% conversion of the

can also be synthesized with the hydantoinase–

carbamoylase system on a 100-kg scale.56

Figure 5.11. Multiple-enzyme catalyzed production of L-methionine from racemic MTEH.
205

In addition to aforementioned advantages, whole-cell systems allow the use of cofactor
pools intrinsic to the cells, and thus, makes cofactor regeneration much easier. Synthesis of chiral
alcohols using asymmetric reduction of ketones catalyzed by NAD(P)H-dependent alcohol
dehydrogenases is often limited by the higher cost of adding the cofactor exogenously. Therefore
an enzyme-coupled cofactor-regeneration system has been developed for the in situ regeneration
of NAD(P)H.57 R-Specific alcohol dehydrogenase (ADH) from Lactobacillus kefir and S-specific
ADH from Rhodococcus erythropolis were coupled with a cofactor-regenerating enzyme glucose
dehydrogenase from Bacillus subtilis in a whole-cell biocatalytic system (Figure 5.12). This
system operates in aqueous media at substrate concentrations >100 g/L without an added external
cofactor to produce various functionalized alcohols with >99% ee and >90% conversion.

Figure 5.12. Cofactor regeneration strategy using glucose dehydrogenase.

Despite the numerous advantages, reactions catalyzed by whole-cell biocatalytic systems
are much slower than those catalyzed by free enzymes.58 The envelopes of microbial cells often
retard the penetration of substrates into the cell and thus prevents the product from being
released to the culture medium for an easy recovery.58,59 Gram negative bacteria is encased by
inner and outer cell membranes.60 The inner cell membrane is a phospholipid bilayer containing
various membrane proteins. Hydrophobic molecules can penetrate the lipophilic bilayer faster
206

than hydrophilic compounds. However, hydrophilic molecules need specific transport proteins to
cross the inner cell membrane. The outer membrane contains a densely packed amphiphilic
lipopolysaccharide layer and acts as a barrier to both hydrophobic and hydrophilic molecules.60
Small hydrophilic molecules (<600 Da) are transported into the cells via non-specific
transmembrane pores called porins.61 Nevertheless, passive diffusion through the highly ordered
lipopolysaccharide layer is the only method for large synthetic substrates (>600 Da) to cross the
outer cell membrane. Consequently, the E. coli cell membrane acts as a mass transport barrier
between the large, non-natural substrates in the bulk medium and the biocatalyst(s) enclosed in
the cell.
Numerous studies have been carried out to alter the cell membrane permeability for
chosen substrates. Solvents62 (ethanol, isopropyl alcohol, toluene, diethyl ether, chloroform),
detergents such as Triton X-100,63 salt stress,64 freeze-thaw cycles,65 EDTA, and
electropermeabilization66 are common physical and chemical cell permeabilization methods. In
addition, molecular engineering approaches have been exploited to modulate the cell
permeability. Mutations in genes that altered the composition of the lipopolysaccharide layer of
the outer cell membrane has significantly accelerated the whole-cell biocatalysis reactions by
reducing the permeability barrier.59 Furthermore, cell surface display or displaying enzymes on
the microbial cells membranes has circumvented the need for cell permeabilization methods. 67
This method allows an enzyme free access to substrates without the need to cross the cell
membrane. In addition, the surface display approach eliminates the undesired side reactions that
could possibly occur inside the cell which comprises a broad variety of other enzymes.

207

Due to the numerous advantages, whole-cell biocatalytic systems are emerging as
efficient alternative for conventional catalysis. Currently, various whole-cell biocatalysts are
implemented in the manufacture of fine chemicals.68,69 However, state-of-the-art biocatalytic
method for β-arylalanine production is based on purified Amano lipase PS, and there are several
limitations associated with this in vitro method (see section 5.1.3.1). The investigation described
herein involved the development of an efficient and sustainable PaPAM based whole-cell
biocatalytic system for producing β-arylalanines.

208

5.2. Experimental

5.2.1. Substrates, Authentic Standards and Reagents
(S)-α-,

p-methoxy-(S)-α-,

p-nitro-(S)-α-,

and

p-chloro-(R/S)-β-phenylalanine and

(trimethylsilyl) diazomethane (2.0 M in diethyl ether) were purchased from Sigma-Aldrich-Fluka
(St. Louis, MO). Racemic p-nitro-β-phenylalanine was purchased from Oakwood Products, Inc.
(West Columbia, SC), and o-methoxy-(S)-α-, m-methoxy-(S)-α-, o-nitro-(S)-α-, m-nitro-(S)-α-, omethoxy-(S)-β-, m-methoxy-(S)-β-, o-nitro-(S)-β-, and m-nitro-(S)-β-phenylalanine were
purchased from Chem-Impex International, Inc. (Wood Dale, IL). All other (S)-α- and -β-amino
acids were purchased from PepTech Corporation (Burlington, MA). All chemicals were used
without further purification, unless noted.

5.2.2. Bacterial Strains, Plasmids, and Culture Media
BL21(DE3) E. coli bacterial strain transformed with expression vector pET24b(+) was
used for the whole-cell biotransformations with PaPAM. E. coli cells were grown in M9 minimal
medium [Na2HPO4.7H2O (12.8 g•L–1), KH2PO4 (3 g•L–1), NaCl (0.5 g•L–1), NH4Cl (1 g•L–1),
MgSO4, (2 mM), CaCl2 (0.1 mM), 100x Basal Medium Eagle vitamins (Sigma-Aldrich, St.
Louis, MO; 10 mL•L–1), and glucose (20%), pH 7.4] supplemented with kanamycin (Gold
Biotechnology Inc., St. Louis, MO; 50 μg•mL–1). Optical density measurements of cell
suspensions were obtained in 1.5 mL polystyrene cuvettes (General Laboratory Supply,
Pasadena, TX) using a UV−visible spectrophotometer (Beckmann DU 640, Beckmann Coulter,
Brea, CA).
209

5.2.3. General Instrumentation: GC/EI-MS Analysis
GC-MS analysis was performed on an Agilent 6890N gas chromatograph equipped with
a capillary GC column (30 m × 0.25 mm × 0.25 μM; HP-5MS; J&W Scientific) with helium as
the carrier gas (flow rate, 1 mL/min). The injector port (at 250 °C) was set to splitless injection
mode. A 1-μL aliquot of each sample was injected using an Agilent 7683 auto-sampler (Agilent,
Atlanta; GA). The column temperature was increased from 50 – 110 °C at 30 °C/min, then
increased by 10 °C/min to 250 °C (total run time of 16 min), and returned to 50 °C over 5 min,
with a 5 min hold. The gas chromatograph was coupled to a mass selective detector (Agilent,
5973 inert) operated in electron impact mode (70 eV ionization voltage). All spectra were
recorded in the mass range of 50 – 400 m/z.

