THERMOTOGA MARITIMA GLYCEROL DEHYDROGENASE AS A CATALYST FOR
DIHYDROXYACETONE PRODUCTION: ENZYME CHARACTERIZATION,
ENGINEERING AND COFACTOR IMMOBILIZATION
By
Justin Liam Beauchamp

A DISSERTATION
Submitted to
Michigan State University
In partial fulfilment of the requirements
For the degree of
Cell and Molecular Biology—Doctor of Philosophy
2015

ABSTRACT
THERMOTOGA MARITIMA GLYCEROL DEHYDROGENASE AS A CATALYST FOR
DIHYDROXYACETONE PRODUCTION: ENZYME CHARACTERIZATION,
ENGINEERING AND COFACTOR IMMOBILIZATION
By
Justin Liam Beauchamp
NAD-dependent dehydrogenases facilitate a diverse range of hydride transfer reactions in
many areas of cellular metabolism. Dehydrogenases catalyze about 12% of metabolic reactions
and are classified into 4 different groups: the medium-chain dehydrogenase superfamily, the
short-chain dehydrogenase superfamily, the long-chain dehydrogenases, and the family III metaldependent polyol dehydrogenases. The broad range of reactions catalyzed makes them attractive
catalysts for synthetic process, especially since many produce chiral products. However, their
use as catalysts has been limited by the requirement that the NAD cofactor be supplied in
stoichiometric amounts. Enzymatic cofactor regeneration has been used for synthesis of high
value products, but is not applicable to many processes because providing an extra enzyme
increases costs. Electrochemical cofactor regeneration could be a lower cost alternative to
enzymatic regeneration. Our overall goal is to immobilize Thermotoga maritima glycerol
dehydrogenase (TmGlyDH) and its NAD cofactor onto an electrode for catalytic production of
dihydroxyacetone (DHA) from glycerol where the electrode regenerates NAD+. Glycerol is a
waste product of the biodiesel industry, while DHA is a more valuable synthetic precursor and
sunless tanning agent.
TmGlyDH’s kinetics, stability and activity were characterized to provide acceptable
operating conditions for an electrochemical reactor. TmGlyDH showed cooperative rather than
Michaelis-Menten kinetics with glycerol and DHA, so the Hill equation was used to determine

limiting rate (Vmax), half saturating substrate concentration (K0.5) and Hill coefficient (n). The
optimum pH for glycerol oxidation was 7.9. We tested alternative substrates similar to glycerol
and TmGlyDH was able to produce 1,2-propanediol from hydroxyacetone at greater than 99%
enantiomeric excess. To test if TmGlyDH can use immobilized NAD, the NAD analogue N6carboxymethyl-NAD (N6-CM-NAD) was synthesized and immobilized on amino-linkermodified sepharose beads (NAD-sepharose). TmGlyDH had low activity with N6-CM-NAD and
mutants were produced to increase activity with the NAD analog 10-fold, but both TmGlyDH
mutants and wild-type showed similar activity with NAD-sepharose. The length of the linker
between NAD and sepharose had no effect on coenzymic activity. To see if other
dehydrogenases can use N6-immobilized NAD, we tested 6 different dehydrogenases and 5 of
the 6 used NAD-sepharose as a cofactor; structural analysis of the enzymes binding pockets
predicted activity with soluble N6-carboxymethyl-NAD, but not with NAD-sepharose.
In a catalytic system, the longevity of the catalyst is important because renewing the
catalyst is costly. We observed that TmGlyDH gets inactivated by its product, DHA. TmGlyDH
inactivation by DHA is a result of the enzyme getting modified by Maillard reactions between
the Lys and Arg residues and DHA. We identified which Lys and Arg residues get modified by
DHA and prepared mutants to improve stability in the presence of DHA. The mutant most stable
to DHA at 50°C was K361Q, maintaining activity twice as long as the wild-type enzyme in the
presence of DHA at 50°C.
In summary, TmGlyDH has been characterized and engineered towards use in a
bioelectronic catalytic system. Our immobilization method of NAD is suitable for use by a wide
variety of dehydrogenases, indicating a working bioelectronic system would be adaptable to
many novel applications.

Dedicated to the memory of my father, Michael Beauchamp. He
was thrilled that I chose to go to graduate school, always
excited to hear about my progress and would have loved to see
me finish.

iv

ACKNOWLEDGEMENTS

I would like to thank all former and current members of Dr. Vieille’s lab. In particular, I
would like to thank former postdoc Hoon Soon Song for mentoring me during my first couple
years in the lab and for teaching me many of the lab skills I still use today. I would like to thank
former postdocs (Dr. JJ Park and Dr. Sanhana Lal) and current and former graduate students (Dr.
Bryan Schindler, Nik McPherson, and Rajasi Joshi) for being helpful and supportive lab
members. I would like to thank Maris Laivenieks for sharing his considerable experience and
knowledge with me and helping with many technical aspects of my project. I thank Ian
Bloomquist for cloning the TmGlyDH gene prior to my joining the lab. I would like to thank the
undergraduate researchers that worked with me on my project (Phillip Gross, Christopher Blum,
and Mohamed Esa). Especially, I would like to thank Phillip Gross for his countless hours of
work and dedication to the project.
I would like to thank Dr. R. Mark Worden, Dr. Scott Calabrese Barton, Dr. Brian Hassler,
Dr. Hanzi Li, Dr. Rui Li and Dr. Chloe Liu, my collaborators in chemical engineering, for many
constructive conversations, suggestions and insight into my project. I would like to thank Dr.
Dan Jones, Dr. Lijun Chen and Dr. Scott Smith, members of MSU RTSF Mass Spectrometry
Facility for their technical help with my project. I would like to thank Dr. Michael Feig and Dr.
Srinivasa Gopal for help with the structural protein modeling aspects of my project. I would like
to thank Arving Jagnathan for helping me get started with the NAD analog synthesis.
I would like to thank my committee members. Dr. R. Mark Worden, Dr. Eric Hegg, Dr.
Shelagh Ferguson-Miller and Dr. Kevin Walker for their time and constructive guidance.

v

I would like to thank my friends and family for being there when needed. I thank my
wife, Sami for her unwavering support and encouragement. I am grateful for the friends I have
made while here at MSU; I think of them as part of my extended family.
Lastly I thank my advisor Dr. Claire Vieille. I am honored to have had her as my advisor.
She has always been available when needed and has been very supportive and helpful. I
appreciate the many hours she has spent editing my writing and helping me improve as a writer.
Her insight and guidance has helped me to become a better scientist.

vi

TABLE OF CONTENTS
LIST OF TABLES .......................................................................................................................... x
LIST OF FIGURES ....................................................................................................................... xi
KEY TO ABBREVIATIONS ...................................................................................................... xiii
CHAPTER 1. Introduction – Dehydrogenases as catalysts ............................................................ 1
1.1 INTRODUCTION ................................................................................................................ 2
1.2 DEHYDROGENASE DIVERSITY AND FAMILIES ........................................................ 4
1.2.1 The long-chain dehydrogenases ..................................................................................... 5
1.2.2 The medium-chain dehydrogenases ............................................................................... 5
1.2.3 The short-chain dehydrogenases .................................................................................... 6
1.2.4 The family III polyol dehydrogenases ........................................................................... 6
1.3 GLYCEROL DEHYDROGENASES................................................................................... 6
1.4 DHA PRODUCTION AND APPLICATIONS .................................................................... 8
1.5 ENZYMATIC COFACTOR REGENERATION METHODS ............................................ 9
1.5.1 Formate dehydrogenase ................................................................................................. 9
1.5.2 Phosphite dehydrogenase ............................................................................................. 10
1.5.3 Regeneration of oxidized cofactors.............................................................................. 11
1.6 ELECTROCHEMICAL COFACTOR REGENERATION ............................................... 12
1.6.1 Direct cofactor regeneration......................................................................................... 12
1.6.2 Mediated cofactor regeneration ................................................................................... 13
1.7 MODIFYING ELECTRODES FOR IMMOBILIZATION OF MEDIATOR AND
COFACTOR ............................................................................................................................. 15
1.8 COFACTOR IMMOBILIZATION .................................................................................... 18
1.8.1 Modification of NAD’s N1 and N6-amines .................................................................. 18
1.8.2 Immobilized NAD in affinity chromatography ........................................................... 22
1.8.3 Direct coupling of NAD ............................................................................................... 23
1.9 ENZYME IMMOBILIZATION ......................................................................................... 24
1.9.1 Immobilization of glycerol dehydrogenase.................................................................. 25
1.10 INTRODUCTION TO PROJECT .................................................................................... 26
REFERENCES ......................................................................................................................... 27
CHAPTER 2. Characterization of Thermotoga maritima glycerol dehydrogenase for use in the
production of dihydroxyacetone in a bioelectronics system ......................................................... 34
2.1 ABSTRACT........................................................................................................................ 36
2.1 INTRODUCTION .............................................................................................................. 36
2.3 MATERIALS AND METHODS........................................................................................ 40
2.3.1 Chemicals, bacterial strains, and culture conditions .................................................... 40
2.3.2 Cloning of the gene encoding TmGlyDH .................................................................... 40
2.3.3 Expression and purification of TmGlyDH ................................................................... 40
2.3.4 TmGlyDH activity assays ............................................................................................ 41
2.3.5 Determination of TmGlyDH’s apparent kinetic parameters ........................................ 43
2.3.6 N6-CM-NAD+ synthesis and purification .................................................................... 44
vii

2.3.7 Site-directed mutagenesis of TmGlyDH ...................................................................... 45
2.3.8 Determination of 1,2-propanediol enantiomeric excess .............................................. 45
2.3.9 DHA stability assays .................................................................................................... 46
2.3.10 Activity assays with N6-CM-NAD+ immobilized on Sepharose beads ..................... 46
2.4 RESULTS ........................................................................................................................... 47
2.4.1 Expression, purification, and characterization of TmGlyDH ...................................... 47
2.4.2 Effect of temperature on TmGlyDH stability and activity .......................................... 49
2.4.3 TmGlyDH apparent kinetic parameters ....................................................................... 51
2.4.4 Introduction of NAD+ into the 3D structure of TmGlyDH .......................................... 55
2.4.5 Kinetics of TmGlyDH on N6-CM-NAD+ .................................................................... 56
2.4.6 Activity of TmGlyDH with N6-CM-NAD+ immobilized on Sepharose beads ............ 60
2.5 DISCUSSION ..................................................................................................................... 60
2.6 ACKNOWLEDGEMENTS ................................................................................................ 65
2.7 SUPLEMENTARY MATERIALS..................................................................................... 66
2.8 PERMISSIONS .................................................................................................................. 70
REFERENCES ......................................................................................................................... 78
CHAPTER 3. Activity of select dehydrogenases with sepharose-immobilized N6-CM-NAD .... 83
3.1 ABSTRACT........................................................................................................................ 85
3.2 INTRODUCTION .............................................................................................................. 85
3.3 RESULTS AND DISCUSSION ......................................................................................... 87
3.4 CONCLUSION ................................................................................................................... 92
3.5 ACKNOWLEDGEMENTS ................................................................................................ 93
REFERENCES ......................................................................................................................... 94
CHAPTER 4. Inactivation of Thermotoga maritima glycerol dehydrogenase by its product,
dihydroxyacetone .......................................................................................................................... 97
4.1 ABSTRACT........................................................................................................................ 98
4.2 INTRODUCTION .............................................................................................................. 98
4.3 MATERIALS AND METHODS...................................................................................... 100
4.3.1 Materials .................................................................................................................... 100
4.3.2 Enzyme activity assays .............................................................................................. 100
4.3.3 Two-hour enzyme inactivation assays ....................................................................... 101
4.3.4 Time course enzyme inactivation assays ................................................................... 101
4.3.5 Trypsin and chymotrysin digests of TmGlyDH......................................................... 102
4.3.6 Identification of peptides by electrospray ionization time-of-flight mass spectrometry
(ESI)-TOF ........................................................................................................................... 103
4.3.7 Identification of the TmGlyDH residues modified by DHA ..................................... 104
4.4 RESULTS ......................................................................................................................... 104
4.4.1 TmGlyDH inactivation by DHA at 50°C................................................................... 104
4.4.2 Identification of the TmGlyDH residues modified by DHA ..................................... 105
4.4.3 Screening TmGlyDH mutants for increased stability to DHA .................................. 108
4.4.4 Time course analysis of mutants’ stability to DHA ................................................... 110
4.5 DISCUSSION ................................................................................................................... 115
4.5.1 The importance of TmGlyDH stability ...................................................................... 115
4.5.2 TmGlyDH inactivation by DHA is a complex process ............................................. 115
4.5.3 DHA modification sites ............................................................................................. 116
viii

4.5.4 Stability and biphasic behavior in 50°C inactivation time-courses ........................... 118
4.6 CONCLUSION ................................................................................................................. 120
4.7 SUPPLEMENTARY MATERIALS ................................................................................ 121
REFERENCES ....................................................................................................................... 123
CHAPTER 5. Conclusions and future work ............................................................................... 126
5.1 REVIEW OF PROJECT ................................................................................................... 127
5.1.1 Review of TmGlyDH characterization ...................................................................... 127
5.1.2 Suitability of TmGlyDH for use in bioelectrocatalytic system ................................. 128
5.1.3 Suitability of N6-immobilized NAD for use in bioelectrocatalytic systems .............. 129
5.2 FUTURE WORK .............................................................................................................. 130
5.2.1 Immobilizing NAD on an electrode ........................................................................... 130
5.2.2 Engineering of TmGlyDH ......................................................................................... 131
5.2.3 Adding TmGlyDH to the bioelectronic interface ...................................................... 132
5.3 ADDITIONAL USES OF THE BIOELECTRONIC INTERFACE ................................ 132
REFERENCES ....................................................................................................................... 134

ix

LIST OF TABLES

Table 2.1: TmGlyDH’s oxidizing activity on glycerol-related substrates at 80°C, pH 7.4. ......... 48
Table 2.2: Apparent kinetic parameters of TmGlyDH for NADH and NAD+ at 50°C. ................ 53
Table 2.3: Apparent kinetic parameters of TmGlyDH for DHA and glycerol at 50°C based on the
Hill equation (Eq. 1). .................................................................................................... 54
Table 2.4: Apparent kinetic parameters of TmGlyDH (WT) and its mutant derivatives on NAD+
and N6-CM-NAD+. ....................................................................................................... 58
Table S2.1: Primers used for Chapter 2. ....................................................................................... 67
Table S2.2: Relative effects of different buffers on TmGlyDH activity. ...................................... 67
Table S2.3: pH optima for different glycerol dehydrogenases. .................................................... 68
Table 3.1: Enzymes used in this study. ......................................................................................... 87
Table 3.2: Specific activity of the selected enzymes with 100 μM NAD and N6-CM-NAD and
enzyme ranking by polarity/charge and openness of the area surrounding NAD’s N6amine. ........................................................................................................................... 90
Table 4.1: TmGlyDH residues found to be modified by DHA. .................................................. 106
Table S4.1: Conditions used to detect peptides containing all Lys and Arg residues. ................ 122

x

LIST OF FIGURES

Figure 1.1: Two-chamber bioelectrocatalytic system for DHA and mannitol enzymatic
production. .................................................................................................................... 3
Figure 1.2: A single-chamber flow-through reactor for DHA enzymatic production from glycerol.
....................................................................................................................................... 3
Figure 1.3: Electrochemical reduction of NAD+. ......................................................................... 13
Figure 1.4: Two possible NAD binding orientations on a boronate linker. .................................. 16
Figure 1.5: NAD+ structure. .......................................................................................................... 20
Figure 2.1: TmGlyDH interconverts glycerol and DHA with the NAD(H) cofactor. .................. 37
Figure 2.2: Inactivation of TmGlyDH at 50°C (A) and 80°C (B). ............................................... 50
Figure 2.3: Effect of temperature on TmGlyDH’s DHA reduction activity at pH 7.4.................. 51
Figure 2.4: TmGlyDH’s glycerol oxidizing activity (A, B and C) and DHA reducing activity (D,
E and F) at 50°C, pH 7.4. ............................................................................................ 55
Figure 2.5: (A) Ribbon representation of alignment of BsGlyDH (Blue) and minimized
TmGlyDH (green), (B) active site residues around NAD .......................................... 57
Figure 2.6: Activity of TmGlyDH with N6-CM-NAD+ immobilized on B3APA-Sepharose. ...... 61
Figure S2.1: Structure of glycerol and Tris ................................................................................... 68
Figures S2.2: TmGlyDH activity for DHA reduction in the presence of increasing KCl
concentrations. ............................................................................................................ 68
Figure S2.3: DHA stability at 50°C in phosphate buffer. ............................................................. 69
Figure 3.1: Zoom view of enzyme-bound NAD (8-iodo-NAD for yADH). A. TmMtDH, B.
TmGlyDH, C. yADH, D. LmG6PDH, E. rLDH, F. bGDH. ....................................... 89
Figure 4.1: Maillard reaction scheme. .......................................................................................... 99
Figure 4.2: Time course of TmGlyDH inactivation at 50°C with and without DHA. ................ 105
Figure 4.3: Crystal structure of TmGlyDH showing the residues modified by DHA ................ 107
Figure 4.4: Screening of TmGlyDH mutants for increased resistance to inactivation by DHA at
50°C. ......................................................................................................................... 108
xi

Figure 4.5: Residual activity of TmGlyDH and its mutants after a 16-min incubation with DHA
at 80°C. ..................................................................................................................... 109
Figure 4.6: Time courses of TmGlyDH and mutants’ inactivations at 50°C with and without
DHA. .......................................................................................................................... 111
Figure 4.7: Inactivation time courses of TmGlyDH and its mutants at 80°C with and without
DHA. ..........................................................................................................................112
Figure 4.8: Specific activities and decay constants of wild-type TmGlyDH and its mutants at
80°C with and without DHA. ....................................................................................114

xii

KEY TO ABBREVIATIONS

2°ADH, secondary alcohol dehydrogenase
ADH, alcohol dehydrogenase
B3APA, bis(3-aminopropyl)amine
bGDH, Bovine liver type III L-glutamate dehydrogenase
BsGlyDH, Bacillus stearothermophilus glycerol dehydrogenase
CBA, Carboxyphenylboronic acid
CboFDH, Candida boidinii formate dehydrogenase
CHES, 2-(cyclohexylamino)ethanesulfonic acid
CV, column volumes
DETA, diethylene triamine
DHA, dihydroxyacetone
e.e., enantiomeric excess
ED, ethylene diamine
ESI-TOF, electrospray ionization time of flight
GlyDH, glycerol dehydrogenase
IAA, iodoacetamide
LDR, long-chain dehydrogenases/reductases
LmG6PDH, Leuconostoc mesenteroides glucose-6-phosphate dehydrogenase
MDR, medium-chain dehydrogenases/reductases
N1-CM-NAD, N1-carboxymethyl-NAD
N6-CM-NAD, N6-carboxymethyl-NAD

xiii

PEG, poly(ethylene glycol)
PQQ, pyrroloquinoline quinone
PseFDH, Pseudomonas sp.101 formate dehydrogenase
PTDH, phosphite dehydrogenase
rLDH, Rabbit muscle type XI L-lactate dehydrogenase
SDR, short-chain dehydrogenases/reductases
TBO, toluidine blue O
TFA, trifluoroacetic acid
TmGlyDH, Thermotoga maritima glycerol dehydrogenase
TmMtDH, Thermotoga maritima mannitol dehydrogenase
yADH, Saccharomyces cerevisiae alcohol dehydrogenase

xiv

CHAPTER 1. Introduction – Dehydrogenases as catalysts

1

1.1 INTRODUCTION
My project started as a collaboration between the Vieille laboratory in the Department of
Microbiology and Molecular Genetics, and the Worden and Calabrese Barton laboratories in the
Department of Chemical Engineering and Material Science at Michigan State University. Our
goal was to develop a dual-chamber bioelectrocatalytic reactor that uses dehydrogenase enzymes
to catalyze chemical reactions. The two enzymes selected for this project were Thermotoga
maritima mannitol dehydrogenase (TmMtDH) that interconverts mannitol and fructose and T.
maritima glycerol dehydrogenase (TmGlyDH) that interconverts glycerol and dihydroxyacetone
(DHA). In a reactor, the enzymes and NAD(H) cofactors would be immobilized on electrodes
that have been modified to allow electron transfer to/from NAD(H). TmMtDH was supposed to
catalyze fructose conversion to mannitol in the cathode compartment, and TmGlyDH was
supposed to convert glycerol to DHA in the anode compartment. The electrons and protons
released in the anode compartment were going to be used to regenerate NADH in the cathode
compartment (Figure 1.1), with the protons traveling through a proton-permeable membrane and
the electrons traveling through the external electrical circuit. I Studied the enzymes and
devoloped a method to immobilize NAD(H) onto an electrode surface. Partway into my project,
we were not able to reproduce early results of NAD+ electrochemical reduction, although NADH
oxidation was reproducible and new methods to increase the oxidation rate at lower
overpotentials were promising. Therefore, we decided to focus our project on the anode reaction,
for which a single chamber flow-through reactor may be ideal (Figure 1.2). For my part of the
project, it meant focusing on the study of TmGlyDH and NAD immobilization rather than on
both TmGlyDH and TmMtDH.