5.2.4. General Procedure for Whole-Cell Biocatalytic Incubations
BL21(DE3) E. coli cells (50 mL, OD600 ~1.0) transformed to express the papam gene
from pET24b(+) were used to inoculate M9 minimal medium (1 L) supplemented with
kanamycin (50 μg/mL). The cells were grown at 37 C until to OD600 ~0.6, isopropyl-β-Dthiogalactopyranoside (100 μM) was added to induce the expression of papam, and the cultures
were incubated for 16 h at 16 C. The next steps were performed at 4 C, unless indicated
otherwise. The cells were harvested by centrifugation at 3,230g (10 min), and the cell pellet was
resuspended separately in three different whole-cell feeding media (M9 minimal medium, 50
mM phosphate buffer at pH 7, and 50 mM phosphate buffer at pH 8), each typically adjusted
separately to an optical density of 35. For analyses at higher biomass, the optical density was
adjusted to 70 and 280 in M9 minimal medium. The α-amino acid substrate (1 mM) was added

210

to the cell suspension and incubated separately at 16, 25 and 30 C. In control experiments,
transformed E. coli cells were incubated without substrate added. Additional control experiments
included assays with E. coli transformed with an empty pET-24 vector incubated with or without
the substrate. All the biotransformation feeding assays were done in triplicate.
The minimum incubation volume for whole-cell biocatalysis was assessed by adding 20
mM of α-phenylalanine to a cell suspension of E. coli (OD600 ~35) engineered to express papam
at 16 C, in 5, 100, or 1000 mL of M9 minimal medium. The cultures were incubated,
respectively, in 50-mL screw-cap conical centrifuge tubes (Corning Incorporated Life Sciences,
Tewksbury, MA), 250 mL Erlenmeyer flasks, and 2.8 L Fernbach flasks with agitation on a
shaker (225 rpm, MaxQ 5000, Thermo Scientific, Waltham, MA) for 6 h. The cells were
removed by centrifugation, and the β-phenylalanine in the culture medium of each assay was
derivatized and quantified by GC/EI-MS. The yield of β-phenylalanine produced was nominally
~2 mg•L–1 for bacteria incubated in each volume of culture medium. Thus, a 5-mL volume of
culture medium was selected for the bacterial incubations in pilot-scale assays to assess the
production levles of the α-arylalanines to their β-isomers in vivo.

5.2.5. Derivatization and Quantification of Amino Acids
To terminate the incubation, the reaction medium was separated from the cells by
centrifuging at 3,230g for 10 min, and the supernatant (1 mL) from each assay was basified to
pH 10 (6 M NaOH). Internal standards m-fluoro-β-phenylalanine, p-methyl-β-phenylalanine, and
β-phenylalanine at 20 μM were added, respectively, to quantify three sets of biosynthetic βamino acids products–Set 1: β-phenylalanine; o-, m-, and p-methyl-; o-, m-, and p-methoxy-; m-

211

and p-nitro-; m- and p-chloro-β-phenylalanine; Set 2: o- and p-fluoro; m-, and p-bromo-βphenylalanine; and (2-thienyl)- and (3-thienyl)-β-alanine; and Set 3: m-fluoro-β-phenylalanine.
Each solution was treated with ethylchloroformate (50 μL) for 10 min, basified again to pH 10,
and reacted with a second batch of ethylchloroformate (50 μL) for 10 min. The solutions were
acidified to pH 2-3 (6 M HCl) and extracted with diethyl ether (2 × 2 mL). The organic fraction
was removed under vacuum, and to the resulting residue dissolved in ethyl acetate:methanol
(3:1, v/v) (200 μL) was added (trimethylsilyl)diazomethane until the yellow color persisted. The
derivatized aromatic amino acids were quantified by GC/EI-MS. The peak area was converted to
concentration by solving the linear equation obtained from the standard curves constructed with
the corresponding authentic standards.

5.2.6. Analysis of Substrate Uptake and Product Release by E .coli Cells
To E. coli cells expressing PaPAM incubated in M9 minimal medium (45 mL, OD600
~35) was added α-phenylalanine (1 mM) at 16 C. Aliquots (5 mL) were withdrawn at 1, 2, 4, 6,
and 8 h, the cells were harvested by centrifugation (3,230g, 5 min), and the culture medium was
twice serially basified (6 M NaOH) and treated with ethyl chloroformate. The solution was
acidified (pH 2-3, 6 M HCl), extracted with diethyl ether (2 × 2 mL), concentrated in vacuo, and
the amino acids were methyl esterified with (trimethylsilyl)diazomethane for quantification by
GC/EI-MS analysis. In parallel, the cell pellet was resuspended in 50 mM sodium phosphate
buffer (pH 8.0), the cells were lysed by sonication to release the soluble amino acids, and the
cellular debris was removed by centrifugation (3,230g, 30 min). The supernatant was decanted
and the amino acids therein were derivatized by the same reactions used for the culture medium
and analyzed by GC/EI-MS.
212

5.2.7. Effect of Temperature, Time, and Culture Medium Type
After expressing the papam gene in E. coli for 16 h at 16 °C in M9 minimal medium (12
× 1 L), the cultures were centrifuged, the supernatants were decanted, and the cells from each
batch were resuspended separately in M9 minimal medium or 50 mM phosphate buffer (each at
pH 7 or 8) to OD600 ~35. α-Phenylalanine was added (1 mM) to 45-mL cell suspended in each
medium, and the feeding studies were conducted at 16, 25 and 30 C. Aliquots (5 mL) were
withdrawn from each suspension at 0.5, 1, 2, 4, 6, 8, 10, and 12 h. The supernatants were
clarified by centrifugation, decanted, and separately treated with derivatizing reagents to form
the N-(ethoxycarbonyl) α- and β-phenylalanine methyl esters and the cinnamic acid methyl ester,
which were quantified by GC/EI-MS.

5.2.8. Effect of Substrate Concentration
To a cell suspension of E. coli (OD600 ~35), engineered to express papam in 5 mL of M9
minimal medium was added separately 1, 5, 10, 15, 20, and 25 mM of α-phenylalanine. The
cultures were incubated for 6 h at 16 C. The cells were removed by centrifugation, and the βphenylalanine in the culture medium of each assay was quantified after derivatization and
analysis by GC/EI-MS.

5.2.9. Effect of the Biocatalyst Amount
The E. coli cells harboring papam were resuspended in M9 minimal medium to an OD600
of 35, 70 or 280. α-Phenylalanine was added to a final concentration of 1, 5, 10, 15, 20, and 25
mM in separate assays at each cell density, and the assays (5 mL) were incubated for 6 h at 16
213

C. The culture medium was clarified by centrifugation, and the β-phenylalanine in the
supernatant (1 mL aliquot) of each assay was derivatized for and quantified by GC/EI-MS
analysis.