2

Figure 1.1: Two-chamber bioelectrocatalytic system for DHA and mannitol enzymatic
production.

Figure 1.2: A single-chamber flow-through reactor for DHA enzymatic production from
glycerol.

For dehydrogenase-based bioelectrocatalysis to be viable there needs to be (1) rapid
electron transfer between the electrode, mediator, cofactor, and enzyme; (2) rapid proton
transfer; (3) rapid mass transfer of reactants; and (4) rapid enzyme kinetics (69). Immobilization
3

of the enzyme and cofactor will eliminate the need for regenerated NAD(H) to diffuse through
the bulk liquid to the enzyme, reducing mass transfer limitations and allowing catalytic quantities
of enzyme and cofactor to be used (15). However, for immobilization to be successful, two
conditions must be met. First, NAD must be immobilized in an orientation that allows enzyme
binding. Second, the immobilized enzyme must retain high levels of activity. Another
requirement for bioelectrocatalysis is that the products should be of higher value than the
reactants. In the following sections, I am reviewing the background knowledge on NAD(P)dependent dehydrogenases, glycerol dehydrogenases, DHA as a high value-added chemical,
cofactor regeneration methods, and methods for cofactor and enzyme immobilization methods,
before introducing the main objectives of my project.

1.2 DEHYDROGENASE DIVERSITY AND FAMILIES
Dehydrogenases catalyze oxido-reduction (redox) reactions that transfer an electron pair
and a proton between a cofactor such as NADH and a substrate (68). These enzymes play an
important role in metabolism and are responsible for over 12% of all metabolic reactions in
Escherichia coli (50). NAD(P)-dependant dehydrogenases and reductases belong to one of four
groups, short-chain dehydrogenases/reductases (SDR) (21), medium-chain
dehydrogenases/reductases (MDR) (48), long-chain dehydrogenases/reductases (LDR) (22, 23),
and family III metal-dependent polyol dehydrogenases (52), based on size, sequence, structure,
and features of their catalytic mechanism (21-23, 48, 52). All characterized dehydrogenases
share the NAD binding Rossman fold motif and catalyze a hydride transfer (21, 48). Of note,
aldo-keto reductases, which evolved separately, do not share sequence similarity with
dehydrogenases, do not have a Rossman fold, and belong to the TIM barrel protein family, still
4

have active sites that are nearly superimposable to those of dehydrogenases (2, 21). Most known
dehydrogenases belong to either the MDR or SDR groups (21, 48).

1.2.1 The long-chain dehydrogenases
The LDR group is small, consisting of only 66 members as of 2003 (22). LDRs range in
size from 357 to 544 amino acids, and are primarily categorized based on their sequence
containing the Lys-X4-Asn-X2-His motif (X is any amino acid) (22). In general, LDRs are not
well characterized, but they are primarily active as monomers, are not metal dependent, have a
conserved lysine in their catalytic site, and are distinct structurally and mechanistically from the
MDRs and SDRs (21-23).

1.2.2 The medium-chain dehydrogenases
The MDR superfamily is composed of many large families, and smaller groups, many of
which are zinc dependent (48). MDRs use a catalytic zinc, and members of the alcohol
dehydrogenase (ADH) family also have an additional, structural zinc. There are two main
structural domains in MDRs, a substrate-binding domain and a Rossman fold NAD-binding
domain. The catalytic site sits in a deep cleft in the protein with the cofactor binding site. MDRs
can be monomers or multimers with different numbers of subunits and each monomer is
typically composed of about 350 amino acids. Most known MDRs can be grouped into 8
families, but it is estimated that there are 500 different MDR families (48).

5

1.2.3 The short-chain dehydrogenases
There are at least 140 different SDRs, making up about ¼ of all dehydrogenases, and they
are found in all domains of life (21). SDRs are unique because they are made up of only one
domain that is about 250 amino acids long. Most of them have a catalytic tyrosine in their active
site, but no residue is completely conserved. The one unifying factor is their NAD(P)-binding
Rossman fold. SDRs catalyze a broad range of reactions, being found in half of the different
enzyme class types (21).

1.2.4 The family III polyol dehydrogenases
The family III metal-dependent polyol dehydrogenases, formerly called iron-containing
dehydrogenases are metal and NAD-dependent cytoplasmic enzymes (52). They can either have
zinc or iron bound in their active site, or require iron in solution for activity (49). Many have
similar activity to the medium chain dehydrogenases, but do not show sequence similarity (52).
Although structurally different, all the dehydrogenase families share the Rossman fold for
NAD binding. The similarity in cofactor binding pocket is advantageous because a cofactor
immobilization method that works for one dehydrogenase is likely adaptable to many different
dehydrogenases.

1.3 GLYCEROL DEHYDROGENASES
Glycerol dehydrogenase (GlyDH) is a promising enzyme catalyst for DHA production from
glycerol. Glycerol dehydrogenases can be classified into four groups. The first group is
composed of cytoplasmic zinc-dependent enzymes that belong to the NAD-dependent family II
polyol dehydrogenases in the MDR superfamily (74). The second group comprises cytoplasmic
6

zinc-dependent enzymes that are part of the NAD-dependent family III metal-dependent polyol
dehydrogenases formerly called iron-containing dehydrogenases (52). Although similar in
activity, groups one and two GlyDHs do not show any sequence similarity (52). The third group
has not been well studied and comes from molds and filamentous fungi; these enzymes are
NADP-dependent and can reduce both DHA and glyceraldehyde to glycerol (32). The final
group represents membrane-bound pyrroloquinoline quinone (PQQ)-containing enzymes found
in many bacterial species. These enzymes are hard to purify and are unstable once purified (25).
In contrast, NAD-dependent cytoplasmic GlyDHs can be expressed and purified in high yields.
To reduce costs associated with dehydrogenase-based catalysis and to optimize reaction
conditions to an industrial setting, thermostable dehydrogenases are being investigated. Part of
the reduction in cost comes from simple purification of overexpressed proteins via heat
treatment. (56) Thermostable enzymes also increase operational temperatures for enzymatic
processes by reducing mass transfer limitations, reducing viscosity, increasing substrate and
product solubility, and reducing microbial contamination (8, 64). We chose to study the GlyDH
from T. maritima (TmGlyDH). T. maritima is a hyperthermophilic bacterium that grows
optimally at 80°C, and we expected TmGlyDH to be highly thermostable. TmGlyDH is a family
III NAD-dependent GlyDH whose crystal structure had been solved (7, 58) as part of a structural
genomics project targeting the T. maritima proteome, but whose catalytic properties remained
unstudied.
Family III NAD-dependent GlyDHs use zinc in their active site and share the Rossman
fold. Many have broad substrate specificity, showing high activity levels on 1,2-ethanediol, 1,2propanediol, 2-3-butanediol, 3-mercapto-1,2-propanediol, 3-chloro-1,2-propanediol, 1,2,3propanetriol monoacetate, 3-amino-1,2-propanediol, and 3-methoxy-1,2-propanediol (27, 57,

7

60). These enzymes have also been shown to produce a variety of (R) or (S) diols at 98% or
greater enantiomeric excess (e.e.) (73).

1.4 DHA PRODUCTION AND APPLICATIONS
We chose to synthesize DHA, because it is a value-added product that can be produced
from glycerol, a waste product of the biodiesel industry. DHA is the only FDA-approved active
ingredient used in sunless tanning creams (19), and it is used as a precursor for many valuable
products. DHA can be used as a precursor to 1,2-propylene glycol, which is used in antifreeze,
brake fluids, and as a precursor to polyurethanes and polypropylene glycols (59). DHA is also a
precursor of anti-ulcer, antihypertensive anti-inflammatory, analgesic (47), anti-ischemic, and α2-adrenergic receptor agonist drugs (11). Many chiral, optically active epoxide-based drugs (75),
pesticides (59), and sweeteners (6) use DHA as a precursor. The DHA global market was about
2,000 tons per year as of 2007 (45).
DHA is currently produced by glycerol oxidation in a large stirred tank using
Gluconobacter oxydans (16). A batch-fed process is used so that glycerol and DHA can be kept
below inhibitory concentrations (16). Glycerol oxidation to DHA needs large quantities of
oxygen, so aeration levels must be kept high using oxygen-enriched air (16). An industrial strain
of G. oxydans optimized to overcome DHA inhibition produced DHA at a rate of 3.6 g L-1 h-1
with a 85% yield in a two-step fed-batch process (3). G. oxydans has been further optimized to
produce DHA at a rate ranging from 12 g L-1 h-1 at a concentration of 100 g/L glycerol to 9.5 g
L-1 h-1 at a concentration of 140 g/L glycerol with a 96% yield. In addition the productivity based
on cell dry weight increased by 140% (31). Although optimization of the DHA production
process has increased yields and rate of production, the cost of DHA remains relatively high
8

because of the batch-fed process, high aeration levels required, and downstream purification
costs. A bioelectrocatalytic system has potential to reduce the production and purification costs
of DHA.

1.5 ENZYMATIC COFACTOR REGENERATION METHODS
Because of the variety of reactions catalyzed by dehydrogenases, these enzymes have
potential as catalysts for the production of many different compounds including chiral
compounds and pharmaceutical precursors (67, 68). In cells, the activity of dehydrogenases is
driven by cofactor regeneration in complementary pathways. In vitro, though, dehydrogenase
reactions depend on the cofactor being supplied in stoichiometric amounts or somehow
regenerated. This limitation has significantly hindered dehydrogenase-based industrial
applications (67, 68).

1.5.1 Formate dehydrogenase
Several methods have been developed to regenerate the cofactor, reducing overall cost. For
high value products such as chiral compounds, NADH can be regenerated by coupling the
desired reaction with another enzymatic reaction (67, 68). A promising enzyme for regenerating
NADH is formate dehydrogenase (FDH), which converts formate to CO2 while reducing NAD+
to NADH (55). For example, formate dehydrogenase was first used as a NADH regeneration
system to produce lactate (79% yield after 15 days) from pyruvate using D-lactate
dehydrogenase, producing D-lactate in 92% enantiomeric excess (55). This NADH regeneration
method is useful because CO2 is poorly soluble in water and easily removed from the reaction,
making the formate dehydrogenase reaction almost irreversible. Cofactor recycling with FDH
9

can thus be used to drive thermodynamically unfavorable reactions and formate has little effect
on dehydrogenase activity (67).
The use of FDH to regenerate the reduced cofactor has been limited by the low stability of
the enzyme (61). FDHs have Cys residues whose modification or oxidation, quickly inactivates
the enzyme. Also, FDH does not natively use NADP as a cofactor, enzyme production costs are
high, and specific activity is low. FDH has been extensively engineered to alleviate these
problems. The two most studied FDHs are from Candida boidinii (CboFDH) and Pseudomonas
sp.101 (PseFDH). PseFDH is the most active native FDH with a specific activity of 10 U per mg
(kcat = 7.3 s-1). Several of the PseFDHs Cys residues have been mutated resulting in a 1,000-fold
increased chemical stability. Directed evolution has been used to increase thermal stability of
PseFDH enough to allow a heat treatment step during purification, reducing purification cost.
The cofactor specificity has also been changed to allow FDH to be used with NADP-dependent
enzymes, although the activitiesare only about 10% of the wild-type enzyme with NAD (61).
Overall, with the exception of the PseFDH chemical stabilization, improvement in FDH
thermostability, activity levels, and activity with NADP have been incremental, though.

1.5.2 Phosphite dehydrogenase
Another promising cofactor-recycling enzyme is phosphite dehydrogenase (PTDH) (77).
PTDH catalyzes the oxidation of phosphite to phosphate, reducing NAD to NADH in the
process. With an equilibrium constant of 1011, the reaction is essentially irreversible (65). PTDH
has been engineered for increased activity with NADP, yielding a mutant enzyme with a 1,000fold increased catalytic efficiency with NADP (70), and a kcat with NADP reaching almost 50%
of the kcat with NAD (70). Interestingly, the catalytic efficiency of the mutant enzyme with NAD
10

also increased 3.6-fold. PTDH’s thermostability has been increased by directed evolution,
increasing the half-life at 45°C 7,000-fold (18). PTDH is likely a better recycling enzyme than
FDH for reactions involving NADP, because the catalytic efficiency of PTDH with NADP is 39fold higher than that of the commercially available mutant PseFDH while the kcat is comparable
(70).

1.5.3 Regeneration of oxidized cofactors
Regeneration of the oxidized cofactor NAD(P)+ has not been as well studied as
regeneration of the reduced form because the reduced form is used more often, in particular to
drive enzymatic reactions that produce chiral compounds (63). Regenerating the oxidized
cofactor could allow dehydrogenases to be used for clarifying racemic mixtures of enantiomers,
or synthesizing difficult-to-prepare aldehydes and ketones. Glutamate dehydrogenase, which is
active with both NAD and NADP, could be used to recycle the oxidized cofactor by reducing αketoglutarate. NAD+ could also be recycled using NADH oxidase. Many NADH oxidases
produce hydrogen peroxide, which can damage proteins, so the addition of catalase would be
required. But certain NADH oxidases produce water instead and would be preferred.
A major limitation of enzyme-based cofactor regeneration is the added cost of using two
enzymes and two reactants while still producing only one product. Several other enzymes have
been used to regenerate NADH including glucose-6-phosphate dehydrogenase, glucose
dehydrogenase, alcohol dehydrogenase, and hydrogenase, (67). All of these enzymes require
extra reaction components, often complicating product purification and increasing overall cost.
Regenerating NAD(H) with electrical current could be a lower cost alternative, allowing new
dehydrogenase-based catalytic systems to be developed.
11

1.6 ELECTROCHEMICAL COFACTOR REGENERATION
1.6.1 Direct cofactor regeneration
There are two electrochemical methods of regenerating NAD(H), direct electron transfer
and mediated electron transfer. Direct electron transfer involves electrons being transferred
directly between the electrode and NAD(H). Carbon electrodes can be used to oxidize NADH at
moderate potentials (ca 0.64 V vs standard hydrogen electrode [SHE]) (35), but the reaction
kinetics are quite poor and NAD adsorbs onto the electrode surface, interfering with activity
(35). Using high overpotentials of around 1V greater than the formal potential (-320 mV vs SHE)
(10, 35) increases the catalytic rate but can lead to unwanted reactions taking place, cofactor
degradation, and electrode fouling (4, 5). Direct NADH oxidation has been combined with
different dehydrogenases in biosensors to detect ethanol, lactate, glutamate, hydroxybutyrate,
and cholic acid with the NAD either in solution, entrapped under a membrane, or covalently
attached (35). In general these biosensors had low stability (losing most activity after a few days
to a few weeks) and used applied voltages of ca ~1 V (vs SHE) (35). In contrast, FAD-dependent
enzymes such as glucose oxidase readily allow direct electron transfer (17, 35, 71).
Direct electrochemical reduction of NAD+ also requires high overpotentials and requires two
electrons for each NAD+ molecule (53, 54). A radical intermediate forms during direct reduction
because electrons are transferred one by one (53, 54). Two NAD radicals can then react together
to form an inactive dimer (Figure 1.3) (53, 54).

12

Figure 1.3: Electrochemical reduction of NAD+. R= Adenosine diphosphate ribose.

1.6.2 Mediated cofactor regeneration
The problems associated with direct electrochemical oxidation of NADH have been
addressed using mediated electron transfer. Mediators accept electrons from NADH and then
diffuse to the electrode surface where they are oxidized at a lower potential than NADH (35).
The most studied mediators are quinones, quinoneimines, and azines such as phenazine. One
azine mediator that can be readily covalently bound to an electrode through carbodiimide
coupling is toluidine blue O (TBO) (14, 15). TBO has been coimmobilized with the cofactor and
alcohol dehydrogenase and used for the electrochemical detection of alcohol. Mediated transfer
has also been combined with the enzyme diaphorase (10). Diaphorase regenerates NAD+ with a
benzyl viologen (10) or ferrocene mediator (66), transferring electrons from the electrode to
diaphorase where NADH is oxidized. Diaphorase-coupled regeneration is less desirable than

13

using a chemical mediator because diaphorase is oxygen sensitive and adding a second enzyme
increases cost.
Similarly, the electrode surface can be directly modified with a catalyst that can activate
NADH oxidation. The difference between a mediator and a catalyst can be unclear at times, but
generally a mediator will carry electrons to a distance from the electrode while a catalyst will be
incorporated into the electrode surface. Adding carbon nanotubes to an electrode lowers the
overpotential required for NADH oxidation and increases the surface area of the electrode (76).
A carbon nanotube-modified electrode worked as a glucose biosensor, which operated
continually for 25 hours without a significant reduction in current and maintained 64% of
activity after being stored dry for 1.5 months (76). NADH oxidation took place readily at
voltages of ca 180 mV vs SHE, which is 300 mV lower than a bare glassy carbon electrode (76).
Adding a mediator/catalyst, such as an azine can increase current density and further lower the
oxidation potential (30).
Methylene green and TBO have also been electropolymerized onto a carbon nanotubemodified electrode, catalyzing NADH oxidation (30). Polymethylene green was a more effective
NADH oxidation catalyst than polyTBO. Electropolymerization takes place by cycling the
potential of the electrode from low potential (ca -0.7 V vsSHE) to high potential (c a1.3 V vs
SHE) for many cycles, where the polymerization takes place around 1V. Methylene green
polymerized onto a carbon nanotube-modified electrode is promising as a NAD+ regeneration
system. In addition, electroactivation of carbon electrodes can increase the catalytic rate of
NADH oxidation (29). Electroactivation works by cycling the potential of the electrode from
ca -1.7 V to 2.3 V vs SHE for twenty cycles, where at the higher voltage the carbon can oxidize
forming a variety of functional groups, of which the formation of quinones is the main

14

contributor to NADH oxidation. In addition, carbon nanotubes can be electroactivated,
increasing the surface area and further increasing the catalytic rate. Adding methylene green and
polymethylene green can lead to current densities of 7.0 mA/cm2 and 3.8 mA/cm2, respectively,
in 20 mM NADH at ca 28 mV vs SHE (29). The high current densities produced make
electrochemical regeneration of NAD+ feasible for a dehydrogenase-based catalytic reactor.

1.7 MODIFYING ELECTRODES FOR IMMOBILIZATION OF MEDIATOR AND
COFACTOR
Two approaches to modifying electrodes can be used. In the first approach, a small
chemical with a carboxyl group (e.g., cysteine or glycine) is electrolytically deposited or
polymerized on the electrode, while the second approach involves depositing on the electrode
polymers that have reactive groups that can be further modified; such methods for carbon
electrodes are currently being developed in the Worden and Calabrese-Barton laboratories
(Department of Chemical Engineering and Materials Science, Michigan State University). The
first approach has already been used for cysteine on gold electrodes (15). Firest, a selfassembling monolayer of cysteine was prepared on a gold electrode. The mediator (TBO) was
then bound to cysteine’s carboxylic group, and the boronate linker carboxyphenyl boronic acid
was bound to cysteine’s amino group using amide linkages (Figure 1.4) (15). Then NAD(P) was
bound to the boronate linker through its cis-diols. This interface with immobilized NAD was
tested with Saccharomyces cerevisiae ADH, and immobilized NADP was tested with
Thermoanaerobacter ethanolicus secondary ADH (2°ADH). The enzymes were immobilized
onto the interface using glutaric dialdehyde fixation (15). Current was observed to increase
linearly from 0 to 25 mM of enzyme substrate (ethanol for ADH and 2-propanol for 2°ADH).
15

The 2°ADH interface lost only 2% of its activity after 24 hours of continuous operation at room
temperature and it could be stored dry for 3 weeks without activity loss. The interface started to
lose 5% of activity per rinse after 10 rinses (15).

Figure 1.4: Two possible NAD binding orientations on a boronate linker. Methionine (Met)
is directly bound to the electrode, carboxyphenylboronic acid (CBA) and TBO are bound to
MET through amide bonds, and NAD is bound to CBA through its cis-diols. The dotted lines
represent the limits between different molecules.

The second approach was designed to make easily renewable and higher loading interfaces.
This method involves a layer-by-layer build up using polyelectrolyte polymers (14). Polymers
with carboxylic or amine groups are layered, alternating between positively and negatively
charged polymers. After 3-mercaptopropionic acid attaches to the electrode, the TBO mediator is
bound to its terminal carboxylic group, then a layer of poly(ethyleneimine) is added to the
electrode. A linker such as carboxyphenyl boronic acid can then be attached to the terminal
16

amine groups in a reaction with N-(3-dimethylaminopropyl)-N′-ethylcarbodiimide hydrochloride
(EDC), followed by NAD binding (14). A 0.1 M HCl solution can be used to remove the
poly(ethyleneimine)-attached NAD and allow the electrode interface to be renewed. Then a new
layer can be applied, restoring electrode activity. After regeneration, the saturation current,
sensitivity, and turnover rate were reduced by 8.3%, 16%, and 5% respectively (14). Saturation
current is the maximum current measured as the enzyme substrate is increased and sensitivity is
the change in current observed after a change in substrate concentration. The layer-by-layer build
up method could also be used for an amine-containing linker, if the final polymer layer contains
carboxylic groups.
In both of these methods, a boronate linker was used to attach NAD, but boronate-based
linkages have two problems for the applications proposed for the current work. First, NAD has
two cis-diols, so NAD can bind to the boronate in two possible conformations which could affect
its ability to bind to the enzyme and affect catalytic efficiency (Figure 1.2) (15).
Second, the bond formed between boronate and NAD is relatively unstable, with a
dissociation constant (Kd) of 5.9  10-3 (1). Boronate also has affinity for cis-diols found on
sugars and sugar alcohols; the Kd for mannitol is similar to that for NAD and the Kd for fructose
is smaller than that for NAD (1). In a catalytic reactor using TmGlyDH glycerol will be present
in high concentrations. Because the Kd for glycerol is likely in the same order of magnitude as
that for NAD, the boronate linkage would be unstable, with high concentrations of glycerol
displacing NAD. For a catalytic reactor, stability is a particularly important issue because the
catalytic interface would need to be stable over a long period of time (several days to several
weeks).

17

1.8 COFACTOR IMMOBILIZATION
For dehydrogenase-based catalysis to be economically feasible on a large scale, the cofactor
needs to be easily regenerated and recovered. Cofactor regeneration during the reaction allows
products to accumulate using only small quantities of NAD, while NAD recovery after the
reaction allows it to be used again. Recovering the soluble cofactor for reuse in another reaction
can be difficult and increases reaction costs. Immobilization of the cofactor and enzyme is ideal,
because it allows reuse of the entire catalytic system. In an electrochemical recycling system the
cofactor would ideally be bound to the electrode surface together with the enzyme to facilitate
electron transfer from the electrode to the enzyme active site. Physically entrapping the cofactor
in a matrix is not ideal, because the NAD can slowly diffuse away from the electrode (40).
Membranes have been used to entrap enzyme and cofactor at the electrode surface successfully
for biosensors (24, 28, 46). Using membranes in a catalytic reactor is not ideal, though, because
the membrane would have to be selective to allow reactant and product to diffuse while
preventing cofactor and enzyme from diffusing, and the membrane can reduce mass transfer
rates (34). Covalent attachment of the cofactor to an electrode is more promising for catalytic
use because it is adaptable to high surface area electrodes and the immobilization process allows
the addition of a linker molecule that can reduce steric hindrance between the enzyme and the
surface to which NAD is bound.

1.8.1 Modification of NAD’s N1 and N6-amines
Covalent modification of NAD was studied heavily in the 1970s and 1980s (9, 37). Some of the
earliest NAD immobilizations involved directly coupling NAD to a carboxyl-containing polymer
matrix using a carbodiimide coupler (26) or directly coupling NAD to CNBr-activated agarose
18

(38). The first well-characterized NAD+ analogue was synthesized by Lindberg et al. (33). The
authors alkylated NAD+ with iodoacetic acid to form N1-carboxymethyl-NAD+, which then
rearranges to N6-CM-NAD+ under alkaline conditions (Figure 1.5). N1-CM-NAD+ was reported
to have a broad-shouldered absorbance maximum at 259 nm. N6-CM-NAD+ was reported to
have an absorbance maximum at 266 nm, as opposed to a shoulderless absorbance maximum at
260 nm for NAD+. N6-CM-NAD+ was also coupled to 1,6-diaminohexane using a carbodiimidecoupling reagent giving NAD+-N6-[N-(6-aminohexyl)-acetamide] while maintaining the
absorbance maximum at 266 nm. The NAD+-N6-[N-(6-aminohexyl)-acetamide] was then
immobilized by reacting with CNBr-activated sepharose. The sepharose-bound NAD+ was
stable, retaining activity as a functioning cofactor and affinity chromatography matrix for several
months. When tested in enzyme assays with three dehydrogenases (horse liver ADH, beef heart
type III lactate dehydrogenase, and pig heart malate dehydrogenase), N6-CM-NAD+ showed 5575% of activity as compared to native NAD+, while the NAD+-N6-[N-(6-aminohexyl)-acetamide]
analogue showed 50-100% activity as compared to native NAD+. Enzyme activity with the
sepharose-bound cofactor was also tested with a three-enzyme system where malate
dehydrogenase converted malate to oxaloacetate, citrate synthase converted oxaloacetate to
citrate, and lactate dehydrogenase recycled the cofactor. The reaction rate increased after the
addition of the cofactor recycling enzyme. As an affinity chromatography matrix, sepharosebound NAD+ was able to bind ADH and lactate dehydrogenase, while bovine serum albumin
passed through the column (33).

19

Figure 1.5: NAD+ structure. The N6-amine, phosphate backbone, and cis diols are possible
immobilization points.

Building on the early work by Lindberg et al. (33), NAD was modified at the N1-amine by
various reactive species (i.e., iodoacetic acid, aziridine, propionolactone, and epoxides), creating
a terminal carboxyl, amino, or vinyl group that could be used for immobilization (37). The N1modified NAD could then be immobilized, or rearranged into N6-modified NAD via alkaline
Dimroth rearrangement. One disadvantage of using the N1-modified analogue is that it is
unstable at basic pH, because the Dimroth rearrangement will take place slowly under mild
alkaline conditions. The stability is low enough that, when the alkylation reaction is performed
over fifteen days at pH 9.0, the main product is the N6-derivative instead of the N1-derivative
formed at pH 6.5 (44). Recently, a mixture of chitosan-immobilized N1-CM-NAD and chitosanimmobilized glucose dehydrogenase in a water solution of carbon nanotubes was dried on an
electrode and used as a glucose biosensor (76). The addition of a Nafion membrane over the
sensor increased the sensitivity and current density, giving a sensitivity of 1.8 mA M-1 cm-2 with

20

a linear range of 0.02-2.0 mM. The Nafion-modified electrodes were quite stable, maintaining
100% of activity over 24 hours and 64% of activity after storage for 1.5 months (76).
Another way to modify the N6-position is to use an NAD derivative containing a 6-chloroor 6-mercapto-modified adenine and use nucleophilic substitution by an amine-containing
compound (37). This approach is not ideal, however, because it requires starting with an NAD
analogue. Direct acylation of the N6-amine is also possible, but the acylated derivative is poorly
stable above pH 7 (72). NAD derivatives modified at the C8 of the adenine ring have also been
synthesized, but they generally have lower coenzymic activity than the N1 or N6 derivatives (37).
The N6-amine group on NAD(H) is a promising candidate for linkage, as it can be modified to
have a terminal carboxyl group (33, 39). NAD(P)’s N6 amine has been used to attach large
molecules like poly(ethylene glycol) (PEG) to NAD to facilitate NAD reuse and removal from
an enzymatic reaction (20, 43).
Recently, NAD has also been immobilized onto silica nanoparticles using the bifunctional
linker glycidoxypropyltrimethoxysilane in which the epoxide is able to alkylate NAD at the N1amine and the silane group can be attached to silica (34). Nanoparticle-bound cofactor,
nanoparticle-bound bovine glutamate dehydrogenase, and nanoparticle-bound rabbit lactate
dehydrogenase were prepared by amine reaction with the epoxide-modified nanoparticles. A
mixture of all three types of nanoparticles allowed the two enzymes to recycle the cofactor, while
Brownian motion allowed the cofactor to be transferred between the two kinds of enzymeparticles. A similar method was used to immobilize NAD through its N6-amine into a sol-gelbased biosensor (66). In this case, a diaphorase and ferrocene mediator were used to
electrochemically sense and regenerate the cofactor on the electrode, in the presence of a coimmobilized dehydrogenase. The epoxide of glycidoxypropyltrimethoxysilane was used to

21

modify NAD at the N6-amine and the silane was covalently incorporated into the sol-gel matrix.
This biosensor assembly worked for D-sorbitol dehydrogenase, D-glucose dehydrogenase, and Llactate dehydrogenase. Polyethyleneimine was added to stabilize the diaphorase and sorbitol
dehydrogenase, and ferrocene was immobilized to the polyethyleneimine (66).

1.8.2 Immobilized NAD in affinity chromatography
N6-immobilized NAD has been thoroughly studied for affinity chromatography using a
kinetic locking-on strategy (13, 41, 42, 62). In this strategy, NAD is immobilized on a resin and
NAD-dependent enzymes can be “locked on” to the NAD-resin in the presence of a competitive
inhibitor, typically a substrate analogue.Thus, if NAD+ is used, a non-reactive oxidized substrate
analogue will also be used (41). This approach helps the enzyme form a ternary enzymesubstrate-cofactor complex but no reaction takes place, holding the enzyme onto the NAD-resin.
A variety of NAD immobilization methods have been tested for this purpose, including 8’-azo,
N6, S6, N1, and linkages to NAD’s phosphate backbone (13). Enzyme retention was the strongest
for five of the six dehydrogenases tested when N6-immobilized NAD was used as part of the
affinity matrix compared to the other linkages (13). The one enzyme that did not follow this
pattern was bovine glutamate dehydrogenase, which showed strong binding to the S6- and 8’azo-linked NAD but not to NAD bound through the N6-amine. All of the linkages for
immobilizing NAD showed some ability to be used for affinity chromatography. The N6 linkage
showed higher NAD loading and higher levels of enzyme redox-active NAD (up to 2.1 µmol/g
wet sepharose resin) (13), indicating that N6 linkages are the best. Because of the apparent broad
acceptance for binding of N6-immobilized NAD by dehydrogenases, immobilization of NAD
through the N6-amine was selected for this project.

22

1.8.3 Direct coupling of NAD
An alternative method of immobilizing NAD involves direct coupling of NAD to a
carboxy-terminating spacer arm (42). This method results in NAD immobilized primarily
through its phosphate groups, but because adenine can be directly coupled to carboxy-sepharose,
some immobilization likely takes place through the N6-amine as well (42). To elucidate the point
of attachment, NAD, NADP, adenine, and inosine were immobilized by direct coupling in
different conditions (42). When coupled in pyridine using dicyclohexylcarbodiimide,
immobilization appeared to occur at the ribosyl hydroxyls, but coupling in water at pH 4.8 using
EDC seemed to take place through NAD’s pyrophosphate backbone or NADP’s phosphate. In
water at pH 4.8, inosine could not be coupled directly, indicating that the ribosyl moiety is not
involved in immobilization. Adenine could be coupled directly in water at pH 4.8, with an
observed absorbance maximum shift from 260 nm to 230 nm. In affinity chromatography,
phenylalanine dehydrogenase did not bind to directly immobilized NAD, but it could bind to N6immobilized NAD, indicating a different mode of attachment. The absorption maximum of
directly coupled NAD shifted to around 245 nm, indicating that some coupling took place
through the adenine ring (42).
Although simple to prepare, directly coupled cofactors were somewhat unstable, losing
between 30-70% of coupled NAD over a 1 month time and the absorbance maximum slowly
shifted from 245 nm to 260 nm over a 3 month time (42). The amount of enzyme bound to the
affinity chromatography matrix was generally lower with directly coupled NAD than with N6immobilized NAD. The lower stability and potential mixture of immobilized products would
make it difficult to characterize directly coupled NAD-modified electrodes. Direct coupling also
yields lower levels of accessible NAD than linking to N6-CM-NAD (13, 42). However, it could

23

be useful because it eliminates the need to synthesize and purify N6-CM-NAD, simplifying the
linkage synthesis process. In addition, the lower NAD loading may not affect our electrode
system because enzyme loading might be the limiting factor for catalytic rate, since the enzyme
is substantially larger than NAD.

1.9 ENZYME IMMOBILIZATION
Enzymes can be immobilized through adsorption, covalent binding, affinity
immobilization, and entrapment (12). Generally, immobilized enzymes maintain activity and are
more stable. Immobilization can increase storage stability, thermostability, and stability in
organic solvents, and reduce proteolysis, aggregation, and product inhibition (12, 51). Some of
the support matrices used for enzyme immobilization include chitosan, collagen, carrageenan,
gelatin, cellulose, starch, pectin, sepharose, and other synthetic polymers and inorganic support
materials (12). Enzyme adsorption includes both liquid adsorption and drying an enzyme
solution on a matrix. Adsorption of enzymes results from salt bridges, hydrophobic interactions
and hydrogen bonds depending on the matrix used (12). Although generally, adsorbed enzymes
slowly leach from the matrix, the interaction can be irreversible if the right matrix is used. Fibers
with good water holding capacity work well for adsorbing enzymes (12). Covalent
immobilization involves using a matrix with reactive groups that can react with protein side
chains such as Lys and Glu (12). Often a matrix is modified by a reactive species followed by the
addition of enzyme. A variety of linking agents have been used, but glutaraldehyde is often used
because it can bind the enzyme to the matrix and stabilize the enzyme by forming crosslinks
(12). Enzymes can be immobilized by entrapping them in gels, fibers, or under membranes (12).
Entrapped enzymes are physically held in the matrix by the structure of the matrix. Enzymes
24

entrapped in gels and fibers often leach out of the matrix, but selecting different support
materials has somewhat alleviated these problems (12). Affinity immobilization results from the
enzyme having affinity for the support matrix (12). The matrix can be modified to carry an
enzyme ligand, or the enzyme can be modified to have affinity to the matrix. An example would
be adding streptavidin to an enzyme and attaching biotin to the matrix or vice versa.

1.9.1 Immobilization of glycerol dehydrogenase
The glycerol dehydrogenases from Geobacillus stearothermophilus, citrobacter braakii,
and cellulomonas sp have recently been covalently immobilized on agarose (51). Agarose
activated by CNBr, agarose activated with glyoxyl groups, and agarose with a combination of
glyoxal and amine groups were tested. Immobilization was tested in two buffers (sodium
bicarbonate and CAPS) and with two additives (glycerol and PEG600). GlyDH immobilized on
the CNBr-activated agarose gave higher activity levels (42-65%) than on glyoxyl-modified
agarose (5-29%). The addition of glycerol or PEG to the glyoxal immobilization reaction
increased activity, but glycerol hampered the glyoxal chemistry. The half-life of the immobilized
enzymes increased up to 4 times. Of the three enzymes tested, G. stearothermophilus GlyDH
was the least stable in soluble form, but was the most stable after immobilization, retaining 100%
of its activity after 18 hours at 65°C (51).
In our reactor system, we foresee that the cofactor will be covalently immobilized on the
electrode first, followed by binding the enzyme to the immobilized cofactor by affinity. Then
glutaraldehyde fixation can be used to covalently attach and stabilize the enzyme.
Glutaraldehyde has been shown to immobilize other dehydrogenases while maintaining their
activity and increasing their stability (15, 36).
25

1.10 INTRODUCTION TO PROJECT
A large part of my project has been the characterization of TmGlyDH. Although
TmGlyDH had been previously crystallized, the enzyme had not been kinetically characterized.
In Chapter 2, I describe the characterization of TmGlyDH and demonstrate that the enzyme was
able to use N6-immobilized NAD. Kinetic characterization is important for determining what
conditions are suitable for a bioelectronic reactor, and the enzymes kinetic data can be compared
to electrochemical rates. Another part of my project was to immobilize NAD while maintaining
its coenzymic activity (Chapter 2). Although a bioelectronic interface had been previously set up
in the Worden laboratory, NAD was immobilized onto a boronate linker through its cis-diols,
which—as I described earlier—is unstable in the presence of other cis-diols. I chose a different
method (modifying NAD at the N6-amine) of immobilizing NAD and then ensured that the
enzyme maintained activity with the immobilized NAD. In Chapter 3, I describe testing several
different dehydrogenases with the immobilized NAD to show that N6-immobilized NAD in an
electrochemical system will likely be usable for many other dehydrogenases; I also check if the
structure of the enzymes’ binding pocket can predict activity with the immobilized cofactor. In
Chapter 4, I describe how TmGlyDH is inactivated by its product (DHA), and engineer the
enzyme for higher stability in the presence of DHA. In Chapter 5, I discuss the suitability of
using TmGlyDH in a bioelectronic system, the suitability of N6-immobilized NAD, and
recommended future work on the project.