5.2.10. Assessing the Substrate Scope of the Biocatalytic System
α-Phenylalanine, its analogues (ortho/meta/para-methyl, -methoxy, -fluoro, -chloro, bromo, and –nitro), and 2- and 3-thienylalanine were separately incubated in a cell suspension of
engineered E. coli (OD600 ~35) in M9 minimal medium (5 mL) at 16 C for 6 h. The cells were
pelleted by centrifugation, and an aliquot (1 mL) from each supernatant was separately treated
with derivatizing reagents to form the N-(ethoxycarbonyl) methyl esters of the α-arylalanines and
biosynthetic β-arylalanines, and the methyl esters of the biosynthetic arylacrylates for
quantification by GC/EI-MS.

5.2.11. Sustainability of the Biocatalytic System
E. coli cells (at OD600 ~35) engineered to express papam were resuspended in M9
minimal medium (5-mL), and α-phenylalanine at 1 and 5 mM was added separately to different
batches of cell suspensions. The cells were incubated for 6 h at 16 C, the culture medium was
clarified by centrifugation at 3,230g (10 min). The cell pellets were serially washed (2 × 5 mL)
with M9 minimal medium and clarified by centrifugation between each wash to remove residual
substrate/product from the cells. The washed cell pellets were resuspended in the culture medium
(5 mL), a new batch of α-phenylalanine substrate was added to each suspension, and the
biotransformation reactions were incubated for 6 h at 16 C. The supernatant from each reaction
214

was twice serially basified (6 M NaOH) and treated with ethyl chloroformate. The solution was
acidified (pH 2-3, 6 M HCl), extracted with diethyl ether (2 × 2 mL), concentrated in vacuo, and
treated with (trimethylsilyl)diazomethane to methyl esterify the amino acids for quantification by
GC/EI-MS. Cell harvesting, cell pellet washing, incubation, and GC/EI-MS quantification of the
synthetically derivatized β-phenylalanine from the supernatant were repeated for 4 cycles.

5.2.12. Calculation of Colony-Forming Units (CFU)
An aliquot (50 µL) from each cell suspension was serially diluted between 10-fold and
108-fold with the culture medium (in triplicate). The 108-fold diluted culture suspension (100 µL)
was spread on an agar plate supplemented with kanamycin (50 μg/mL) and incubated for 16 h at
37 C. The colonies on each plate were counted, and the CFU was calculated.

215

5.3. Results and Discussion
5.3.1. General Assay Conditions
To construct a whole-cell biocatalyst for the production of β-arylalanines, the papam
gene from Pantoea agglomerans was sub-cloned into the pET24b(+) expression vector. A
BL21(DE3) E. coli strain transformed with pET24b(+) harboring papam was used as the
microbial host. LB medium, which is a rich nutrient broth was avoided since it contains αphenylalanine and would therefore confound in vivo biocatalysis results.70 Thus, papam was
expressed in host cells grown in M9 minimal medium free of amino acids to avoid background
catalysis of α-phenylalanine during gene expression. After overnight expression of papam, a 1-L
cell culture was pelleted and concentrated 10-fold (OD600 ~35) in the culture medium to increase
the effective concentration of PaPAM by increasing the cell density towards fermentation
conditions.

5.3.2. Assessment of the Whole-cell Biocatalytic Properties of PaPAM
Ideally, the starting materials feed into a whole cell biotransformation system should be
transported into the cells with low impedance.58 Reciprocally, the biosynthesized products
should be released into the culture medium for easy recovery without breaking the cells, thus
enabling the cells to be used sustainably in additional biocatalytic cycles. Recent investigations
found that active transport of phenylalanine and other aromatic amino acids into the E. coli cells
across the inner membrane is mediated by five distinct transport systems.71-73 All arylalanines are
transported via a general aromatic amino acid permease (AroP),72 while the permease PheP74
216

specifically transports phenylalanine. A permissive ATP-binding cassette type transporter, LIVI/LS system, principally conveys branched chain amino acids yet also occasionally transports
phenylalanine .75
In this study, each productive arylalanine fed to cultures of engineered E. coli was
biosynthesized to a β-aryl-β-alanine and released into the culture medium for recovery. βPhenylalanine (benchmark product) was quantified in the medium at 0.5, 1, 2, 4, 6, and 8 h after
incubation to evaluate the proportion of β-amino acid remaining in the E. coli cells and culture
medium. After brief zero-order production of β-phenylalanine (i.e., δ[β-Phe]/δt ~0.46 mg·h–1,
between 0 and 0.5 h), the change in production rate (δ[β-Phe]/δt) approached 0 after 1 h, and βphenylalanine reached 90 mg in a 1-L culture (OD600 ~35) after 8 h. The β-phenylalanine (3.2
mg) obtained from the cells (80 g wet wt) in 1-L of medium accounted for ~3% of the total βphenylalanine biosynthesized after 8 h (Figure 5.13). These data suggested the substrate crossed
the E. coli membranes into the cells, while the product released reciprocally into the medium
without the need to disrupt the cells.
The colony-forming units (CFU) in the E. coli samples at each time point (0.5, 1, 2, 4, 6,
and 8 h) remained constant (Figure 5.13), suggesting no time-dependent accumulation of dead
cells and cellular protein in the medium at stationary growth phase. Thus, biocatalysis did not
occur fortuitously in the culture medium by release of PaPAM from dead cells.

217

5
4

0.6

3
0.4
2
0.2

0.0

CFU (mL-1) x 10-11

-Phenylalanine (mg)

0.8

1

0

2

4

Time (h)

6

8

0

Figure 5.13. Distribution of β-phenylalanine in culture medium and E. coli cells. Time course and
distribution of β-phenylalanine released to the M9 minimal medium (pH 7.4, 5 mL) (■ on solid line) and
remaining in the E. coli cell pellet (0.4 g) (♦ on dotted line) after conversion of α-phenylalanine (1 mM)
by PaPAM whole-cell biocatalysis at 16 C (OD600 ~35). CFU values () for the bacterial cells at each
time point are shown.

5.3.3. Enantiomeric Excess of the Biosynthetic (S)-β-Phenylalanine
Driven by the application of chiral building blocks for the synthesis of pharmaceutically
active drugs, the development of industrially viable processes for the synthesis of
enantiomerically pure compounds continues to accelerate.76 Hence, optimizing the enantiomeric
excess of a catalytic process is an important goal. To analyze the enantioselectivity of the wholecell biocatalyst used herein, biosynthetic β-phenylalanine was converted to its N-(1(S)camphanoyl) methyl ester. The derivatized diastereomeric β-amino acid eluted at a retention time
(14.43 min) identical to that of authentic N-[(1S)-camphanoyl]-(3S)-β-phenylalanine (14.43 min)
(Figure 5.14) confirming the 3S-product stereochemistry. There was no indication of N-[(1S)camphanoyl]-(3R)-β-phenylalanine (14.41 min) (Figure 5.14) present in the biosynthetic product
suggesting >99.9% enantiomeric excess of the biosynthetic (3S)-β-amino acid, which was
consistent with the product stereochemistry of the PaPAM reaction in earlier in vitro assays.39,77

218

Relative Total Ion Count
11

12

1.0
14.43

0.8
0.6

14.10

14.43

0.4
0.2
0.0

13

14

15

Time (min)

16

17

18

Figure 5.14. Gas chromatogram profiles of the N-[(1'S)-camphanoyl] methyl esters of authentic
racemate 3R– (14.10 min) and 3S–β–phenylalanine (14.43 min) (solid line), and biosynthetic β–
phenylalanine (14.43 min) (dotted line).

5.3.4. Effect of Temperature on β-Phenylalanine Production
The activity, stability, and selectivity of a whole-cell biocatalyst are important for
industrial scale biosynthesis.78 To examine the thermal stability of the recombinant E. coli
whole-cell biocatalyst, the cell cultures were incubated at 16, 25 or 30 °C. Aliquots from each
thermal sample were withdrawn at 30 min, 1 h, and then every 2 h up to 10 h to measure the βphenylalanine in the medium (Figure 5.15A). Cultures (OD600 ~35) incubated at 16 °C
maximally produced β-phenylalanine (0.092 g·L–1) at 6 h, yet cultures incubated at 25 and 30 °C
nominally produced β-phenylalanine at ~1.5-fold lower maximum (~0.065 g·L–1) at 1 h (Figure
5.15A). β-Phenylalanine increased over 6 h at 16 °C and then equilibrated; while at 25 and 30
°C, the production increased over 1 h and then remain constant up to 8 h. The higher reaction
temperatures likely affected overall gene expression,79 enzyme activity, substrate binding,
increased the flux of α-phenylalanine towards post-exponential Pex protein synthesis,80 and/or
amplified phenylalanine catabolism81 in E. coli at stationary phase. Therefore, the recombinant
E. coli biocatalyst expressing PaPAM was incubated at 16 °C to assess the permissivity for
219

converting other unnatural β-arylalanines. At all three temperatures, the β-phenylalanine
decreased markedly after the equilibrium phase beyond 8-h incubation. Analysis of the product
distribution at each time point at 16 °C revealed that the concentration of cinnamic acid (a 5%
by-product from α-phenylalanine during PaPAM catalysis) increased over time, while the
amount of α-phenylalanine rapidly decreased (Figure A.3.1). An increase in cinnamic acid
suggested that the occurrence of the reverse reaction catalyzed by PaPAM82 is converting βphenylalanine to cinnamate as the β- phenylalanine concentration increased over that of αphenylalanine in the medium. However, α- phenylalanine did not equilibrate with βphenylalanine in the reaction medium, likely due to its partitioning to other metabolic pathways.

A

B
0.10

-Phenylalanine (g/L)

-Phenylalanine (g/L)

0.10
0.08
0.06
0.04
0.02
0.00

0.08
0.06
0.04
0.02
0.00

0

2

4

6

8

10

0

Time (h)

2

4

6

8

10

Time (h)

Figure 5.15. Effect of reaction temperature and reaction medium. A) The production of βphenylalanine in PaPAM whole-cell biocatalysis in M9 minimal medium (pH 7.4) containing αphenylalanine (1 mM) at 16 C (■), 25 C (●), and 30 C (▲). B) The production of β-phenylalanine in
PaPAM whole-cell biocatalysis in M9 minimal medium (pH 7.4) (■), phosphate buffer at pH 7 (♦), and
pH 8 (►), each containing α-phenylalanine (1 mM).

220

5.3.5. The Effect of Reaction Medium on α- to β-Phenylalanine Isomerization
The influence of pH and buffer type on the stability and activity of the whole-cell
biocatalyst was evaluated in phosphate buffer (50 mM, pH 7 and 8) and M9 minimal medium
(pH 7.4). The M9 minimal medium contains a supply of nitrogen, phosphorous and sulfur
source, mono- and di-valent cations (K+, Ca2+, Mg2+ and Fe2+), BME-vitamins, and a carbon
source needed for bacterial cell survival and growth. By comparison, the phosphate buffer lacks
a carbon source as well as other nutrients needed for cell growth and viability. In M9 minimal
medium, the production rate of β-phenylalanine by the whole-cell biocatalyst (OD600 ~35) was
~0.062 g·L–1·h–1 between 0.5 – 1 h, and then slowly approached equilibrium at ~0.092 g·L–1 βphenylalanine between 6 – 8 h. β-Phenylalanine was produced in 50 mM phosphate buffer (pH 7
and 8) at an initial rate (~0.059 g·L–1·h–1) over 1 h, similar to that made by E. coli incubated in
M9 minimal medium (cf. Figure 5.15B). By contrast, between 2 – 8 h, β-phenylalanine
decreased linearly in phosphate buffer at ~0.004 g·L–1·h–1 (pH 7) and ~0.006 g·L–1·h–1 (pH 8),
and slowly increased in minimal medium (~0.005 g·L–1·h–1). α-Phenylalanine depleted faster in
50 mM phosphate buffer (0.024 g·L–1·h–1 and 0.028 g· L–1·h–1 at pH 7 and 8, respectively),
compared to when the bacteria was incubated in M9 minimal medium (0.012 g· L–1·h–1) (Figure
A.3.2). We also observed that α-phenylalanine disappeared from the phosphate buffer at a rate
greater than β-phenylalanine was exported into the medium over 8 h. This observation suggested
that the whole-cell biocatalyst metabolized the exogenous α-phenylalanine (and likely βphenylalanine) for cell growth and survival in the nutrient-deprived phosphate buffers. Thus, the
moderately fortified M9 minimal medium was used for other whole-cell PaPAM biocatalyst
assays.

221

In addition, to assess the cell viability in the different media, CFU·mL-1 were calculated
for bacteria incubated in M9 minimal medium (pH 7.4) and phosphate buffer (pH 7) containing
α-phenylalanine (1 mM). The CFU (4.1 × 1011 mL−1) of E. coli incubated for 4 h in minimal
medium was ~1.5-fold higher than the CFU (2.9 × 1011 mL−1) of cells incubated in phosphate
buffer (pH 7) (Table 5.2). E. coli viability in minimal medium did not change when αphenylalanine was omitted, whereas the CFU reduced 2-fold to 1.6 × 1011 mL−1 in phosphate
buffer without α-phenylalanine supplementation (Table 5.2). Thus, the exogenous amino acid
was likely only slightly mesotrophic for the bacteria in the phosphate medium, causing the cell
viability as well as the production rate of β-phenylalanine to decrease (cf. Figure 5.15B).
Table 5.2. Colony Forming Units (CFU) of E. coli in different reaction media.