26

REFERENCES

27

REFERENCES

1.

Bailon, P., G. K. Ehrlich, W.-J. Fung, and W. Berthold (ed.). 2000. Methods in
molecular biology: affinity chromatography methods and protocols, vol. 147. Humana
Press Inc, Totowa NJ.

2.

Barski, O. A., S. M. Tipparaju, and A. Bhatnagar. 2008. The aldo-keto reductase
superfamily and its role in drug metabolism and detoxification. Drug. Metab. Rev.
40:553–624.

3.

Bauer, R., N. Katsikis, S. Varga, and D. Hekmat. 2005. Study of the inhibitory effect
of the product dihydroxyacetone on Gluconobacter oxydans in a semi-continuous twostage repeated-fed-batch process. Bioprocess Biosys. Engin. 28:37–43.

4.

Bergel, A., and M. Comtat. 1991. Thin-layer spectrochemical study of the reversible
reaction bewteen nicotinamide adenine dinucleotide and flavin adenine dinucleotide. J.
Electroanal. Chem. 302:219–231.

5.

Blaedel, W. J., and R. A. Jenkins. 1975. Study of the electrochemical oxidation of
reduced nicotinamide adenine dinucleotide. Anal. Chem. 47:1337–1343.

6.

Bliznak, J. B., and R. E. Harcarufka. 1981. Sweetener and flavoring compositions and
method of producing same. United States Patent:4,277,511.

7.

Brinen, L. S., J. M. Canaves, X. Dai, A. M. Deacon, M. A. Elsliger, S. Eshaghi, R.
Floyd, A. Godzik, C. Grittini, S. K. Grzechnik, C. Guda, L. Jaroszewski, C. Karlak,
H. E. Klock, E. Koesema, J. S. Kovarik, A. Kreusch, P. Kuhn, S. A. Lesley, D.
McMullan, T. M. McPhillips, M. A. Miller, M. D. Miller, A. Morse, K. Moy, J.
Ouyang, A. Robb, K. Rodrigues, T. L. Selby, G. Spraggon, R. C. Stevens, H. v. d.
Bedem, J. Velasquez, J. Vincent, X. Wang, B. West, G. Wolf, S. S. Taylor, K. O.
Hodgson, J. Wooley, and I. A. Wilson. 2003. Crystal structure of a zinc-containing
glycerol dehydrogenase (TM0423) from Thermotoga maritima at 1.5 Å resolution.
Proteins: Struct., Funct., Bioinf. 50:371–374.

8.

Bruins, M., A. Janssen, and R. Boom. 2001. Thermozymes and their applications.
Appl. Biochem. Biotechnol. 90:155–186.

9.

Bückmann, A., and G. Carrea. 1989. Synthesis and application of water-soluble
macromolecular derivatives of the redox coenzymes NAD(H), NADP(H) and FAD. Adv.
Biochem. Eng. Biotechnol. 39:97–152.

10.

Calabrese Barton, S., J. Gallaway, and P. Atanassov. 2004. Enzymatic biofuel cells
for implantable and microscale devices. Chem. Rev. 104:4867–4886.

11.

Cossement, E., J. P. Geerts, J. Gobert, P. Michel, and E. Wulfert. 1990. Substituted
1-(1H-imidazol-4-yl)alkyl-benzamides as anti-ischemics and as α-2-adrenergic receptor
agonists. United States Patent:4,923,865

12.

Datta, S., L. R. Christena, and Y. R. S. Rajaram. 2013. Enzyme immobilization: an
overview on techniques and support materials. Biotech. 3:1–9.
28

13.

Forde, J., L. Oakey, L. Jennings, and P. Mulchahy. 2005. Fundamental differences in
bioaffinity of amino acid dehydrogenases for N6- and S6-linked immobilized cofactors
using kinetic-based enzyme capture strategies. Anal. Biochem. 338:102–112.

14.

Hassler, B. L., N. Kohli, J. G. Zeikus, I. Lee, and R. M. Worden. 2007. Renewable
dehydrogenase-based interfaces for bioelectronic applications. Langmuir 23:7127–7133.

15.

Hassler, B. L., and R. M. Worden. 2006. Versatile bioelectronic interfaces based on
heterotrifunctional linking molecules. Biosens. Bioelectron. 21:2146–2154.

16.

Hekmat, D., R. Bauer, and J. Fricke. 2003. Optimization of the microbial synthesis of
dihydroxyacetone from glycerol with Gluconobacter oxydans. Bioprocess Biosys. Eng.
26:109–116.

17.

Jasti, L. S., S. R. Dola, N. W. Fadnavis, U. Addepally, S. Daniels, and S.
Ponrathnam. 2014. Co-immobilized glucose oxidase and β-galactosidase on bovine
serum albumin coated allyl glycidyl ether (AGE)-ethylene glycol dimethacrylate
(EGDM) copolymer as a biosensor for lactose determination in milk. Enzyme Microb.
Technol. 64–65:67–73.

18.

Johannes, T. W., R. D. Woodyer, and H. Zhao. 2005. Directed evolution of a
thermostable phosphite dehydrogenase for NAD(P)H regeneration. Appl. Environ.
Microbiol. 71:5728–5734.

19.

Johnson, J. A., and R. M. Fusaro. 1994. Alteration of skin surface protein with
dihydroxyacetone: A useful application of the Maillard browning reaction, p. 114–119. In
T. P. Labuza, G. A. Reineccius, V. M. Monnier, J. O'Brien, and J. W. Baynes (ed.),
Maillard Reactions in Chemistry, Food, and Health. The Royal Society of Chemistry,
Cambridge.

20.

Katayama, N., I. Urabe, and H. Okada. 1983. Steady-state kinetics of coupled twoenzyme reactor with recycling of poly(ethylene glycol)-bound NAD. Eur. J. Biochem.
132:403–409.

21.

Kavanagh, K. L., H. Jörnvall, B. Persson, and U. Oppermann. 2008. The SDR
superfamily: functional and structural diversity within a family of metabolic regulatory
enzymes. Cell. Mol. Life Sci. 65:3895–3906.

22.

Kavanagh, K. L., M. Klimacek, B. Nidetzky, and D. K. Wilson. 2003. Crystal
structure of Pseudomonas fluorescens mannitol 2-dehydrogenase: evidence for a very
divergent long-chain dehydrogenase family. Chem.-Biol. Interact. 143–144:551–558.

23.

Klimacek, M., and B. Nidetzky. 2002. A catalytic consensus motif for D-mannitol 2dehydrogenase, a member of a polyol-specific long-chain dehydrogenase family,
revealed by kinetic characterization of site-directed mutants of the enzyme from
Pseudomonas fluorescens. Biochem. J. 367:13–18.

24.

Kulys, J. J., U. Bilitewski, and R. D. Schmid. 1991. Reagentless biosensors for
substrates of dehydrogenases. Anal. Lett. 24:181–189.

25.

Lapenaite, I., A. Ramanaviciene, and A. Ramanavicius. 2006. Current trends in
enzymatic determination of glycerol. Crit. Rev. Anal. Chem. 36:13–25.

29

26.

Larsson, P.-O., and K. Mosbach. 1971. Preparation of a NAD(H)-polymer matrix
showing coenzymic function of the bound pyridine nucleotide. Biotechnol. Bioeng.
13:393–398.

27.

Lee, L. G., and G. M. Whitesides. 1986. Preparation of optically-active 1,2-diols and hydroxy ketones using glycerol dehydrogenase as catalyst - limits to enzyme-catalyzed
synthesis due to noncompetitive and mixed inhibition by product. J. Org. Chem. 51:25–
36.

28.

Li, G., N. Z. Ma, and Y. Wang. 2005. A new handheld biosensor for monitoring blood
ketones. Sens. Actuators, B 109:285–290.

29.

Li, H., R. Li, R. M. Worden, and S. Calabrese Barton. 2014. Facilitation of high-rate
NADH electrocatalysis using electrochemically activated carbon materials. ACS Appl.
Mater. Inter. 6:6687–6696.

30.

Li, H., H. Wen, and S. Calabrese Barton. 2012. NADH oxidation by
electropolymerized azines on carbon nanotube modified electrodes. Electroan. 24:398–
406.

31.

Li, M., J. Wu, X. Liu, J. Lin, D. Wei, and H. Chen. 2010. Enhanced production of
dihydroxyacetone from glycerol by overexpression of glycerol dehydrogenase in an
alcohol dehydrogenase-deficient mutant of Gluconobacter oxydans. Bioresour. Technol.
101:8294–8299.

32.

Liepins, J., S. Kuorelahti, M. Penttila, and P. Richard. 2006. Enzymes for the
NADPH-dependent reduction of dihydroxyacetone and D-glyceraldehyde and Lglyceraldehyde in the mould Hypocrea jecorina. FEBS J. 273:4229–4235.

33.

Lindberg, M., P. Larsson, and K. Mosbach. 1973. A new immobilized NAD+
analogue, its application in affinity chromatography and as a functioning coenzyme. Eur.
J. Biochem. 40:187–193.

34.

Liu, W., S. Zhang, and P. Wang. 2009. Nanoparticle-supported multi-enzyme
biocatalysis with in situ cofactor regeneration. J. Biotechnol. 139:102–107.

35.

Lobo, M. J., A. J. Miranda, and P. Tuñón. 1997. Amperometric biosensors based on
NAD(P)-dependent dehydrogenase enzymes. Electroan. 9:191–202.

36.

López-Gallego, F., L. Betancor, C. Mateo, A. Hidalgo, N. Alonso-Morales, G.
Dellamora-Ortiz, J. M. Guisán, and R. Fernández-Lafuente. 2005. Enzyme
stabilization by glutaraldehyde crosslinking of adsorbed proteins on aminated supports. J.
Biotechnol. 119:70–75.

37.

Lowe, C. R. 1978. Immobilised coenzymes. Trends Biochem. Sci. 3:134–137.

38.

Lowe, C. R., and P. D. G. Dean. 1971. Affinity chromatography of enzymes on
insolubilized cofactors. FEBS Lett. 14:313–316.

39.

Muramatsu, M., I. Urabe, Y. Yamada, and H. Okada. 1977. Synthesis and kinetic
properties of a new NAD+ derivative carrying a vinyl group. Eur. J. Biochem. 80:111–
117.

30

40.

Noguer, T., D. Szydlowska, J.-L. Marty, and M. Trojanowicz. 2004. Sol-gel
immobilization of aldehyde dehydrogenase and NAD+ on screen-printed electrodes for
designing of amperometric acetaldehyde biosensor. Polish J. Chem. 78:1679–1689.

41.

O'Flaherty, M., M. McMahon, and P. Mulcahy. 1999. A kinetic locking-on strategy
for bioaffinity purification: further studies with alcohol dehydrogenase. Protein Expr.
Purif. 15:127–145.

42.

Oakey, L., and P. Mulcahy. 2004. Immobilized cofactor derivatives for kinetic-based
enzyme capture strategies: direct coupling of NAD(P)+. Anal. Biochem. 335:316–325.

43.

Okuda, K., I. Urabe, and H. Okada. 1985. Synthesis of poly(ethylene glycol)-bound
NADP by selective modification at the 6-amino group of NADP. Eur. J. Biochem.
151:33–38.

44.

Ottolina, G., G. Carrea, S. Riva, and A. Bückmann. 1990. Coenzymatic properties of
low molecular-weight and macromolecular N6-derivatives of NAD+ and NADP+ with
dehydrogenases of interest for organic synthesis. Enzyme Microb. Technol. 12:596–602.

45.

Pagliaro, M., R. Ciriminna, H. Kimura, M. Rossi, and C. Della Pina. 2007. From
glycerol to value-added products. Angew. Chem., Int. Ed. 46:4434–4440.

46.

Parellada, J., A. Narváez, M. A. López, E. Domı́nguez, J. J. Fernández, V. Pavlov,
and I. Katakis. 1998. Amperometric immunosensors and enzyme electrodes for
environmental applications. Anal. Chim. Acta 362:47–57.

47.

Parhi, S. S. L. 1986. Substituted 2-mercapto-imidazoles and their preparation. United
States Patent:4,584,383.

48.

Persson, B., J. Hedlund, and H. Jörnvall. 2008. Medium- and short-chain
dehydrogenase/reductase gene and protein families: The MDR superfamily. Cell. Mol.
Life Sci. 65:3879–3894.

49.

Radianingtyas, H., and P. C. Wright. 2003. Alcohol dehydrogenases from thermophilic
and hyperthermophilic archaea and bacteria. FEMS Microbiol. Rev. 27:593–616.

50.

Reed, J. L., T. D. Vo, C. H. Schilling, and B. O. Palsson. 2003. An expanded genomescale model of Escherichia coli K-12 (iJR904 GSM/GPR). Genome Biol. 4:R54.

51.

Rocha-Martin, J., A. Acosta, J. Berenguer, J. M. Guisan, and F. Lopez-Gallego.
2014. Selective oxidation of glycerol to 1,3-dihydroxyacetone by covalently immobilized
glycerol dehydrogenases with higher stability and lower product inhibition. Bioresour.
Technol. 170:445–453.

52.

Ruzheinikov, S. N., J. Burke, S. Sedelnikova, P. J. Baker, R. Taylor, P. A. Bullough,
N. M. Muir, M. G. Gore, and D. W. Rice. 2001. Glycerol dehydrogenase: Structure,
specificity, and mechanism of a family III polyol dehydrogenase. Structure 9:789–802.

53.

Samec, Z., W. T. Bresnahan, and P. J. Elving. 1982. Theoretical analysis of
eletrochemical reactions involving two successive one-electron transfers with
dimerization of intermediate-applicaiton to NAD+/NADH redox couple. J. Electroanal.
Chem. 133:1–23.

31

54.

Schmakel, C. O., K. S. V. Santhanam, and P. J. Elving. 1975. Nicotinamide adenine
dinucleotide (NAD+) and related compounds. electrochemical redox pattern and allied
chemical behavior. J. Am. Chem. Soc. 97:5083–5092.

55.

Shaked, Z. e., and G. M. Whitesides. 1980. Enzyme-catalyzed organic synthesis:
NADH regeneration by using formate dehydrogenase. J. Am. Chem. Soc. 102:7104–
7105.

56.

Song, S. H., N. Ahluwalia, Y. Leduc, L. T. Delbaere, and C. Vieille. 2008.
Thermotoga maritima TM0298 is a highly thermostable mannitol dehydrogenase. Appl.
Microbiol. Biotechnol. 81:485–495.

57.

Spencer, P., K. J. Brown, M. D. Scawen, T. Atkinson, and M. G. Gore. 1989.
Isolation and characterization of the glycerol dehydrogenase from Bacillus
stearothermophilus. Biochim. Biophys. Acta 994:270-279.

58.

Srinivasan, V., K. S. Ma, M. W. W. Adams, M. G. Newton, J. P. Rose, and B. C.
Wang. 2002. Towards the crystal structure of glycerol dehydrogenase from Thermotoga
maritima. Acta Crystallog. D. D58:867–869.

59.

Studer, M. 2003. Production of optically active -hydroxyacetals. United States
Patent:6,660,890.

60.

Tang, C. T., F. E. Ruch, and E. C. C. Lin. 1979. Purification and properties of a
nicotinamide adenine dinucleotide-linked dehydrogenase that serves an Escherichia coli
mutant for glycerol catabolism J. Bacteriol. 140:182–187.

61.

Tishkov, V. I., and V. O. Popov. 2006. Protein engineering of formate dehydrogenase.
Biomol. Eng. 23:89–110.

62.

Tynan, J., J. Forde, M. McMahon, and P. Mulcahy. 2000. Synthesis of a highly
substituted N6-linked immobilized NAD+ derivative using a rapid solid-phase modular
approach: suitability for use with the kinetic locking-on tactic for bioaffinity purification
of NAD+-dependent dehydrogenases. Protein Expression Purif. 20:421–434.

63.

van der Donk, W. A., and H. Zhao. 2003. Recent developments in pyridine nucleotide
regeneration. Curr. Opin. Biotechnol. 14:421–426.

64.

Vieille, C., and G. J. Zeikus. 2001. Hyperthermophilic enzymes: sources, uses and
molecular mechanisms for thermostability. Microbiol. Mol. Biol. Rev. 65:1–43.

65.

Vrtis, J. M., A. K. White, W. W. Metcalf, and W. A. van der Donk. 2002. Phosphite
dehydrogenase: a versatile cofactor-regeneration enzyme. Angew. Chem., Int. Ed.
41:3257–3259.

66.

Wang, Z., M. Etienne, F. Quilès, G.-W. Kohring, and A. Walcarius. 2012. Durable
cofactor immobilization in sol-gel bio-composite thin films for reagentless biosensors and
bioreactors using dehydrogenases. Biosens. Bioelectron. 32:111–117.

67.

Weckbecker, A., H. Groger, and W. Hummel. 2010. Regeneration of nicotinamide
coenzymes: principles and applications for the synthesis of chiral compounds. Adv.
Biochem. Engin./Biotechnol. 120:195–242.

32

68.

Wichmann, R., and D. Vasic-Racki. 2005. Cofactor regeneration at the lab scale. Adv.
Biochem. Engin./Biotechnol. 92:225–260.

69.

Wollenberger, U., R. Spricigo, S. Leimkühler, and K. Schröder. 2008. Protein
electrodes with direct electrochemical communication. Adv. Biochem.
Engin./Biotechnol. 109:19–64.

70.

Woodyer, R., W. A. van der Donk, and H. Zhao. 2003. Relaxing the nicotinamide
cofactor specificity of phosphite dehydrogenase by rational design. Biochemistry
42:11604–11614.

71.

Wu, X., F. Zhao, J. R. Varcoe, A. E. Thumser, C. Avignone-Rossa, and R. C. T.
Slade. 2009. Direct electron transfer of glucose oxidase immobilized in an ionic liquid
reconstituted cellulose-carbon nanotube matrix. Bioelectrochemistry 77:64–68.

72.

Wykes, J. R., P. Dunnill, and M. D. Lilly. 1972. The preparation of soluble high
molecular weight NAD derivative active as a cofactor. Biochim. Biophys. Acta 286:260–
268.

73.

Yamada-Onodera, K., N. Kawahara, Y. Tani, and H. Yamamoto. 2004. Synthesis of
optically active diols by Escherichia coli transformant cells that express the glycerol
dehydrogenase gene of Hansenula polymorpha DL-1. Engin. Life Sci. 4:413–417.

74.

Yamada-Onodera, K., A. Nakajima, and Y. Tani. 2006. Purification, characterization,
and gene cloning of glycerol dehydrogenase from Hansenula ofunaensis, and its
expression for production of optically active diol. J. Biosci. Bioengin. 102:545–551.

75.

Yang, D., J. H. Zhang, M. K. Wong, Y. C. Yip, and M. W. Tang. 1998. Process for
catalytic epoxidation of olefinic compounds, novel cyclic ketone catalysts useful in said
process. United States Patent:5,763,623.

76.

Zhang, M., C. Mullens, and W. Gorski. 2007. Co-immobilization of dehydrogenases
and their cofactors in electrochemical biosensors. Anal. Chem. 79:2446–2450.

77.

Zou, Y., H. Zhang, J. S. Brunzelle, T. W. Johannes, R. Woodyer, J. E. Hung, N.
Nair, W. A. van der Donk, H. Zhao, and S. K. Nair. 2012. Crystal structures of
phosphite dehydrogenase provide insights into nicotinamide cofactor regeneration.
Biochemistry 51:4263–4270.

33

CHAPTER 2. Characterization of Thermotoga maritima glycerol dehydrogenase for use in the
production of dihydroxyacetone in a bioelectronics system

34

This work is previously published and is reprinted with permission: Beauchamp, J., P. G.
Gross, and C. Vieille. 2014. Characterization of Thermotoga maritima glycerol
dehydrogenase for the enzymatic production of dihydroxyacetone. Appl. Microbiol.
Biotechnol. 98:7039–7050.

35

2.1 ABSTRACT
NAD-dependent Thermotoga maritima glycerol dehydrogenase (TmGlyDH) converts
glycerol into dihydroxyacetone (DHA), a valuable synthetic precursor and sunless tanning agent.
In this work, recombinant TmGlyDH was characterized to determine if it can be used to catalyze
DHA production. The pH optima for glycerol oxidation and DHA reduction at 50°C were 7.9
and 6.0, respectively. Under the conditions tested, TmGlyDH had a linear Arrhenius plot up to
80°C. TmGlyDH was more thermostable than other glycerol dehydrogenases, remaining over
50% active after 7 h at 50°C. TmGlyDH was active on racemic 1,2-propanediol, and produced
(R)-1,2-propanediol from hydroxyacetone with an enantiomeric excess above 99%, suggesting
that TmGlyDH can also be used for chiral synthesis. (R)-1,2-propanediol production from
hydroxyacetone was demonstrated for the first time in a one-enzyme cycling reaction using
glycerol as the second substrate. Negative cooperativity was observed with glycerol and DHA,
but not with the cofactor. Apparent kinetic parameters for glycerol, DHA, and NAD(H) were
determined over a broad pH range. TmGlyDH showed little activity with N6-carboxymethylNAD+ (N6-CM-NAD), an NAD+ analog modified for easy immobilization to amino groups, but
the double mutation V44A/K157G increased catalytic efficiency with N6-CM-NAD+ ten-fold.
Finally we showed for the first time that a GlyDH is active with immobilized N6-CM-NAD+,
suggesting that N6-CM-NAD+ can be immobilized on an electrode to allow TmGlyDH activity in
a system that reoxidizes the cofactor electrocatalytically.

2.1 INTRODUCTION
Dehydrogenases catalyze redox reactions that reversibly transfer an electron pair and a
proton from a cofactor such as NADH to the substrate (42). These enzymes have potential as
36

catalysts for the production of many different compounds including chiral molecules and
pharmaceutical precursors (40, 42). However, the fact that the cofactor must be supplied in
stoichiometric amounts to the reaction or be regenerated (40, 42) strongly limits current
applications of these enzymes. Several cofactor regeneration methods have been used to drive
dehydrogenase-based reactions including using another enzyme reaction and a second substrate
(40). The high cost associated with supplying a second substrate and enzyme has limited the use
of dehydrogenases to the production of high-value specialty products. Electrochemical
regeneration promises to be a low-cost NAD(H) regeneration method allowing broader use of
dehydrogenases. In the past, electrochemical regeneration has been limited by the high
overpotentials (1.09 V vs. reversible hydrogen electrode) required for activity, which results in
side reactions and electrode fouling (20). Recently polyazines have been electropolymerized onto
carbon nanotube-modified electrodes, reducing the overpotential required down to ca 28 mV vs
SHE while increasing NADH oxidation activity, thus providing a lower cost NAD+ regeneration
method (20).
The next step in the development of a bioelectrocatalytic reactor is to select a useful
enzyme and immobilize the enzyme and cofactor onto an electrode. Glycerol dehydrogenase
(GlyDH, EC 1.1.1.6), which interconverts glycerol and dihydroxyacetone (DHA) using NAD(H)
as the cofactor, could be a good enzyme for this purpose (Figure 2.1).

Figure 2.1: TmGlyDH interconverts glycerol and DHA with the NAD(H) cofactor.

37

DHA is the only FDA-approved active ingredient used in sunless tanning creams (17),
and it is used as a precursor to many valuable products (14). For example, DHA can be used to
produce 1,2-propylene glycol, which is used in antifreeze and brake fluids, and to produce
polyurethanes and polypropylene glycols (12). It is a precursor of anti-ulcer, antihypertensive,
anti-inflammatory, analgesic (30), anti-ischemic, and α-2-adrenergic receptor agonist drugs (9).
Many optically active epoxide-based drugs (44), pesticides (37), and sweeteners (3) also use
DHA as a precursor.
DHA is currently produced by Gluconobacter oxidans grown on glycerol in a large
stirred tank (14). Glycerol oxidation to DHA requires large quantities of oxygen, so high aeration
levels must be maintained using oxygen-enriched air (14). An industrial strain of G. oxidans
optimized to overcome DHA inhibition produced DHA at a rate of 3.6 g L-1 h-1 with 85% yield
in a two-step fed-batch process (2). G. oxidans has been further optimized to produce DHA with
a 96% yield at a rate ranging from 12 g L-1 h-1 (at 100 g/L glycerol) to 9.5 g L-1 h-1 (at 140 g/L
glycerol). Productivity based on cell dry weight increased by 140% (21). Although optimization
of the DHA production process has increased yields and production rates, the cost of DHA
remains relatively high because of the fed-batch process, high aeration levels required, and
downstream purification costs. An enzymatic catalytic system promises to be a good alternative
to bacterial DHA production.
Escherichia coli GlyDH was recently characterized. DHA reduction activity was
optimum at pH 8.0, while glycerol oxidation activity was optimum at pH 11 (31). The enzyme
was fairly stable for a mesophilic enzyme, maintaining 100% glycerol oxidation activity after a
60-min incubation at pH 8.0 and 60°C, but it lost over 95% of its glycerol oxidation activity at
pH 10.3, its optimum pH for glycerol oxidation. At pH 8.0, its glycerol oxidation activity was

38

also only about 10% of its maximum activity at pH 10.3. The lack of stability of E. coli GlyDH
precludes its use in a biocatalytic system, and an enzyme is needed that is stable in conditions
optimum for glycerol oxidation activity.
Enzymes originating from (hyper)thermophilic microbes are not only thermostable, but
they often show increased resistance to other conditions such as high pressure, high salt
concentrations, and high solvent concentrations (6, 32). Thus, they are likely to be more stable
than enzymes from mesophiles in conditions required for biocatalysis. Thermostable
dehydrogenases are being investigated to reduce costs associated with dehydrogenase-based
catalysis and to optimize reaction conditions for industrial applications. Part of the cost reduction
associated with thermostable enzymes comes from simple purification of overexpressed proteins
via heat treatment (34). Thermostable enzymes typically have long shelf lives. They also allow
higher operational temperatures for enzymatic processes than mesophilic enzymes do, thus
reducing mass transfer limitations, reducing viscosity, increasing substrate and product
solubility, and reducing the risk of microbial contamination (6, 39). Here we characterize the
GlyDH from the hyperthermophilic bacterium Thermotoga maritima (TmGlyDH). Our rationale
was that TmGlyDH should be more thermostable than the corresponding E. coli enzyme,
allowing longer electrode life and higher operating temperatures. TmGlyDH was previously
cloned, expressed, and crystallized, but only the zinc content of the enzyme and the specific
activity of the enzyme for glycerol oxidation at pH 8.6 were reported (4, 36). Our studies aim to
determine TmGlyDH’s potential as a catalyst for DHA production. In this chapter, forward
reaction refers to glycerol oxidation and reverse reaction refers to DHA reduction.

39

2.3 MATERIALS AND METHODS
2.3.1 Chemicals, bacterial strains, and culture conditions
All chemicals were purchased from Sigma-Aldrich (St. Louis, MO), unless stated
otherwise. T. maritima strain MSB8 (ATCC 43589) was grown as described (34) and was the
source of genomic DNA to clone gldA. E. coli strains DH5 (Invitrogen, Grand Island, NY),
BL21(DE3) (EMD chemicals, Gibbstown, NJ), and XL2 Blue (Stratagene, La Jolla, CA) were
used as cloning and plasmid propagation host, protein expression host, and transformation host
after mutagenesis, respectively. E. coli strains were routinely grown in lysogeny broth (LB) in
the presence of 25 g/ml kanamycin when needed for plasmid maintenance.

2.3.2 Cloning of the gene encoding TmGlyDH
T. maritima's genomic DNA was extracted as described (34). The TmGlyDH-encoding
gene, gldA, was amplified from T. maritima genomic DNA using primers gldAfor and gldArev
(Table S2.1). The PCR product was cloned into the pCR2.1-TOPO vector (TOPO TA Cloning
Kit, Invitrogen), sequenced, and subcloned into the NdeI and XhoI restriction sites of pET24a(+)
(Novagen, Madison, WI), yielding plasmid pTmGlyDH. From this construct, TmGlyDH is
expressed with a C-terminal His6 tag. DNA manipulations followed standard protocols (1).

2.3.3 Expression and purification of TmGlyDH
TmGlyDH was purified from a 1-L culture of BL21(DE3)(pTmGlyDH). At an OD600 of
0.8, the culture was induced with 1.2 mM isopropyl-β-D-thiogalactopyranoside. After 4 h
induction the cells were harvested by centrifugation (4,500  g, 20 min) and stored at -80°C. The

40

pellet was resuspended in 40 mL of wash buffer (50 mM sodium phosphate buffer [pH 8.0], 0.5
M NaCl) containing 60 mg lysozyme, 1,000 units DNase I, and 1,000 units RNase, and
incubated for 1 h at 25°C. The cells were sonicated on ice for six 10-s bursts on high power
using a microtip (50% duty cycle, output control 4.5, Branson Sonifier 450). The cell debris were
removed by centrifuging at 15,000  g for 30 min. Non-thermostable proteins were removed by a
20-min heat treatment at 85°C followed by cooling on ice and centrifugation at 15,000  g for 30
min. The lysate was added to a 7-mL nickel nitriloacetic acid agarose (Ni-NTA, Qiagen,
Valencia, CA) column equilibrated with wash buffer. Unbound proteins were eluted with 20
column volumes (CV) of wash buffer followed by 10 CV of wash buffer containing 20 mM
imidazole and 10 CV of wash buffer. TmGlyDH was eluted in nine 10-mL fractions with wash
buffer containing 250 mM imidazole. The purified protein was concentrated by ultrafiltration
and dialyzed 3 times against 20 mM Tris-HCl (pH 8.0) containing 250 mM NaCl and 5 mM βmercaptoethanol. Because the enzyme lost over 90% of its activity after dialysis into Tris buffer
as compared to the eluted protein in wash buffer containing 250 mM imidazole, the enzyme was
dialyzed three times against 50 mM sodium phosphate buffer (pH 8.0) containing 250 mM NaCl.
Protein concentrations were determined using the Bio-Rad protein assay kit (Bio-Rad,
Richmond, CA), with bovine serum albumin as the standard.

2.3.4 TmGlyDH activity assays
All activity assays were performed in triplicate in 1-mL volumes in a Varian Cary 300
UV/Vis spectrophotometer (Palo Alto, CA) equipped with a Peltier heating system. Standard
glycerol oxidation assays were performed at 80°C with 100 mM buffer (either phosphate or 2[cyclohexylamino]ethanesulfonic acid [CHES]) adjusted to the desired pH, 678 mM glycerol,
41

500 μM NAD+, and 6.7 g TmGlyDH. Activity assays with alternative substrates were
performed in 50 mM phosphate buffer (pH 7.4) containing 300 μM NAD+, 33.5 g enzyme, and
500 mM glycerol, 1,2-propanediol, 1,2-butanediol, ethylene glycol, isopropanol, 1,3propanediol, 1-propanol, or glycolic acid. For all assays, water, buffer, and substrate were heated
in the cuvette at 80°C for 15 min, the cofactor was added, and the initial absorbance was
monitored at 340 nm for 20 s. The reaction was started by adding 10 μL enzyme. The reaction
was monitored for 20 s, and initial velocity was calculated using the initial linear portion of the
slope. The extinction coefficient for NADH (6,220 M-1 cm-1 at 340 nm) was used to calculate
activity.
The effect of temperature on TmGlyDH kinetic stability was determined by incubating
6.7 g TmGlyDH in 50 mM phosphate buffer (pH 7.4) for increasing amounts of time at 50°C
and 80°C, and measuring the remaining activity at 50°C by adding 200 mM DHA and 300 μM
NADH. The samples heated at 80°C were cooled on ice for at least 30 min, then preheated at
50°C for 15 min before measuring remaining activity at 50°C. The effect of DHA on TmGlyDH
stability was measured similarly by adding 200 mM DHA to the incubation solution. The effect
of temperature on activity was determined by measuring TmGlyDH activity at temperatures
ranging from 30°C to 90°C with 200 mM DHA, 200 μM NADH, and 3.35 g TmGlyDH. The
effect of KCl and NH4Cl on DHA reduction was measured at 50°C by adding increasing
concentrations of KCl or 100 mM NH4Cl to the assay mixture. The effect of KCl and NH4Cl on
glycerol oxidation was measured by adding 100 mM of either salt to the assay mixture.

42

2.3.5 Determination of TmGlyDH’s apparent kinetic parameters
Apparent kinetics of TmGlyDH for NAD+ were measured at 50°C to provide kinetic data
at a temperature likely to be used to operate a bioelectronic reactor. Reaction mixtures contained
50 mM phosphate buffer (pH 6.0‒7.9), 678 mM glycerol, 3.35 g TmGlyDH, and 5 µM to 300
µM NAD+. Apparent kinetics of TmGlyDH for NADH were measured as described for NAD+,
except that 200 mM DHA and 5 µM to 200 µM NADH were substituted for glycerol and NAD+,
respectively, and that 50 mM phosphate (pH 6.0-7.9) or CHES buffer (pH 8.0-9.0) was used.
TmGlyDH’s apparent kinetic parameters for DHA were measured with 50 mM phosphate or
CHES buffer, 200 µM NADH, 3.35 g TmGlyDH, and 0.05 mM to 300 mM DHA. TmGlyDH
apparent kinetic parameters for glycerol were measured with 50 mM phosphate or CHES buffer,
300 µM NAD+, 3.35 g TmGlyDH, and 10 mM to 810 mM glycerol. Apparent kinetic
parameters for NAD+ and NADH were calculated using a non-linear fit of the Michaelis-Menten
equation in Origin 8.0 (Originlab, Northampton, MA). Apparent kinetic parameters for DHA and
glycerol were calculated using a non-linear fit to the Hill equation (equation 2.1). The Hill
equation adequately describes cooperative kinetics, and it allows the determination of Vmax and
K0.5. (K0.5 is similar to the Km, expressing the concentration of substrate required to produce half
of the Vmax.) The Hill coefficient, n, describes the degree of cooperativity. A value of n above 1
indicates positive cooperativity, whereas n below 1 indicates negative cooperativity. MichaelisMenten kinetics apply when n=1 (8). Apparent kinetics of TmGlyDH for N6-carboxymethylNAD+ (N6-CM-NAD+) were measured with 50 mM phosphate buffer (pH 7.4), 810 mM
glycerol, 33.5 μg TmGlyDH, and 40 μM to 2,000 µM N6-CM-NAD+.

43

Eq. 2.1
𝑉=

𝑉𝑚𝑎𝑥 ∗ (𝑆)𝑛
𝑛
𝐾0.5
+ (𝑆)𝑛

2.3.6 N6-CM-NAD+ synthesis and purification
N6-CM-NAD+ was synthesized from 1 g NAD+ as described (11, 22) with the following
modifications: (i) N1-CM-NAD+ was filtered on Whatman #1 filter paper and washed twice each
time with cold acetone and ether; (ii) in the N1-CM-NAD+ reduction step, the solution was
deaerated with N2 gas for 5 min instead of 2 min; and (iii) precipitated N6-CM-NAD+ was
washed twice each with cold acetone and ether. N1-CM-NAD+ was found to be unstable when
stored dissolved in water. Molar yield and molecular weight were calculated using the extinction
coefficient 19,000 M-1 cm-1 (22) for N6-CM-NAD+ and the measured mass.
N6-CM-NAD+ was dissolved in water and one third of the volume was loaded onto a
Dowex 1X2 Cl-form column (2.5 cm  35 cm) that had been washed with 2 CV of 3 M HCl
followed by deionized water until the pH was above 4.0. The column loaded with N6-CM-NAD+
was washed with 2 CV of H2O, and then 2 CV of 5 mM CaCl2 (pH 2.7). N6-CM-NAD+ was
eluted with a linear 5 mM (pH 2.7) to 100 mM (pH 2.0) CaCl2 gradient over 13 CV. The
fractions containing N6-CM-NAD+ as measured by OD266 were lyophilized, dissolved in a
minimum amount of H2O, desalted in 2-mL batches on a sephadex G10 column (2.5 cm  23
cm), lyophilized again, and stored at -20°C.

44

2.3.7 Site-directed mutagenesis of TmGlyDH
Point mutations were introduced in T. maritima gldA with plasmid pTmGlyDH as the
template using the QuikChange® (Stratagene, La Jolla, CA) protocol. Single mutations V44A,
I154A, K157N, and K157G as well as double mutations I154A/K157N, V44A/K157G,
V44A/K157N, V44A/I154A, and I154A/K157G, plus the triple mutation V44A/I154A/K157G
were created with the primers listed in Table S2.1. PCR reactions were transformed into
electrocompetent XL2 Blue. Mutated genes were sequenced, and the mutated plasmids were
transformed into BL21(DE3) for protein expression as described above.

2.3.8 Determination of 1,2-propanediol enantiomeric excess
Twenty mM hydroxyacetone was converted to (R)-1,2-propanediol at 50°C with NADH
recycling in a 10-mL reaction containing 50 mM phosphate buffer (pH 7.4), 600 mM glycerol,
200 μM NADH, and 0.0153 mg/mL TmGlyDH. A 20 mM racemic mixture of (R) and (S) 1,2propanediol was resolved by removing (R)-1,2-propanediol with NAD+ recycling at 50°C in a
10-mL reaction containing 400 mM DHA, 50 mM phosphate buffer (pH 7.4), 200 μM NADH,
and 0.0153 mg/mL TmGlyDH. Reaction samples were diluted 1:10 in acetonitrile, incubated on
ice for 20 min, centrifuged at 15,000  g for 10 min, and analyzed using an Agilent GC-MS
(6890N GC coupled to a 5973 Inert Mass Selective Detector) equipped with a CP-Chirasil-Dex
CB column (25 m, 0.25 mm inner diameter) as described (43). The enantiomeric excess (ee) was
calculated using the peak areas of the enantiomers. If no peak was observed, the amount present
was assumed to be equal to the limit of detection for the GC-MS, which was determined as
described (41).

45

2.3.9 DHA stability assays
DHA stability was assayed in 50 mM phosphate buffer at 50°C for up to 6 h at pH 6.4,
7.4, and 7.9. Samples collected every 90 min were analyzed on a Breeze high-performance liquid
chromatograph (HPLC, Waters, Milford, MA) equipped with an Aminex-87C carbohydrate
analysis column (Bio-Rad, Hercules, CA). The molecules were separated at 80°C using water as
the mobile phase. DHA and glyceraldehyde were detected with a Waters 2478 Dual λ
Absorbance UV detector at 210 nm.