Assay Medium
M9 minimal medium
Phosphate buffer
(pH 7, 50 mM)

α-Phenylalanine
(mM)
0
1
0
1

CFU (mL−1) × 10−11
4.2±0.33
4.1±0.49
1.6±0.17
2.9±0.38

5.3.6. The Effect of α-Phenylalanine Concentration on β-Phenylalanine Production
Theoretically, the production rates of catalytic systems can increase by providing higher
substrate concentrations.83 However, higher concentration of substrates is often deleterious to
microbial cells used in whole-cell biocatalytic systems.83 Thus, α-phenylalanine (1 – 25 mM)
was used as the benchmark for all the substrates tested to examine its effect on β-phenylalanine
production and E. coli whole-cell viability. β-Phenylalanine production in the whole-cells
increased 0.09 – 1.2 g·L–1 with increasing substrate concentration in 6-h incubations at 16 °C
(Figure 5.16A). The highest measured β-phenylalanine production (1.2 g·L–1) was at 25 mM
222

substrate. However, the percent conversion of α- to β-phenylalanine decreased from 55% at 1
mM substrate to a lower limit of 29% at 25 mM α-phenylalanine (Figure 5.16A). Mass transfer
likely reached its limit when the biocatalyst was limiting or the higher concentration of substrate
became toxic to the cells.

A

B
80

1.75

8

1.00
40
0.75

30

0.50

6

4

20
2

0.25
0.00

OD600 = 35

1.6

-Phenylalanine (g/L)

50

Conversion (%)

60
1.25

CFU (mL-1) x 10-11

1.50

-Phenylalanine (g/L)

10

70

OD600 = 280

1.2

0.8

0.4

10

0

5

10

15

20

25

0

0

0.0
0

-Phenylalanine (mM)

5

10

15

20

25

-Phenylalanine (mM)

Figure 5.16. Effect of increased substrate concentration and biocatalyst amount. A) The production
of β-phenylalanine in PaPAM whole-cell biocatalysis in M9 minimal medium (pH 7.4) containing
various concentrations of α-phenylalanine (■ on dashed line) at 16 C, percent conversion of α- to βphenylalanine (● on solid line), and CFU (●) at each susbtrate concentration. B) Comparison of PaPAM
whole-cell biocalyst concentration OD600 ~35 (■ on dashed line) and 280 (♦ on soild line) incubated at
various substrate concentrations.

Substrate toxicity was analyzed by incubating E. coli cells in increasing substrate
concentration. Calculating the CFU of cultures at increasing substrate concentration revealed that
the cells remain viable even at 25 mM α-phenylalanine (Figure 5.16A). When the amount E. coli
cells was increased from OD600 ~35 to ~70 and 100, the higher biomasses did not significantly
improve the β-phenylalanine production (data not shown). By comparison, at OD600 ~280 (an 8223

fold increase in cell biomass over OD600 ~35), the β-phenylalanine production increased by 33%
when the substrate concentration was between 15 – 25 mM (Figure 5.16B). This suggested
moderate scalability of the substrate for the whole-cell PaPAM biocatalyst.

5.3.7. Substrate Scope of the Whole-cell Biocatalytic System
Limited substrate scope is commonly a main disadvantage of enzyme biocatalysis.84
PaPAM, however, has a broad substrate scope in vitro,49 and thus was anticipated to perform
similarly in vivo. The PaPAM whole-cell biocatalyst was probed with 21 α-arylalanines,
including two heteroaromatic compounds. The productive and inactive substrates were similar to
those identified earlier for PaPAM in vitro.49 The biocatalyst did not isomerize o-chloro-, obromo- and o-nitro-α-phenylalanine (Table 5.3) in vivo; likely, the proximity of the orthosubstituent to the alanine side chain caused ineffective substrate binding.49 Five α-arylalanine
substrates (m-bromo (96%), p-chloro (93%), p-bromo (92%), 3-thienyl (92%), and m-methyl
(90%)) were converted efficiently to their β-amino acids. In general, the whole-cell PaPAM
biocatalyst preferentially isomerized meta- and para-substituted substrates over their orthosubstituted isomers; ortho-regioisomers (fluoro- (41%), methyl- (27%),and methoxy- (6%)) were
converted more poorly.
The rank order between the in vitro catalytic rates (kcat) and the production levels of the
whole-cell biocatalyst was redistributed for the substrates tested (Table 5.3). For example, the
intrinsic kcat values 0.484 and 0.420 s−1 of PaPAM in vitro for m-chloro (8) and m-bromo (1)
(ranked 1st and 2nd), respectively, were followed by the natural substrate α-phenylalanine (11)

224

(ranked 3rd) in an earlier study.49 By comparison, based on the substrate turnover by the wholecell biocatalyst, m-bromo (1) ranked 1st (235 mg·L−1), m-chloro (8) ranked 8th (128 mg·L−1), and
the natural substrate (11) ranked 11th (96 mg·L−1); interestingly, 11 was isomerized in vivo
slower than ten other substrates (Table 5.3). Other instances that highlight the differences in
turnover by PaPAM in whole cells and in vitro are with the p-chloro (3) and p-bromo (2)
substrates. The kcat values for 3 and 2 (0.053 and 0.045 s−1, respectively) ranked 9th and 10th in
vitro, while their turnovers unpredictably ranked 3rd and 2nd (186 and 224 mg·L−1 for 3 and 2,
respectively) in the whole-cells (Table 5.3). It is interesting that all fluoro-substituted
regioisomer substrates (5, 9, and 13) were isomerized among the slowest in vitro (~0.02 s−1),
over 10-fold slower than the natural substrate.49 However, the whole-cell biocatalyst converted
the same fluorinated substrates (ortho-13, 74 mg·L−1; para-5, 159 mg·L−1; and meta-9, 113
mg·L−1) at a level similar to several other substrates, including α-phenylalanine (Table 5.3).
In vitro biocatalysts are governed by cofactors (if any), physical conditions (temperature,
pH, ionic strength), and the intrinsic properties of the enzymes when reconstituted in buffer. In
vivo whole-cell biocatalysts are dependent however on factors such as mass transfer across cell
membranes, cellular metabolism, protein synthesis, stimulation and inactivation of microbial cell
growth, toxicity, and by-product formation.85 The relative transport rates of molecules across the
E. coli cell membrane can depend on steric and electronic properties of the substrate.60 The
aromatic amino acid:H+–symporter permeases (PheP and AroP) in the inner membrane of E. coli
transport phenylalanine display high affinity (KM = 2 μM and 0.5 μM, respectively) towards the
natural substrate.86 The AroP permease has similar affinity (KM = 0.6 μM) for the phenylalanine
analogue tyrosine and the bulkier heteroaromatic tryptophan,87 whereas PheP is more specific for
phenylalanine.88
225

Thus, we expect that AroP is the principal transporter for the α-arylalanines used in this
whole-cell feeding study. This hypothesis is supported by an earlier inhibition study of the
general aromatic amino acid transporter (AroP) by p-fluoro-α-phenylalanine and 3-(2-thienyl)-αalanine,89 also used as substrates herein. These analogues competitively inhibited the transporter
as strongly as natural aromatic amino acids, suggesting that at least these two have equal affinity
for AroP.89 In another study, arylalanines with substituents on the ring (not specified) were also
reported as being good substrates for AroP, yet had lower affinity (data not provided) for the
permease than the natural substrate.90 We hypothesize the various arylalanines (1 mM) in the
medium, likely bind the permeases with variable affinities and thus transfer through the E. coli
inner membrane at different maximum velocities. In addition, α-phenylalanine is also a primary
metabolite of protein synthesis and likely (including other unnatural arylalanines) is a precursor
of catabolic phenylpropanoid degradation. Thus, several combined physiological causes are
expected to have affected whole-cell biocatalyst when making each β-arylalanine. As a result,
the rank-order correlation was scrambled between the in vivo production and the kcat values of
PaPAM for the substrates (Table 5.3).