2.3.10 Activity assays with N6-CM-NAD+ immobilized on Sepharose beads
N6-CM-NAD+ was immobilized on cyanogen bromide-activated Sepharose 4b as
described (38) using bis(3-aminopropyl)amine (B3APA) as the linker, yielding NAD-CMB3APA-Sepharose. B3APA loading on the Sepharose was quantified using 2,4,6-trinitrobenzene
sulfonic acid using the G-Biosciences protocol (Cat # BC86). N6-CM-NAD+ binding to B3APASepharose was quantified by phosphorus analysis after digestion with H2SO4 and H2O2 (26).
The enzyme cycling reaction was performed in 50 mM sodium phosphate (pH 7.4) containing
200 mM glycerol, 100 mM hydroxyacetone, 37‒41 mg wet NAD-CM-B3APA-Sepharose, and
0.345 mg TmGlyDH in 10-mL. Ethylene diamine (ED) and diethylene triamine (DETA) were
also used as linkers following the above protocols. The amounts of NAD-ED-Sepharose and
NAD-DETA-Sepharose used in cycling reactions were calculated to give the same NAD
concentration as that present in the NAD-B3APA-Sepharose reactions. DHA and 1,2propanediol produced were quantified by GC-MS as described above. Glycerol and
hydroxyacetone consumed were quantified on a HPLC as described above except that analytes
were detected with a Waters 410 refractive index detector.
46

2.4 RESULTS
2.4.1 Expression, purification, and characterization of TmGlyDH
TmGlyDH purified using a Tris buffer system gave high yields of pure enzyme (72
mg/L) with very low glycerol oxidation activity (3 ± 2 µmol min-1 mg-1 at 80°C, pH 7.9),
compared to the enzyme assayed in N-(2-hydroxyethyl)piperazine-N-(3-propanesulfonic acid)
(96 µmol min-1 mg-1 at pH 8.6) (36). Dialyzing TmGlyDH against phosphate buffer increased
activity to 45 ± 4 µmol min-1 mg-1 at 80°C (pH 7.9). TmGlyDH inhibition by Tris was likely due
to the structural similarity of Tris to glycerol. Tris is a four-carbon molecule with a primary
amine on the second carbon. Its three alcohol groups (Figure S2.1) make its structure look very
similar to that of glycerol and likely allow it to act as a competitive inhibitor of GlyDH. Bicine
and glycine buffers also inhibited enzyme activity (Table S2.2). GlyDH from
Schizosaccharomyces pombe is also reported to be inhibited by Tris (24). TmGlyDH showed 45
± 2 µmol min-1 mg-1 glycerol oxidation activity (80°C, pH 7.4) when all purification steps were
performed in phosphate buffer.
Because GlyDHs from other organisms are activated by KCl and NH4Cl (24, 25, 27, 31),
TmGlyDH activity in the presence of those salts was tested. DHA reduction activity was
partially inhibited by KCl, with 60% activity retained in the presence of 40 mM KCl and 70%
activity at KCl concentrations above 100 mM (Figure S2.2). DHA reduction activity was also
inhibited by NH4Cl, decreasing by 27% in 100 mM NH4Cl. Glycerol oxidation activity increased
by 39% and 18% in the presence of 100 mM NH4Cl and 100 mM KCl, respectively. In DHA
reduction assays performed in the presence of NH4Cl, the background rate shifted from a
negative slope due to autoxidation of NADH to a positive slope 3.4-fold higher in magnitude. A
study of DHA stability found DHA to be more stable near neutral pH (pH 6.4 and 7.4) than at pH
47

7.9 (Figure S2.3). During these studies DHA degradation corresponded with an increase in
absorbance at 280 nm with a broad shoulder overlapping 340 nm (not shown). The rate of OD280
increase when 50 mM DHA was incubated in 50 mM phosphate buffer (pH 7.4) at 50°C was 3fold higher in the presence of 100 mM NH4Cl than in its absence (not shown). Thus, we
concluded that NH4Cl likely increases the degradation rate of DHA. For this reason, no further
studies with NH4Cl were performed because its presence would reduce DHA yields in a catalytic
reactor.
Table 2.1: TmGlyDH’s oxidizing activity on glycerol-related substrates at 80°C, pH 7.4.
Results shown are the average ± standard deviation of three independent experiments.
Substrate

Specific activity

Activity

(μmol min-1 mg-1)

(%)

Glycerol

49 ± 2

100

rac-1,2-propanediol

23 ± 1

47

rac-1,2-butanediol

17 ± 3

35

3.5 ± 0.1

7

Isopropanol

0.29 ± 0.05

0.6

1,3-propanediol

0.22 ± 0.03

0.4

0.0

0

Ethylene glycol

1-propanol

While TmGlyDH exhibited significant activity on 1,2-propanediol and 1,2-butanediol, it
was less active on ethylene glycol, and barely active to completely inactive on the other
substrates tested (Table 2.1). Since enzymes typically are only active on one enantiomer of a
racemate, batch reactions were set up to determine which 1,2-propanediol enantiomer is
48

preferred. TmGlyDH converted 20 mM hydroxyacetone to 5.9 ± 0.3 mM (R)-1,2-propanediol in
180 min with ee greater than 99%. A 20 mM racemic mixture of (R and S) 1,2-propanediol was
resolved leaving the (S)-1,2-propanediol with an ee of 46% after 360 min.
2.4.2 Effect of temperature on TmGlyDH stability and activity
TmGlyDH was stable when stored at -80°C, maintaining 97% of its activity after at least
three months at -80°C. When incubated at 50°C, TmGlyDH linearly lost 42% of its activity over
7 h (R2 = 0.9731, Figure 2.2). At 80°C, TmGlyDH inactivation appeared to follow bi-exponential
decay (Figure 2.2). The enzyme lost 66% of its activity after 20 min and 80% of its activity after
120 min. At 50°C TmGlyDH was initially more stable in the presence of DHA than in its
absence, maintaining 100% activity over 120 min, but it quickly inactivated after 120 min, losing
over 84% of its activity in the next 5 h (Figure 2.2). The effect of temperature on TmGlyDH’s
DHA reduction activity is shown in Figure 2.3. The Arrhenius plot (Figure 2.3 inset) was linear
between 30°C and 80°C (R2 = 0.999) with an apparent activation energy of 45.7 ± 0.7 kJ/mol.

49

Figure 2.2: Inactivation of TmGlyDH at 50°C (A) and 80°C (B). Open circles: inactivation in
the absence of DHA; Closed circles and dashed lines: inactivation in the presence of 200 mM
DHA. The linear regression is shown for TmGlyDH inactivation at 50°C in the absence of DHA,
starting at 1 h. In panel B, the best fit for enzyme inactivation was to the biexponential decay
equation y = 0.44e-0.0067x + 0.56e-0.133x, with R2 = 0.985. Results shown are the average ±
standard deviation of three independent experiments.

1.4

A

Relative activity

1.2

1
0.8
0.6
0.4
0.2

0
0

60

120

180
240
300
Incubation time (min)

360

B

1.0

Relative activity

420

0.8

0.6

0.4

0.2

0.0
0

25

50

75

100

Incubation time (min)

50

125

150

Figure 2.3: Effect of temperature on TmGlyDH’s DHA reduction activity at pH 7.4. Inset:
Arrhenius plot. Results shown are the average ± standard deviation of three independent
experiments.

2.4.3 TmGlyDH apparent kinetic parameters
The Vmax of TmGlyDH was larger for NAD+ in the glycerol oxidation reaction than for
NADH in the DHA reduction sequence at pH 7.9 (Table 2.2). The Km however, was 2–4 times
higher for NAD+ than NADH at all pH values. The initial rate velocities of TmGlyDH for DHA
and glycerol did not fit the Michaelis-Menten equation well (not shown). Because the crystal
structure shows TmGlyDH to be a tetramer (36), and because the non-linear shapes of the
Hofstee plots (Figure 2.4C and 2.4F) are characteristic of cooperative behavior, a cooperative
model was used instead. The data fit the Hill equation well (Figure 2.4). The largest Vmax for
DHA reduction was at pH 6.0, while that for glycerol oxidation was at pH 7.9 (Table 2.3). The
apparent Vmax for glycerol oxidation was 1.8-fold higher than the apparent Vmax for DHA
reduction at their respective pH optima. The K0.5 for glycerol was significantly higher than the
K0.5 for DHA at all the pH values tested. At pH 7.9, the Vmax for DHA reduction was half the

51

Vmax for glycerol oxidation and the K0.5 for glycerol was twice the K0.5 for DHA (Table 2.3).
Because an initial screen of apparent Vmax values vs pH for NAD+ indicated that the pH optimum
was pH 7.4, all other experiments performed at a single pH value were performed at pH 7.4.

52

Table 2.2: Apparent kinetic parameters of TmGlyDH for NADH and NAD+ at 50°C.
pH
6.0a
6.4a
6.9a
7.4a

7.9a

8.5b

9.0b

NADH
Vmax (μmol min-1 mg-1)

ND

11.3 ± 0.3

13.1 ± 0.5

12.2 ± 0.3

13.4 ± 0.6

11.9 ± 0.6

9.8 ± 0.3

Km (μM)

ND

14 ± 2

18 ± 2

17 ± 2

20 ± 3

41 ± 7

49 ± 4

Vmax (μmol min-1 mg-1)

2.5 ± 0.3

4.1 ± 0.2

7.3 ± 0.3

11.4 ± 0.3

29.8 ± 0.9

6.5 ± 0.2

ND

Km (μM)

80 ± 20

62 ± 10

59 ± 7

31 ± 3

51 ± 5

78 ± 7

ND

NAD+

a

Phosphate buffer

b

CHES buffer

ND: Not determined

53

Table 2.3: Apparent kinetic parameters of TmGlyDH for DHA and glycerol at 50°C based on the Hill equation (Eq. 1).
pH
6.0a
6.4a
6.9a
7.4a
7.9a
8.5b
9.0b
DHA
Vmax (μmol min-1 mg-1)

13.4 ± 0.7

10.2 ± 0.4

9.7 ± 0.5

11.4 ± 0.8

11 ± 1

7.7 ± 0.9

7.4 ± 0.8

30 ± 5

13 ± 2

12 ± 2

11 ± 3

19 ± 8

19 ± 9

12 ± 5

0.81 ± 0.06

0.85 ± 0.08

0.71 ± 0.07

0.59 ± 0.07

0.55 ± 0.07

0.56 ± 0.08

0.7 ± 0.2

Vmax (μmol min-1 mg-1)

3.0 ± 0.8

7.7 ± 1.58

7.0 ± 1.9

10.8 ± 1.2

24.0 ± 1.4

6.0 ± 0.3

ND

K0.5 (mM)

393 ± 250

611 ± 315

204 ± 161

51 ± 22

39 ± 7

87 ± 14

ND

n

0.88 ± 0.16

0.76 ± 0.07

0.71 ± 0.14

0.58 ± 0.09

0.86 ± 0.13

0.63 ± 0.04

ND

K0.5 (mM)
n
Glycerol

a

Phosphate buffer

b

CHES buffer

54

Figure 2.4: TmGlyDH’s glycerol oxidizing activity (A, B and C) and DHA reducing activity
(D, E and F) at 50°C, pH 7.4. (A and D) Initial velocity fitted to the Hill equation. (B and E)
Residuals of curve fit in A and C. (C and F) Hofstee plots of v vs. v/[glycerol] (C) and of v vs.
v/[DHA] (F). Data from triplicate experiments are shown in A, B, D, and E. Averages are shown
in C and F.
1.0

9

v (µmol min-1 mg-1)

7

B

A
0.5

6

Residuals

-1

-1

v (mol min mg )

8

5
4
3
2

0.0

-0.5

1
-1.0

0
0

200

400

0

[glycerol] (mM)

200

9
8
7
6
5
4
3
2
1
0

C

0

400

0.1

0.2

0.3

0.4

v/[glycerol] (mL min-1 mg-1)

[glycerol] (mM)

1

E

v (µmol min-1 mg-1)

D
Residuals

-1

-1

v (mol min mg )

12

10
9
8
7
6
5
4
3
2
1
0

0

-1

-2
0

100

200

[DHA] (mM)

300

0

100

200

300

[DHA] (mM)

F

10
8
6
4
2
0
0

5

10

15

v/[DHA] (mL min-1 mg-1)

2.4.4 Introduction of NAD+ into the 3D structure of TmGlyDH
In an electrochemical reactor the cofactor would ideally be bound to the surface of the
electrode. N6, S6, 8’-Azo, and direct coupling are published methods for immobilizing NAD (11,
22, 28, 29, 38). Direct coupling gives binding at a mixture of the phosphate backbone and the N6
locations (29). The N6 position was selected because studies of affinity chromatography matrices
demonstrated that N6-linked NAD could capture a wider variety of NAD-dependent enzymes
than NAD immobilized using the other methods mentioned here (11). N6-CM-NAD+ can be
linked to amino-terminal linker molecules attached onto an electrode surface through a simple
procedure of amide bond formation, and several dehydrogenases have been shown to be active
with N6-CM-NAD+ in solution (22). We introduced NAD+ into the 3D structure of TmGlyDH to

55

predict how active the enzyme would be with N6-CM-NAD+. The crystal structure of TmGlyDH
does not contain NAD+ (4, 36). TmGlyDH shares 48% sequence identity (with 1% gaps) with
BsGlyDH (protein sequences were aligned using Clustal Omega at
http://www.ebi.ac.uk/Tools/msa/clustalo/) (13, 33), and the crystal structure of BsGlyDH is
available in complex with NAD+. After superposing the two proteins structures in PyMol, the
calculated RMSD for the backbone atoms was 1.03 Å. When the active site zinc and the 27
amino acids with atoms within 4 Å of NAD were used to align the structures, the all-atom
RMSD between the two active sites was 0.70 Å, indicating that the alignment was suitable for
importing NAD+ into the TmGlyDH structure. Minimization of the TmGlyDHNAD complex
further improved the conformation of TmGlyDH-bound NAD (Figure 2.5A). The N6 amine
group in the TmGlyDHNAD+ complex appeared only slightly solvent accessible (Figure 2.5B),
thus we do not expect TmGlyDH to be highly active on N6 amine-modified analogs.

2.4.5 Kinetics of TmGlyDH on N6-CM-NAD+
N6-CM-NAD+ was synthesized from NAD+ with a purified molar yield of 13.7% and a
molecular weight of 1,114 g/mol (expected molecular weight 725 g/mol), indicating the presence
of H2O or CaCl2. TmGlyDH’s Vmax with N6-CM-NAD+ was 2% of that with NAD+, and its Km
for N6-CM-NAD+ was 15-fold higher than for NAD+ (Table 2.4). By modeling the
carboxymethyl group onto TmGlyDH-bound NAD+ with PyMol, three residues (V44, I154, and
K157) were identified that were very close to the carboxymethyl group and potentially caused
steric clashes with the NAD+ analog. Single, double, and triple mutants of V44, I154, and K157
were produced to reduce steric hindrance. Several of the TmGlyDH mutants showed increased
Vmax and lower Km values for N6-CM-NAD+ compared to the wild-type enzyme (Table 2.4).
56

Compared to TmGlyDH, the V44A/I154A mutant had a 5-fold higher Vmax, but its Km increased
8-fold; the K157G mutant had a 5-fold lower Km, but its Vmax decreased 25%. The V44A/K157G
mutant had the best catalytic efficiency (Vmax/Km) with a 4-fold higher Vmax and a 32% lower Km
than TmGlyDH.

Figure 2.5: (A) Ribbon representation of alignment of BsGlyDH (Blue) and minimized
TmGlyDH (green), (B) active site residues around NAD. The red NAD is the conformation
from BsGlyDH, the yellow NAD is the conformation from minimization of TmGlyDH. The
crystal structure of TmGlyDH (Protein Data Bank [PDB] entry 1KQ3) was structurally aligned
with 27 residues within 4 Å of NAD and the active site zinc in the 3D structure of Bacillus
stearothermophilus GlyDH (BsGlyDH, PDB entry 1JQ5) using the PyMOL Molecular Graphics
System, Version 1.5.0.4 (Schrödinger, LLC), allowing NAD+, which is present in the BsGlyDH
structure, to be imported into the TmGlyDH binding pocket. The structure of the
TmGlyDHNAD+ complex was solvated and the energy minimized at 100 K for 2,000 steps
using CHARMM, version c38a2 (5).
A

B

57

Table 2.4: Apparent kinetic parameters of TmGlyDH (WT) and its mutant derivatives on NAD+ and N6-CM-NAD+. Assays
were performed in phosphate buffer (pH 7.4) at 50°C
Enzyme

WT

K157N

K157G

V44A

I154A

V44A/ K157N V44A/ I154A V44A/K157G

12.2 ± 0.8

6.9 ± 0.3

19.0 ± 0.7

20.2 ± 0.7

36 ± 2

17.4 ± 0.5

16.6 ± 0.8

3.6 ± 0.1

24 ± 5

25 ± 4

31 ± 3

70 ± 6

88 ± 9

39 ± 3

351 ± 25

21 ± 2

0.5 ± 0.1

0.28 ± 0.05

0.61 ± 0.06

0.29 ± 0.03 0.41 ± 0.05

0.45 ± 0.04

0.047 ± 0.004

0.17 ± 0.02

0.21 ± 0.04 0.208 ± 0.007 0.61 ± 0.03 0.48 ± 0.03

0.41 ± 0.01

1.4 ± 0.2

1.11 ± 0.2

NAD+
Vmax
(μmol min-1 mg-1)
Km
(μM)
Vmax/Km
-1

-1

(L min mg )
N6-CM-NAD+
Approx. Vmax
-1

0.28 ± 0.02
-1

(μmol min mg )
Approx. Km

370 ± 50

360 ± 120

80 ± 10

150 ± 30

730 ± 80

186 ± 14

2,900 ± 500

250 ± 20

0.8 ± 0.1

0.6 ± 0.2

2.6 ± 0.3

4.1 ± 0.8

0.66 ± 0.08

2.2 ± 0.2

0.5 ± 0.1

4.4 ± 0.9

(μM)
Vmax/Km (x10-3)
-1

-1

(L min mg )

58

Table 2.4 (cont’d)

Enzyme

I154A/K157G

V44A/I154A/K157G

31.6 ± 0.8

22 ± 2

108 ± 6

600 ± 100

0.29 ± 0.02

0.037 ± 0.007

0.40 ± 0.07

0.5 ± 0.2

1,500 ± 350

2,300 ± 1200

0.027 ± 0.008

0.02 ± 0.01

NAD+
Vmax
-1

-1

(μmol min mg )
Km
(μM)
Vmax/Km
-1

-1

(L min mg )
N6-CM-NAD+
Approx. Vmax
-1

-1

(μmol min mg )
Approx. Km
(μM)
Vmax/Km (x10-3)
-1

-1

(L min mg )

59

2.4.6 Activity of TmGlyDH with N6-CM-NAD+ immobilized on Sepharose beads
As a proof of concept that TmGlyDH is active with immobilized N6-CM-NAD+,
TmGlyDH activity was measured with N6-CM-NAD+ immobilized on Sepharose beads. Loading
of the linker B3APA on Sepharose was 4.3 ± 0.9 µmol/g wet Sepharose. Loading of N6-CMNAD+ on B3APA-Sepharose was 1.3 ± 0.2 µmol/g wet Sepharose, occupying approximately one
in four linker molecules. TmGlyDH activity was measured by cycling NAD-CM-B3APASepharose in two reactions catalyzed by the enzyme: glycerol oxidation to DHA and
hydroxyacetone reduction to (R)-1,2-propanediol. 13.6 ± 1.0 mM glycerol and 16.5 ± 0.9 mM
hydroxyacetone were consumed and 7.3 ± 0.2 mM DHA and 11.5 ± 0.5 mM (R)-1,2-propanediol
were produced after 22.3 h (Figure 2.6). Some of the DHA produced could have reacted with the
free amines on B3APA-Sepharose, explaining why less DHA was detected than (R)-1,2propanediol. Cycling reactions using N6-CM-NAD+ immobilized with the shorter DETA and ED
linkers produced 11.2 ± 0.1 mM and 10.72 ± 0.01 mM (R)-1,2-propanediol, respectively,
indicating that the length of the linker had no effect on catalysis. TmGlyDH mutant derivatives
V44A/K157G, V44A, and K157G tested with NAD-CM-ED-Sepharose produced 1.5 ± 0.2 mM,
6.3 ± 0.7 mM, and 10.8 ± 0.7 mM of (R)-1,2-propanediol, respectively, after 22.5 h.