226

Table 5.3. Production levels and conversions of PaPAM whole-cell biocatalytic system for various
α-arylalanine substrates.

Product

Production
(mg·L−1)

Yield
(%)

235±3.6

Ranking*
Product

Production
(mg·L−1)

Yield
(%)

In
vivo

In
vitro

96

1

2

109±9.2

224±5.8

92

2

10

186±4.2

93

3

162±7.5

90

159±0.84

Ranking*
In
vivo

In
vitro

56

10

16

96±2.1

60

11

3

9

77±2.4

43

12

17

4

8

74±3.6

41

13

15

87

5

14

68±2.2

40

14

13

157±4.6

92

6

6

49±3.2

27

15

7

140±4.3

72

7

4

17±1.4

8

16

5

128±6.1

64

8

1

13±0.77

6

17

18

113±1.5

62

9

12

8.5±1.1

4

18

11

*Ranking order is based on the β-phenylalanine production of the in vivo system and the catalytic rate (kcat) of the in
vitro system.

5.3.8. Sustainability of the PaPAM Whole-cell Biocatalytic System
Recycling of recombinant cells is an important factor in developing a sustainable
industrial process for producing fine chemicals.78 To test the operational stability of the
227

recombinant E. coli whole-cell biocatalyst towards recycling, a set of consecutive biocatalytic
cycles was performed at 16 °C for 30 h using a single sample of the biocatalyst. After each 6-h
reaction cycle, the cells were recovered by centrifugation, washed with M9 minimal medium to
remove residual substrate and product, and to reduce cell aggregates (at OD600 ~35). The cells
were recovered by centrifugation and resuspended in M9 minimal medium before the next
reaction in the series. β-Phenylalanine production remained almost constant for whole-cell
biocatalyst fed α-phenylalanine at 1 mM and 5 mM (0.075 and 0.9 g·L–1, respectively) in each of
the five batch reaction cycles (Figure 5.17). CFU measurements suggested continued cell

1.50

6

1.25

5

1.00

4

0.75

3

0.50

2

0.25

1

0.00

CFU (mL-1) x 10-11

-Phenylalanine (g/L)

viability even after centrifugation and cell resuspension during each cycle (Figure 5.17).

0
1

2

3

4

5

Number of Cycles

Figure 5.17. Recycling of the E. coli whole-cell biocatalyst (OD600 ~35) for β-phenylalanine
production. The production of β-phenylalanine in PaPAM whole-cell biocatalysis in M9 minimal
medium (pH 7.4) incubated with α-phenylalanine (1 mM (blue bars) and 5 mM (white bars)) at 16 C for
each catalytic cycle. CFU values () for the bacterial cells after each reaction cycle in M9 minimal
medium (pH 7.4) containing α-phenylalanine (5 mM) at 16 C are shown.

228

Using sustainable whole-cell biocatalysts is attractive over biocatalytic routes employing
purified enzymes for in vitro assays, in part, because of the potential higher risk for enzyme
denaturation during purification and incubation with the substrate. For example, a purified
Amano lipase PS was used in vitro for a commercial (Evonik-Degussa, Hanau-Wolfgang,
Germany) biocatalytic process to resolve racemic β-aryl-β-amino acids.32 The enzyme is
inactivated partially after one 15-h batch reaction cycle, yielding a paltry 10-11% conversion in
the second batch reaction. To compensate for the loss of activity after each of five biocatalytic
cycles, fresh enzyme was added to a fraction of the previous batch reaction.32 The whole-cell
PaPAM biocatalyst, described herein, however, can likely produce several (3S)-β-aryl-β-amino
acids at a sustainable concentration over multiple, recycled batches of biocatalyst (shown by
using β-phenylalanine as the model compound) (Figure 5.17).

229

5.4. Conclusion

A PaPAM whole-cell biocatalyst was shown to produce several unnatural (3S)-β-aryl-βamino acids at >99.9% ee, with the highest turnover rate in M9 minimal medium at 16 °C. The
whole-cell biocatalyst biosynthesized 18 β-arylalanines with moderate to excellent converted
yield (4-96%) at production levels of 8.5 – 235 mg·L–1 over 6 h, respectively. More notably, E.
coli cells are reusable over at least five reaction cycles without a noticeable loss in activity and
cell viability. This biocatalyst offers notable advantages over conventional synthetic methods
because of its excellent enantioselectivity, broad substrate scope, single-step conversion, and
sustainability. In addition, the small-scale production yields of the whole-cell biocatalyst can
likely improve by using a bioreactor, increasing the number of bacterial membrane permeases,
optimizing the outer membrane permeability,59 and reducing aromatic amino acid flux through
catabolic pathways in engineered E. coli.

230

APPENDIX

231

0.10

Mass (g/L)

0.08

0.06

0.04

0.02

0.00
0

2

4

6

8

Time (h)

10

Figure A.3.1. Comparison of product accumulation and substrate depletion over time. βPhenylalanine (■ on dashed line) and cinnamic acid ( on solid line) production, and α- phenylalanine (▲
on short dashed line) depletion by engineered E. coli (OD600 ~35) incubated at 16 C in M9 minimal
medium (pH 7.4) containing α- phenylalanine (1 mM).

-Phenylalanine (g/L)

0.05
0.04
0.03
0.02
0.01
0.00

0

2

4

6

8

Time (h)

Figure A.3.2. Comparison of substrate depletion in different reaction media. Amount of αphenylalanine remaining in M9 minimal medium (pH 7.4, ▲ on short dashed line), phosphate buffer at
pH 7 (● on solid line), and pH 8 (■ on dashed line) incubated with engineered E. coli (OD600 ~ 35) at 16
C. Each medium was supplied with α-phenylalanine (1 mM).
232

REFERENCES

233

REFERENCES
(1)

Liljeblad, A.; Kanerva, L. T. Tetrahedron 2006, 62, 5831-5854.

(2)

Seebach, D.; Matthews, J. L. Chem. Commun. 1997, 2015-2022.

(3)

Jin, M.; Fischbach, M. A.; Clardy, J. J. Am. Chem. Soc. 2006, 128, 10660-10661.

(4)

Crews, P.; Manes, L. V.; Boehler, M. Tetrahedron Lett. 1986, 27, 2797-2800.

(5)

Nishizawa, R.; Saino, T.; Takita, T.; Suda, H.; Aoyagi, T.; Umezawa, H. J. Med. Chem.
1977, 20, 510-515.