2.5 DISCUSSION
TmGlyDH was expressed at high levels in E. coli and purified to high purity.
TmGlyDH’s glycerol oxidation activity compared well to that of other characterized GlyDHs. In
the absence of added cations, TmGlyDH showed higher glycerol oxidizing specific activity at
80°C (49 ± 2 µmol min-1 mg-1) than the GlyDHs from S. pombe at 24°C (Vmax = 4.8 µmol min-1
mg-1) (24), K. pneumoniae at 27°C (Vmax = 13.5 µmol min-1 mg-1) (25), B. stearothermophilus at
60

30°C (specific activity = 25 µmol min-1 mg-1) (35), and E. coli at 30°C (Vmax = 33 µmol min-1
mg-1) (31). TmGlyDH glycerol oxidation activity increased moderately in the presence of
NH4Cl, but the added salt increased the rate of DHA degradation and was deemed unsuitable for
a reactor process producing DHA from glycerol.

Figure 2.6: Activity of TmGlyDH with N6-CM-NAD+ immobilized on B3APA-Sepharose.
(■): glycerol; (●): hydroxyacetone; (□): DHA; (○): R-1,2-propanediol.

TmGlyDH’s high glycerol oxidation activity levels, its independence from the presence
of activating cations, together with its high thermostability and its optimum activity near neutral
pH, make this enzyme an excellent candidate for DHA production. Another benefit of using
TmGlyDH is the similar pH optima for glycerol oxidation and DHA stability (Figure S2.3). In a
bioreactor, the pH could be held near pH 7 to maximize DHA stability without sacrificing much
enzymatic activity.
High activity on 1,2-propanediol and high enantiomeric excess observed when producing
(R)-1,2-propanediol from hydroxyacetone indicate that TmGlyDH could also be a useful catalyst
for producing optically active products. Synthesis of optically active 1,2-propanediol from
61

hydroxyacetone has been reported using resting E. coli cells expressing GlyDH and glucose
dehydrogenase with glucose in the reaction mixture to recycle NAD+ (43). Here we have
demonstrated for the first time a one-enzyme, cell-free system for the production of (R)-1,2propanediol using glycerol (a cheap byproduct of biodiesel production) as the recycling
substrate. In optimized reaction conditions, TmGlyDH could also be a useful catalyst for
resolving a racemic mixture of 1,2-propanediol leaving the (S)-enantiomer. Similar enantiomer
preference is expected for 1,2-butanediol.
All characterized GlyDHs have pH optima near 8‒12 and 4‒8 for the forward and reverse
directions, respectively (Table S2.3) (24, 25, 27, 31, 35). Which makes sense because the
oxidation of glycerol releases a proton to solution. TmGlyDH was similar to other GlyDHs in its
pH requirements for DHA reduction, but it required a lower pH than other GlyDHs for optimal
glycerol oxidation (Table S2.3). All the residues located within 10 Å (approximately 3-fold the
length of a hydrogen bond) of the TmGlyDH active site Zn2+ were identical or chemically
conserved (not shown) in the sequences of the B. stearothermophilus, E. coli, K. pneumoniae,
and S. pombe enzymes, suggesting that the pKAvalues of the catalytic residues are in similar
ranges in all these enzymes. Note that TmGlyDH differs kinetically from other GlyDHs. At
around 30 µM, TmGlyDH’s Km for NAD+ was over 25-fold lower than those of the E. coli and
K. pneumoniae enzymes (around 800 µM) (25, 31).
The cooperative behavior of TmGlyDH can be explained by the fact that the enzyme is a
homotetramer. The E. coli enzyme is an octameric enzyme, also showing negative cooperativity
for DHA, but no cooperativity for glycerol, NAD+, or NADH, except that it shows positive
cooperativity for NADH in the presence of KCl (31). Cooperativity has also been observed in
other enzymes involved in glycerol metabolism such as glyceraldehyde-3-phosphate

62

dehydrogenase and 1,3-propanediol dehydrogenase (15, 23). Cooperativity arises when substrate
binding to one subunit affects the affinity of the other subunits for that substrate. It can also be
the result of an enzyme following a bi bi sequential random kinetic mechanism. The steady statederived kinetic equation for the random mechanism contains squared terms (not shown) that, in
themselves, can lead to apparent cooperative behavior (10, 16). Cooperative behavior is
widespread with about 25% of all multi-subunit enzymes showing cooperativity, and
approximately half of those showing negative cooperativity (19). Although not well understood,
negative cooperativity is thought to help regulate homeostasis. Negatively cooperative enzymes
have been found to correlate with branch-point enzymes, and simulations have demonstrated that
metabolic pathways can be better regulated by cooperative enzymes (7, 19). Cooperativity makes
kinetic characterization more difficult because the enzymes do not follow Michaelis-Menten
kinetics. To allow comparison to Michaelis-Menten parameters of other enzymes, TmGlyDH’s
apparent kinetic parameters were determined from the Hill equation. Although the Hill equation
is purely empirical, the parameters given provide information similar to the Michaelis-Menten
equation. Mathematical modeling of a reactor system that includes NADH regeneration on the
electrode, TmGlyDH’s kinetic mechanism and corresponding parameters, as well as diffusion
parameters (18) is in progress (Li et al., in progress).
A bioelectrocatalytic reactor will require specific conditions for optimal performance.
The properties of TmGlyDH determined here will allow reactor conditions to be selected for
optimal enzyme activity as well as for optimizing cofactor and enzyme immobilization
conditions. The broad, near-neutral pH range for glycerol oxidation activity is beneficial because
it allows the pH conditions to be more broadly determined by other reactor constraints such as
electron transfer to the cofactor, DHA stability, or the immobilization interface stability. Ideally,

63

the cofactor and enzyme will be immobilized on the electrode of a bioelectrocatalytic system.
The cofactor should be immobilized on the electrode first, and the enzyme would then be
immobilized by affinity for the cofactor. The entire complex could be immobilized by
crosslinking. TmGlyDH had similar Km values for NAD+ and NADH (Table 2.2), indicating that
either cofactor form would be suitable for immobilizing the enzyme on an electrode.
Liver alcohol dehydrogenase, lactate dehydrogenase (beef heart, type III), and malate
dehydrogenase (pig heart) have been shown to have 55%, 65%, and 75% activity, respectively,
with soluble N6-CM-NAD+ compared to their activity with NAD+, and immobilized NAD+
analogs have been shown to be reducible by yeast alcohol and aldehyde dehydrogenases,
suggesting that N6-immobilized NAD+ can be used with a number of different dehydrogenases
(22, 38). TmGlyDH had very low activity with N6-CM-NAD+, but mutant V44A/K157G’s
catalytic efficiency on the NAD+ analog was 10-fold higher. The next step towards building a
bioelectrocatalytic system will be to immobilize N6-CM-NAD+ on an electrode in a manner that
allows enzyme binding and activity. Our proof-of-concept immobilization experiments indicate
that TmGlyDH is active with N6-CM-NAD+ immobilized on B3APA-, ED-, and DETASepharose. The yields of DHA and (R)-1,2-propanediol were low in these first experiments, but
improving the interface by optimizing the chemical properties (e.g., hydrophobicity) of the linker
and optimizing N6-CM-NAD+ loading on the linker are beyond the scope of this study. An
unexpected result was that the V44A/K157G and V44A mutants were less active with the
immobilized cofactor than the wild-type enzyme. These results suggest that, although
convenient, screening mutants with NAD+ or NAD+ analogs in solution is not necessarily a good
predictor of activity with the immobilized cofactor.

64

TmGlyDH is a promising catalyst for the production of DHA from glycerol. It is as active
and more stable than other known GlyDHs. The broad, near neutral, pH range for TmGlyDH
activity will be useful for optimizing other reactor constraints. TmGlyDH double mutant
V44A/K157G had a ten-fold increased catalytic efficiency with N6-CM-NAD+. Finally we
showed for the first time that a GlyDH is active with immobilized N6-CM-NAD+.

2.6 ACKNOWLEDGEMENTS
We thank Dr. Hanzi Li for providing expertise in the use of the IGOR program. We thank the
Michigan State University Mass Spectrometry Core facility for providing the GC-MS, helping
with experimental design and instrument specific training. We thank Dr. Michael Feig for
helping with the energy minimization of the TmGlyDH*NAD complex, and we thank Drs. Scott
Calabrese Barton and R. Mark Worden for constructive discussions. This work was supported by
grants from the National Science Foundation grant (no. CBET 0756703) and the United States
Department of Agriculture Cooperative State Research, Education, and Extension Service
National Research Initiative (no. 2008-35504-04611), and by Michigan State University start up
funds.

65

2.7 SUPLEMENTARY MATERIALS

66

Table S2.1: Primers used for Chapter 2.
Primer

Sequence (5' to 3')

gldAfor

GCCATATGATAACAACCACCATATTTCC (NdeI site underlined)

gldArev

GCCTCGAGAGTGAGGTTTTTTCTCATCC (XhoI site underlined)

V44Afor

GATTTTGTGGACAAAAACGCGCTAGGAGAAAATTTCTTCAGC

V44Arev

GCTGAAGAAATTTTCTCCTAGCGCGTTTTTGTCCACAAAATC

I154Afor

CTGGTAGACACGGAGGCTGTGGCGAAAGCCCCC

I154Arev

GGGGGCTTTCGCCACAGCCTCCGTGTCTACCAG

K157Nfor

CGGAGATTGTGGCGAATGCCCCCGCGAGGTTTC

K157Nrev

GAAACCTCGCGGGGGCATTCGCCACAATCTCCG

K157Gfor

CGGAGATTGTGGCGGGAGCCCCCGCGAGGTTTC

K157Grev

GAAACCTCGCGGGGGCTCCCGCCACAATCTCCG

I154A/K157Nfor

CTGGTAGACACGGAGGCTGTGGCGAATGCCCCC

I154A/K157Nrev

GGGGGCATTCGCCACAGCCTCCGTGTCTACCAG

I154A/K157Gfor

CTGGTAGACACGGAGGCTGTGGCGGGAGCCCCC

I154A/K157Grev

GGGGGCTCCCGCCACAGCCTCCGTGTCTACCAG

Table S2.2: Relative effects of different buffers on TmGlyDH activity.
Buffer
Activity (%)
Phosphate

100

Tris

4

Bicine

71

Glycine

49

67

Table S2.3: pH optima for different glycerol dehydrogenases.
Species

DHA reduction

Glycerol oxidation

Reference

T. maritima

6.0

7.9

This study

E. coli

7‒8

10‒11

(31)

K. pneumoniae

4‒8

8‒9

(25)

S. pombe

6

10‒12

(24)

Cellulomonas sp. NT3060

6

9

(27)

Figure S2.1: Structure of glycerol and Tris

Figures S2.2: TmGlyDH activity for DHA reduction in the presence of increasing KCl
concentrations. Results shown are the average ± standard deviation of three independent
experiments.
1.2

Relative activity

1
0.8
0.6
0.4
0.2
0
0

100

200

300

400

KCl concentration (mM)

68

500

6

20

5

DHA (mM)

15

4
3

10

2
1

5

0
0

-1
0

60

120

180

240

300

Incubation time at 50 C (min)

69

360

Glyceraldehyde and other products
(mM)

Figure S2.3: DHA stability at 50°C in phosphate buffer. Circles and uninterrupted lines: pH
6.4; squares and dashed lines: pH 7.4; triangles and dotted lines, pH 7.9; Open symbols: DHA;
white-filled symbols: glyceraldehyde; and gray-filled symbols: other DHA degradation products.
At time t, other degradation products are calculated as
[DHA]initial – [DHA]t – [glyceraldehyde]t. Results shown are the average ± standard deviation
of three independent experiments.

2.8 PERMISSIONS

70

SPRINGER LICENSE
TERMS AND CONDITIONS
Apr 10, 2015

This is a License Agreement between Justin L Beauchamp ("You") and Springer ("Springer")
provided by Copyright Clearance Center ("CCC"). The license consists of your order details, the
terms and conditions provided by Springer, and the payment terms and conditions.
All payments must be made in full to CCC. For payment instructions, please see
information listed at the bottom of this form.
License Number
3600491122584
License date
Apr 01, 2015
Order Content Publisher
Springer
Order Content Publication
Applied Microbiology and Biotechnology
Characterization of Thermotoga maritima glycerol
Order Content Title
dehydrogenase for the enzymatic production of
dihydroxyacetone
Order Content Author
Justin Beauchamp
Order Content Date
Jan 1, 2014
Volume number
98
Issue number
16
Type of Use
Thesis/Dissertation
Portion
Full text
Number of copies
6
Author of this Springer article
Yes and you are the sole author of the new work
Order reference number
None
Thermotoga Maritima Glycerol Dehydrogenase as
a Catalyst for Dihydroxyacetone Production:
Title of your thesis / dissertation
Enzyme Characterization, engineering and
Cofactor Immobilization.
Expected completion date
May 2015
Estimated size(pages)
100
0.00 USD
Total
Terms and Conditions
Introduction
The publisher for this copyrighted material is Springer Science + Business Media. By clicking
"accept" in connection with completing this licensing transaction, you agree that the following
terms and conditions apply to this transaction (along with the Billing and Payment terms and
71

conditions established by Copyright Clearance Center, Inc. ("CCC"), at the time that you opened
your Rightslink account and that are available at any time at http://myaccount.copyright.com).
Limited License
With reference to your request to reprint in your thesis material on which Springer Science and
Business Media control the copyright, permission is granted, free of charge, for the use
indicated in your enquiry.
Licenses are for one-time use only with a maximum distribution equal to the number that you
identified in the licensing process.
This License includes use in an electronic form, provided its password protected or on the
university’s intranet or repository, including UMI (according to the definition at the Sherpa
website: http://www.sherpa.ac.uk/romeo/). For any other electronic use, please contact Springer
at (permissions.dordrecht@springer.com or permissions.heidelberg@springer.com).
The material can only be used for the purpose of defending your thesis limited to university-use
only. If the thesis is going to be published, permission needs to be re-obtained (selecting
"book/textbook" as the type of use).
Although Springer holds copyright to the material and is entitled to negotiate on rights, this
license is only valid, subject to a courtesy information to the author (address is given with the
article/chapter) and provided it concerns original material which does not carry references to
other sources (if material in question appears with credit to another source, authorization from
that source is required as well).
Permission free of charge on this occasion does not prejudice any rights we might have to
charge for reproduction of our copyrighted material in the future.
Altering/Modifying Material: Not Permitted
You may not alter or modify the material in any manner. Abbreviations, additions, deletions
and/or any other alterations shall be made only with prior written authorization of the author(s)
and/or Springer Science + Business Media. (Please contact Springer at
(permissions.dordrecht@springer.com or permissions.heidelberg@springer.com)
Reservation of Rights
Springer Science + Business Media reserves all rights not specifically granted in the
combination of (i) the license details provided by you and accepted in the course of this
licensing transaction, (ii) these terms and conditions and (iii) CCC's Billing and Payment terms
and conditions.
Copyright Notice:Disclaimer
You must include the following copyright and permission notice in connection with any
reproduction of the licensed material: "Springer and the original publisher /journal title, volume,
year of publication, page, chapter/article title, name(s) of author(s), figure number(s), original
copyright notice) is given to the publication in which the material was originally published, by
72

adding; with kind permission from Springer Science and Business Media"
Warranties: None
Example 1: Springer Science + Business Media makes no representations or warranties with
respect to the licensed material.
Example 2: Springer Science + Business Media makes no representations or warranties with
respect to the licensed material and adopts on its own behalf the limitations and disclaimers
established by CCC on its behalf in its Billing and Payment terms and conditions for this
licensing transaction.
Indemnity
You hereby indemnify and agree to hold harmless Springer Science + Business Media and CCC,
and their respective officers, directors, employees and agents, from and against any and all
claims arising out of your use of the licensed material other than as specifically authorized
pursuant to this license.
No Transfer of License
This license is personal to you and may not be sublicensed, assigned, or transferred by you to
any other person without Springer Science + Business Media's written permission.
No Amendment Except in Writing
This license may not be amended except in a writing signed by both parties (or, in the case of
Springer Science + Business Media, by CCC on Springer Science + Business Media's behalf).
Objection to Contrary Terms
Springer Science + Business Media hereby objects to any terms contained in any purchase
order, acknowledgment, check endorsement or other writing prepared by you, which terms are
inconsistent with these terms and conditions or CCC's Billing and Payment terms and
conditions. These terms and conditions, together with CCC's Billing and Payment terms and
conditions (which are incorporated herein), comprise the entire agreement between you and
Springer Science + Business Media (and CCC) concerning this licensing transaction. In the
event of any conflict between your obligations established by these terms and conditions and
those established by CCC's Billing and Payment terms and conditions, these terms and
conditions shall control.
Jurisdiction
All disputes that may arise in connection with this present License, or the breach thereof, shall
be settled exclusively by arbitration, to be held in The Netherlands, in accordance with Dutch
law, and to be conducted under the Rules of the 'Netherlands Arbitrage Instituut' (Netherlands
Institute of Arbitration).OR:
All disputes that may arise in connection with this present License, or the breach thereof,
shall be settled exclusively by arbitration, to be held in the Federal Republic of Germany,

73

in accordance with German law.
Other terms and conditions:
v1.3
Questions? customercare@copyright.com or +1-855-239-3415 (toll free in the US) or +1978-646-2777.
Gratis licenses (referencing $0 in the Total field) are free. Please retain this printable
license for your reference. No payment is required.