(6)

Savile, C. K.; Janey, J. M.; Mundorff, E. C.; Moore, J. C.; Tam, S.; Jarvis, W. R.; Colbeck,
J. C.; Krebber, A.; Fleitz, F. J.; Brands, J.; Devine, P. N.; Huisman, G. W.; Hughes, G. J.
Science 2010, 329, 305-309.

(7)

Nicolaou, K. C.; Yang, Z.; Liu, J. J.; Ueno, H.; Nantermet, P. G.; Guy, R. K.; Claiborne, C.
F.; Renaud, J.; Couladouros, E. A.; Paulvannan, K.; Sorensen, E. J. Nature 1994, 367, 630634.

(8)

Haycock-Lewandowski, S. J.; Wilder, A.; Ahman, J. Org. Process Res. Dev. 2008, 12,
1094-1103.

(9)

Seebach, D.; Overhand, M.; Kiihnle, F. N. M.; Martinoni, B. Helv. Chim. Acta 1996, 79,
913-941.

(10) Frackenpohl, J.; Arvidsson, P. I.; Schreiber, J. V.; Seebach, D. ChemBioChem 2001, 2,
445-455.
(11) Seebach, D.; Abele, S.; Gademann, K.; Bernhard, J. Angew. Chem. Int. Ed. 1999, 38, 15951597.
(12) Seebach, D.; Abele, S.; Gademann, K.; Guichard, G.; Hintermann, T.; Jaun, B.; Matthews,
J. L.; Schreiber, J. V. Helv. Chim. Acta 1998, 81, 932-982.
(13) Gademann, K.; Ernst, M.; Hoyer, D.; Seebach, D. Angew. Chem. Int. Ed. 1999, 38, 12231226.
(14) Liu, R. H.; Chen, X. Y.; Hayouka, Z.; Chakraborty, S.; Falk, S. P.; Weisblum, B.; Masters,
K. S.; Gellman, S. H. J. Am. Chem. Soc. 2013, 135, 5270-5273.
(15) Liu, M.; Sibi, M. P. Tetrahedron 2002, 58, 7991-8035.
(16) Lelais, G.; Seebach, D. Biopolymers 2004, 76, 206-243.
234

(17) Weiner, B.; Szymanski, W.; Janssen, D. B.; Minnaard, A. J.; Feringa, B. L. Chem. Soc.
Rev. 2010, 39, 1656-1691.
(18) Yuan, C. G.; Williams, R. M. J. Am. Chem. Soc. 1997, 119, 11777-11784.
(19) Pinho, V. D.; Gutmann, B.; Kappe, C. O. RSC Adv. 2014, 4, 37419-37422.
(20) Sewald, N. Amino Acids 1996, 11, 397-408.
(21) Davies, S. G.; Ichihara, O. Tetrahedron-Asymmetr 1991, 2, 183-186.
(22) Davies, S. G.; Foster, E. M.; McIntosh, C. R.; Roberts, P. M.; Rosser, T. E.; Smith, A. D.;
Thomson, J. E. Tetrahedron-Asymmetr 2011, 22, 1035-1050.
(23) Costello, J. F.; Davies, S. G.; Ichihara, O. Tetrahedron-Asymmetr 1994, 5, 1999-2008.
(24) Dumas, F.; Mezrhab, B.; d’Angelo, J. J. Org. Chem. 1996, 61, 2293-2304.
(25) Matsuyama, H.; Itoh, N.; Yoshida, M.; Kamigata, N.; Sasaki, S.; Iyoda, M. Chem. Lett.
1997, 375-376.
(26) Falborg, L.; Jorgensen, K. A. J. Chem. Soc., Perkin Trans. 1 1996, 2823-2826.
(27) Sibi, M. P.; Liu, M. Org. Lett. 2000, 2, 3393-3396.
(28) Lubell, W. D.; Kitamura, M.; Noyori, R. Tetrahedron-Asymmetr 1991, 2, 543-554.
(29) Zhang, W. C.; Chi, Y. X.; Zhang, X. M. Acc. Chem. Res. 2007, 40, 1278-1290.
(30) Holz, J.; Monsees, A.; Jiao, H. J.; You, J. S.; Komarov, I. V.; Fischer, C.; Drauz, K.;
Borner, A. J. Org. Chem. 2003, 68, 1701-1707.
(31) Kadyrov, R.; Ilaldinov, I. Z.; Almena, J.; Monsees, A.; Riermeier, T. H. Tetrahedron Lett.
2005, 46, 7397-7400.
(32) Grayson, J. I.; Roos, J.; Osswald, S. Org. Process Res. Dev. 2011, 15, 1201-1206.
(33) de Carvalho, C. C. Biotechnol. Adv. 2011, 29, 75-83.
(34) Arnold, F. H. Nature 2001, 409, 253-257.
(35) Schmid, A.; Dordick, J. S.; Hauer, B.; Kiener, A.; Wubbolts, M.; Witholt, B. Nature 2001,
409, 258-268.
(36) Faulconbridge, S. J.; Holt, K. E.; Sevillano, L. G.; Lock, C. J.; Tiffin, P. D.; Tremayne, N.;
Winter, S. Tetrahedron Lett. 2000, 41, 2679-2681.
235

(37) Rudat, J.; Brucher, B. R.; Syldatk, C. AMB Express 2012, 2, 11.
(38) Walker, K. D.; Klettke, K.; Akiyama, T.; Croteau, R. J. Biol. Chem. 2004, 279, 5394753954.
(39) Magarvey, N. A.; Fortin, P. D.; Thomas, P. M.; Kelleher, N. L.; Walsh, C. T. ACS Chem.
Biol. 2008, 3, 542-554.
(40) Soloshonok, V. A.; Fokina, N. A.; Antonyna V. Rybakova, A. V.; Shishkina, I. P.;
Galushko, S. V.; Sorochinsky, A. E.; Kukhar, V. P. Tetrahedron-Asymmetr 1995, 6, 16011610.
(41) Behshad, E.; Ruzicka, F. J.; Mansoorabadi, S. O.; Chen, D.; Reed, G. H.; Frey, P. A.
Biochemistry 2006, 45, 12639-12646.
(42) Ruzicka, F. J.; Frey, P. A. Biochim. Biophys. Acta 2007, 1774, 286-296.
(43) Rachid, S.; Krug, D.; Weissman, K. J.; Muller, R. J. Biol. Chem. 2007, 282, 21810-21817.
(44) Christenson, S. D.; Wu, W.; Spies, M. A.; Shen, B.; Toney, M. D. Biochemistry 2003, 42,
12708-12718.
(45) Chirpich, T. P.; Zappia, V.; Costilow, R. N.; Barker, H. A. J. Biol. Chem. 1970, 245, 17781789.
(46) Walker, K. D.; Klettke, K.; Akiyama, T.; Croteau, R. J. Biol. Chem. 2004, 279, 5394753954.
(47) Christenson, S. D.; Liu, W.; Toney, M. D.; Shen, B. J. Am. Chem. Soc. 2003, 125, 60626063.
(48) Klettke, K. L.; Sanyal, S.; Mutatu, W.; Walker, K. D. J. Am. Chem. Soc. 2007, 129, 69886989.
(49) Ratnayake, N. D.; Liu, N.; Kuhn, L. A.; Walker, K. D. ACS Catal. 2014, 4, 3077-3090.
(50) Wu, B.; Szymanski, W.; Wietzes, P.; de Wildeman, S.; Poelarends, G. J.; Feringa, B. L.;
Janssen, D. B. ChemBioChem 2009, 10, 338-344.
(51) Szymanski, W.; Wu, B.; Weiner, B.; de Wildeman, S.; Feringa, B. L.; Janssen, D. B. J.
Org. Chem. 2009, 74, 9152-9157.
(52) Wu, B.; Szymanski, W.; Wybenga, G. G.; Heberling, M. M.; Bartsch, S.; de Wildeman, S.;
Poelarends, G. J.; Feringa, B. L.; Dijkstra, B. W.; Janssen, D. B. Angew. Chem. Int. Ed.
2012, 51, 482-486.

236

(53) Wanninayake, U.; DePorre, Y.; Ondari, M.; Walker, K. D. Biochemistry 2011, 50, 1008210090.
(54) Nikolova, P.; Ward, O. P. J. Ind. Microbiol. 1993, 12, 76-86.
(55) May, O.; Nguyen, P. T.; Arnold, F. H. Nat. Biotechnol. 2000, 18, 317-320.
(56) Breuer, M.; Ditrich, K.; Habicher, T.; Hauer, B.; Kesseler, M.; Sturmer, R.; Zelinski, T.
Angew. Chem. Int. Ed. 2004, 43, 788-824.
(57) Groger, H.; Chamouleau, F.; Orologas, N.; Rollmann, C.; Drauz, K.; Hummel, W.;
Weckbecker, A.; May, O. Angew. Chem. Int. Ed. 2006, 45, 5677-5681.
(58) Chen, R. R. Z. Appl. Microbiol. Biotechnol. 2007, 74, 730-738.
(59) Ni, Y.; Chen, R. R. Biotechnol. Bioeng. 2003, 87, 804-811.
(60) Silhavy, T. J.; Kahne, D.; Walker, S. Cold Spring Harb. Perspect. Biol. 2010, 2, 1-16.
(61) Nikaido, H. Mol. Microbiol. 1992, 6, 435-442.
(62) Liu, Y.; Hama, H.; Fujita, Y.; Kondo, A.; Inoue, Y.; Kimura, A.; Fukuda, H. Biotechnol.
Bioeng. 1999, 64, 54-60.
(63) Vanderwerf, M. J.; Hartmans, S.; Vandentweel, W. J. J. Appl. Microbiol. Biotechnol. 1995,
43, 590-594.
(64) Canovas, M.; Torroglosa, T.; Kleber, H. P.; Iborra, J. L. J. Basic Microbiol. 2003, 43, 259268.
(65) Matsumoto, T.; Takahashi, S.; Kaieda, M.; Ueda, M.; Tanaka, A.; Fukuda, H.; Kondo, A.
Appl. Microbiol. Biotechnol. 2001, 57, 515-520.
(66) Yang, R. Y.; Bayraktar, O.; Pu, H. T. J. Biotechnol. 2003, 100, 13-22.
(67) Schuurmann, J.; Quehl, P.; Festel, G.; Jose, J. Appl. Microbiol. Biotechnol. 2014, 98, 80318046.
(68) Becker, U.; Doderer, K.; Osswald, S.; Verseck, S.; Wienand, W. In Pharm. Technol.
www.pharmtech.com, 2008.
(69) Schwarm, M. In Application of Whole-Cell Biocatalysis in the Manufacture of Fine
Chemicals; Shioiri, T., Izawa, K., Konoike, T., Eds.; Wiley-VCH Verlag GmbH & Co:
Weinheim, 2011, p 183-205.

237

(70) Strom, S.; Wanninayake, U.; Ratnayake, N. D.; Walker, K. D.; Geiger, J. H. Angew. Chem.
Int. Ed. 2012, 51, 2898-2902.
(71) Piperno, J. R.; Oxender, D. L. J. Biol. Chem. 1968, 243, 5914-5920.
(72) Chye, M. L.; Guest, J. R.; Pittard, J. J. Bacteriol. 1986, 167, 749-753.
(73) Whipp, M. J.; Pittard, A. J. J. Bacteriol. 1977, 132, 453-461.
(74) Pi, J.; Pittard, A. J. J. Bacteriol. 1996, 178, 2650-2655.
(75) Koyanagi, T.; Katayama, T.; Suzuki, H.; Kumagai, H. J. Bacteriol. 2003, 186, 343-350.
(76) Petrovic, S. D.; Misic-Vukovic, M. M.; Mijin, D. Z. Chem. Indust. 2002, 1, 10-16.
(77) Ratnayake, N. D.; Wanninayake, U.; Geiger, J. H.; Walker, K. D. J. Am. Chem. Soc. 2011,
133, 8531-8533.
(78) Li, Q. S.; Li, G. Q.; Ma, F. Q.; Zhang, Z. M.; Zheng, B. S.; Feng, Y. Process Biochem.
2011, 46, 477-481.
(79) Gadgil, M.; Kapur, V.; Hu, W. S. Biotechnol. Progr. 2005, 21, 689-699.
(80) Kolter, R.; Siegele, D. A.; Tormo, A. Annu. Rev. Microbiol. 1993, 47, 855-874.
(81) Teufel, R.; Mascaraque, V.; Ismail, W.; Voss, M.; Perera, J.; Eisenreich, W.; Haehnel, W.;
Fuchs, G. Proc. Natl. Acad. Sci. USA 2010, 107, 14390-14395.
(82) Fortin, P. D.; Walsh, C. T.; Magarvey, N. A. Nature 2007, 448, 824-811.
(83) van Bloois, E.; Dudek, H. M.; Duetz, W. A.; Fraaije, M. W. BMC Biotechnol. 2012, 12.
(84) Reetz, M. T. J. Am. Chem. Soc. 2013, 135, 12480-12496.
(85) Schrewe, M.; Julsing, M. K.; Buhler, B.; Schmid, A. Chem. Soc. Rev. 2013, 42, 6346-6377.
(86) Haney, S. A.; Oxender, D. L. Int. Rev. Cytol. 1992, 137, 37-95.
(87) Brown, K. D. J. Bacteriol. 1970, 104, 177-188.
(88) Pi, J.; Wookey, P. J.; Pittard, A. J. J. Bacteriol. 1991, 173, 3622-3629.
(89) Brown, K. D. J. Bacteriol. 1971, 106, 70-81.
(90) Willshaw, G.; Tristram, H. Biochem. J. 1972, 127, 71.

238