.SPRINGER LICENSE
TERMS AND CONDITIONS
May 04, 2015

This is a License Agreement between Justin L Beauchamp ("You") and Springer ("Springer")
provided by Copyright Clearance Center ("CCC"). The license consists of your order details, the
terms and conditions provided by Springer, and the payment terms and conditions.
All payments must be made in full to CCC. For payment instructions, please see
information listed at the bottom of this form.
License Number
3600500361357
License date
Apr 01, 2015
Order Content Publisher
Springer
Applied Microbiology and
Order Content Publication
Biotechnology
Characterization of Thermotoga
maritima glycerol
Order Content Title
dehydrogenase for the
enzymatic production of
dihydroxyacetone
74

Order Content Author
Order Content Date
Volume number
Issue number
Type of Use
Requestor type
Publisher
Portion
Format
Will you be translating?
Print run

Justin Beauchamp
Jan 1, 2014
98
16
Book/Textbook
Publisher
ProQuest
Full text
Print and Electronic
No
6
Yes and you are the sole author
of the new work
None
Thermotoga Maritima Glycerol
Dehydrogenase as a Catalyst for
Dihydroxyacetone Production:
Enzyme Characterization,
engineering and Cofactor
Immobilization.
Justin L Beauchamp
May 2015
100
0.00 USD

Author of this Springer article
Order reference number

Title of new book

Author of new book
Expected publication date of new book
Estimated size of new book (pages)
Total
Terms and Conditions

Introduction
The publisher for this copyrighted material is Springer Science + Business Media. By clicking
"accept" in connection with completing this licensing transaction, you agree that the following
terms and conditions apply to this transaction (along with the Billing and Payment terms and
conditions established by Copyright Clearance Center, Inc. ("CCC"), at the time that you
opened your Rightslink account and that are available at any time at
http://myaccount.copyright.com).
Limited License
Springer Science + Business Media hereby grants to you a non-exclusive license to use this
material, for the use as indicated in your inquiry. Licenses are for one-time use only with a
maximum distribution equal to the number that you identified in the licensing process.
This License includes use in an electronic form, provided it's password protected, on intranet, or
CD-Rom/E-book. For any other electronic use, please contact Springer at
permissions.dordrecht@springer.com or permissions.heidelberg@springer.com
Although Springer holds copyright to the material and is entitled to negotiate on rights, this

75

license is only valid, provided permission is also obtained from the author (address is given with
the article/chapter) and provided it concerns original material which does not carry references to
other sources (if material in question appears with credit to another source, authorization from
that source is required as well).
Geographic Rights: Scope
Licenses may be exercised anywhere in the world.
Altering/Modifying Material: Not Permitted
However figures and illustrations may be altered minimally to serve your work. Any other
abbreviations, additions, deletions and/or any other alterations shall be made only with prior
written authorization of the author(s) and/or Springer Science + Business Media. (Please
contact Springer at permissions.dordrecht@springer.com or
permissions.heidelberg@springer.com)
Reservation of Rights
Springer Science + Business Media reserves all rights not specifically granted in the
combination of (i) the license details provided by you and accepted in the course of this
licensing transaction, (ii) these terms and conditions and (iii) CCC's Billing and Payment terms
and conditions.
License Contingent on Payment
While you may exercise the rights licensed immediately upon issuance of the license at the end
of the licensing process for the transaction, provided that you have disclosed complete and
accurate details of your proposed use, no license is finally effective unless and until full
payment is received from you (either by Springer Science + Business Media or by CCC) as
provided in CCC's Billing and Payment terms and conditions. If full payment is not received by
Due Date, then any license preliminarily granted shall be deemed automatically revoked and
shall be void as if never granted. Further, in the event that you breach any of these terms and
conditions or any of CCC's Billing and Payment terms and conditions, the license is
automatically revoked and shall be void as if never granted. Use of materials as described in a
revoked license, as well as any use of the materials beyond the scope of an unrevoked license,
may constitute copyright infringement and Springer Science + Business Media reserves the
right to take any and all action to protect its copyright in the materials.
Copyright Notice:
Please include the following copyright citation referencing the publication in which the material
was originally published. Where wording is within brackets, please include verbatim.
"With kind permission from Springer Science+Business Media: <book/journal title,
chapter/article title, volume, year of publication, page, name(s) of author(s), figure number(s),
and any original (first) copyright notice displayed with material>."
Warranties
Springer Science + Business Media makes no representations or warranties with respect to the
licensed material.
Indemnity
You hereby indemnify and agree to hold harmless Springer Science + Business Media and
CCC, and their respective officers, directors, employees and agents, from and against any and

76

all claims arising out of your use of the licensed material other than as specifically authorized
pursuant to this license.
No Transfer of License
This license is personal to you and may not be sublicensed, assigned, or transferred by you to
any other person without Springer Science + Business Media's written permission.
No Amendment Except in Writing
This license may not be amended except in a writing signed by both parties (or, in the case of
Springer Science + Business Media, by CCC on Springer Science + Business Media's behalf).
Objection to Contrary Terms
Springer Science + Business Media hereby objects to any terms contained in any purchase
order, acknowledgment, check endorsement or other writing prepared by you, which terms are
inconsistent with these terms and conditions or CCC's Billing and Payment terms and
conditions. These terms and conditions, together with CCC's Billing and Payment terms and
conditions (which are incorporated herein), comprise the entire agreement between you and
Springer Science + Business Media (and CCC) concerning this licensing transaction. In the
event of any conflict between your obligations established by these terms and conditions and
those established by CCC's Billing and Payment terms and conditions, these terms and
conditions shall control.
Jurisdiction
All disputes that may arise in connection with this present License, or the breach thereof, shall
be settled exclusively by the country's law in which the work was originally published.
Other terms and conditions:
v1.3

Questions? customercare@copyright.com or +1-855-239-3415 (toll free in the US) or +1978-646-2777.
Gratis licenses (referencing $0 in the Total field) are free. Please retain this printable
license for your reference. No payment is required.

77

REFERENCES

78

REFERENCES

1.

Ausubel, F. M., R. Brent, R. E. Kingston, D. D. Moore, J. G. Seidman, J. A. Smith,
and K. Struhl (ed.). 1993. Current Protocols in Molecular Biology, vol. Greene
Publishing & Wiley-Interscience, New York.

2.

Bauer, R., N. Katsikis, S. Varga, and D. Hekmat. 2005. Study of the inhibitory effect
of the product dihydroxyacetone on Gluconobacter oxydans in a semi-continuous twostage repeated-fed-batch process. Bioprocess Biosys. Engin. 28:37–43.

3.

Bliznak, J. B., and R. E. Harcarufka. 1981. Sweetener and flavoring compositions and
method of producing same. United States Patent:4,277,511.

4.

Brinen, L. S., J. M. Canaves, X. Dai, A. M. Deacon, M. A. Elsliger, S. Eshaghi, R.
Floyd, A. Godzik, C. Grittini, S. K. Grzechnik, C. Guda, L. Jaroszewski, C. Karlak,
H. E. Klock, E. Koesema, J. S. Kovarik, A. Kreusch, P. Kuhn, S. A. Lesley, D.
McMullan, T. M. McPhillips, M. A. Miller, M. D. Miller, A. Morse, K. Moy, J.
Ouyang, A. Robb, K. Rodrigues, T. L. Selby, G. Spraggon, R. C. Stevens, H. v. d.
Bedem, J. Velasquez, J. Vincent, X. Wang, B. West, G. Wolf, S. S. Taylor, K. O.
Hodgson, J. Wooley, and I. A. Wilson. 2003. Crystal structure of a zinc-containing
glycerol dehydrogenase (TM0423) from Thermotoga maritima at 1.5 Å resolution.
Proteins: Struct., Funct., and Bioinf. 50:371–374.

5.

Brooks, B. R., C. L. Brooks, A. D. Mackerell, L. Nilsson, R. J. Petrella, B. Roux, Y.
Won, G. Archontis, C. Bartels, S. Boresch, A. Caflisch, L. Caves, Q. Cui, A. R.
Dinner, M. Feig, S. Fischer, J. Gao, M. Hodoscek, W. Im, K. Kuczera, T. Lazaridis,
J. Ma, V. Ovchinnikov, E. Paci, R. W. Pastor, C. B. Post, J. Z. Pu, M. Schaefer, B.
Tidor, R. M. Venable, H. L. Woodcock, X. Wu, W. Yang, D. M. York, and M.
Karplus. 2009. CHARMM: The biomolecular simulation program. J. Comput. Chem.
30:1545–1614.

6.

Bruins, M. E., A. E. M. Janssen, and R. M. Boom. 2001. Thermozymes and their
applications. Appl. Biochem. Biotechnol. 90:155–186.

7.

Bush, E. C., A. E. Clark, C. M. Deboever, L. E. Haynes, S. Hussain, S. Ma, M. M.
McDermott, A. M. Novak, and J. S. Wentworth. 2012. Modeling the role of negative
cooperativity in metabolic regulation and homeostasis. PLOS ONE 7:e48920.

8.

Cornish-Bowden, A. 1995. Fundamentals of enzyme kinetics, vol. Portland Press Ltd.,
London.

9.

Cossement, E., J. P. Geerts, J. Gobert, P. Michel, and E. Wulfert. 1990. Substituted
1-(1H-imidazol-4-yl)alkyl-benzamides as anti-ischemics and as α-2-adrenergic receptor
agonists. United States Patent:4,923,865

10.

Ferdinand, W. 1966. The interpretation of non-hyperbolic rate curves for two-substrate
enzymes. Biochem. J. 98:278–283.

79

11.

Forde, J., L. Oakey, L. Jennings, and P. Mulchahy. 2005. Fundamental differences in
bioaffinity of amino acid dehydrogenases for N6- and S6-linked immobilized cofactors
using kinetic-based enzyme capture strategies. Anal. Biochem. 338:102–112.

12.

Gehrer, E., and H. Wolfgang. 1994. Manufacture of 1,2-propylene glycol. United States
Patent:5,306,847.

13.

Goujon, M., H. McWilliam, W. Z. Li, F. Valentin, S. Squizzato, J. Paern, and R.
Lopez. 2010. A new bioinformatics analysis tools framework at EMBL-EBI. Nucl. Acids
Res. 38:W695–W699.

14.

Hekmat, D., R. Bauer, and J. Fricke. 2003. Optimization of the microbial synthesis of
dihydroxyacetone from glycerol with Gluconobacter oxydans. Bioprocess Biosys. Engin.
26:109–116.

15.

Henis, Y. I., and A. Levitzki. 1980. Mechanism of negative cooperativity in
glyceraldehyde-3-phosphate dehydrogenase deduced from ligand competition
experiments. Proc. Natl. Acad. Sci. USA 77:5055–5059.

16.

Ivanetich, K. M., R. D. Goold, and C. N. Sikakana. 1990. Explanation of the nonhyperbolic kinetics of the glutathione s-transferases by the simplest steady-state random
sequential bi bi mechanism. Biochem. Pharmicol. 39:1999–2004.

17.

Johnson, J. A., and R. M. Fusaro. 1994. Alteration of skin surface protein with
dihydroxyacetone: A useful application of the Maillard browning reaction, p. 114–119. In
T. P. Labuza, G. A. Reineccius, V. M. Monnier, J. O'Brien, and J. W. Baynes (ed.),
Maillard Reactions in Chemistry, Food, and Health. The Royal Society of Chemistry,
Cambridge.

18.

Kar, P., H. Wen, H. Li, S. D. Minteer, and S. Calabrese Barton. 2011. Simulation of
multistep enzyme-catalyzed methanol oxidation in biofuel cells. J. Electrochem. Soc.
158:B580–B586.

19.

Koshland, D. E., and K. Hamadani. 2002. Proteomics and models for enzyme
cooperativity. J. Biol. Chem. 277:46841–46844.

20.

Li, H., H. Wen, and S. Calabrese Barton. 2012. NADH oxidation by
electropolymerized azines on carbon nanotube modified electrodes. Electroan. 24:398–
406.

21.

Li, M., J. Wu, X. Liu, J. Lin, D. Wei, and H. Chen. 2010. Enhanced production of
dihydroxyacetone from glycerol by overexpression of glycerol dehydrogenase in an
alcohol dehydrogenase-deficient mutant of Gluconobacter oxydans. Bioresour. Technol.
101:8294-8299.

22.

Lindberg, M., P. Larsson, and K. Mosbach. 1973. A new immobilized NAD+
analogue, its application in affinity chromatography and as a functioning coenzyme. Eur.
J. Biochem. 40:187–193.

23.

Marçal, D., A. T. Rego, M. A. Carrondo, and F. J. Enguita. 2009. 1,3-Propanediol
dehydrogenase from Klebsiella pneumoniae: Decameric quaternary structure and possible
subunit cooperativity. J. Bacteriol. 191:1143–1151.

80

24.

Marshall, J. H., J. W. May, and J. Sloan. 1985. Purification and properties of
glycerol:NAD+ 2-oxidoreductase (glycerol dehydrogenase) from Schizosaccharomyces
pombe. J. Gen. Microbiol. 131:1581–1588.

25.

McGregor, W. G., J. E. Phillips, and C. H. Suelter. 1974. Purification and kinetic
characterization of a monovalent cation-activated glycerol dehydrogenase from
Aerobacter aerogenes. J. Biol. Chem. 249:3132–3139.

26.

Mosbach, K. 1974. AMP and NAD as general ligands, p. 229–242. Meth. Enzymol., vol.
34. Academic Press.

27.

Nishise, H., A. Nagao, Y. Tani, and H. Yamada. 1984. Further characterization of
glycerol dehydrogenase from Cellulomonas sp. NT3060. Agric. Biol. Chem. 48:1603–
1609.

28.

O'Flaherty, M., M. McMahon, and P. Mulcahy. 1999. A kinetic locking-on strategy
for bioaffinity purification: further studies with alcohol dehydrogenase. Protein Expr.
Purif. 15:127–145.

29.

Oakey, L., and P. Mulcahy. 2004. Immobilized cofactor derivatives for kinetic-based
enzyme capture strategies: direct coupling of NAD(P)+. Anal. Biochem. 335:316–325.

30.

Parhi, S. S. L. 1986. Substituted 2-mercapto-imidazoles and their preparation. United
States Patent:4,584,383.

31.

Piattoni, C. V., C. M. Figueroa, M. D. A. Diez, I. L. Parcerisa, S. Antuna, R. A.
Comelli, S. A. Guerrero, A. J. Beccaria, and A. A. Iglesias. 2013. Production and
characterization of Escherichia coli glycerol dehydrogenase as a tool for glycerol
recycling. Process Biochem. 48:406–412.

32.

Sellek, G. A., and J. B. Chaudhuri. 1999. Biocatalysis in organic media using enzymes
from extremophiles. Enzyme Microbial Technol. 25:471–482.

33.

Sievers, F., A. Wilm, D. Dineen, T. J. Gibson, K. Karplus, W. Z. Li, R. Lopez, H.
McWilliam, M. Remmert, J. Soding, J. D. Thompson, and D. G. Higgins. 2011. Fast,
scalable generation of high-quality protein multiple sequence alignments using Clustal
Omega. Mol. Syst. Biol. 7.

34.

Song, S. H., N. Ahluwalia, L. Yvonne, L. T. J. Delbaere, and C. Vieille. 2008.
Thermotoga maritima TM0298 is a highly thermostable mannitol dehydrogenase. Appl.
Microbiol. Biotechnol. 81:485-495.

35.

Spencer, P., K. J. Brown, M. D. Scawen, T. Atkinson, and M. G. Gore. 1989.
Isolation and characterization of the glycerol dehydrogenase from Bacillus
stearothermophilus. Biochim. Biophys. Acta 994:270–279.

36.

Srinivasan, V., K. S. Ma, M. W. W. Adams, M. G. Newton, J. P. Rose, and B. C.
Wang. 2002. Towards the crystal structure of glycerol dehydrogenase from Thermotoga
maritima. Acta Crystallog. D. D58:867–869.

37.

Studer, M. 2003. Production of optically active -hydroxyacetals. United States
Patent:6,660,890.

81

38.

Tynan, J., J. Forde, M. McMahon, and P. Mulcahy. 2000. Synthesis of a highly
substituted N6-linked immobilized NAD+ derivative using a rapid solid-phase modular
approach: Suitability for use with the kinetic locking-on tactic for bioaffinity purification
of NAD+-dependent dehydrogenases. Protein Express. Purif. 20:421–434.

39.

Vieille, C., and J. G. Zeikus. 2001. Hyperthermophilic enzymes: sources, uses, and
molecular mechanisms for thermostability. Microbiol. Mol. Biol. Rev. 65:1–43.

40.

Weckbecker, A., H. Groger, and W. Hummel. 2010. Regeneration of nicotinamide
coenzymes: principles and applications for the synthesis of chiral compounds. Adv.
Biochem. Engin./Biotechnol. 120:195–242.

41.

Wells, G., H. Prest, and C. W. Russ IV. 2011. Signal, noise, and detection limits in
mass spectrometry. Agilent Technologies. Technical note:5990-7651EN.

42.

Wichmann, R., and D. Vasic-Racki. 2005. Cofactor regeneration at the lab scale. Adv.
Biochem. Engin./Biotechnol. 92:225–260.

43.

Yamada-Onodera, K., N. Kawahara, Y. Tani, and H. Yamamoto. 2004. Synthesis of
optically active diols by Escherichia coli transformant cells that express the glycerol
dehydrogenase gene of Hansenula polymorpha DL-1. Engin. Life Sciences 4:413–417.

44.

Yang, D., J. H. Zhang, M. K. Wong, Y. C. Yip, and M. W. Tang. 1998. Process for
catalytic epoxidation of olefinic compounds, novel cyclic ketone catalysts useful in said
process. United States Patent:5,763,623.

82

CHAPTER 3. Activity of select dehydrogenases with sepharose-immobilized N6-CM-NAD

83

This work is previously published and is reprinted under terms of the open access creative
commons 3.0 license: http://creativecommons.org/licenses/by-nc/3.0/. Beauchamp, J., and C.
Vieille. 2015. Activity of Select Dehydrogenases with Sepharose-Immobilized
N6-carboxymethyl-NAD. Bioengineered, published online: 03 Feb.

84

3.1 ABSTRACT
N6-carboxymethyl-NAD (N6-CM-NAD) can be used to immobilize NAD onto a substrate
containing terminal primary amines. We previously immobilized N6-CM-NAD onto sepharose
beads and showed that Thermotoga maritima glycerol dehydrogenase could use and recycle the
immobilized cofactor. We now show that Saccharomyces cerevisiae alcohol dehydrogenase,
rabbit muscle L-lactate dehydrogenase (type XI), bovine liver L-glutamic dehydrogenase (type
III), Leuconostoc mesenteroides glucose-6-phosphate dehydrogenase, and Thermotoga maritima
mannitol dehydrogenase are active with soluble N6-CM-NAD. The products of all enzymes
except 6-phospho-D-glucono-1,5-lactone were formed when sepharose-immobilized N6-CMNAD was recycled by T. maritima glycerol dehydrogenase, indicating that N6-immobilized NAD
is suitable for use by a variety of different dehydrogenases. Observations of the enzyme active
sites suggest that steric hindrance plays a greater role in limiting or allowing activity with the
modified cofactor than do polarity and charge of the residues surrounding the N6-amine group on
NAD.

3.2 INTRODUCTION
NAD(P)-dependent dehydrogenases participate in a wide variety of metabolic pathways.
They catalyze over 12% of all metabolic reactions in Escherichia coli (8). The requirement that
the cofactor be provided in stoichiometric amounts is the greatest limitation to using
dehydrogenases as catalysts. Widespread industrial use will require regeneration of the cofactor.
Immobilization of the cofactor allows it to be retained in a reaction vessel or recovered more
easily, allowing the same cofactor to be used for many cycles. As a proof of concept, we
previously synthesized the NAD analog N6-carboxymethyl-NAD (N6-CM-NAD), in which the
85

N6 primary amine in the adenine moiety is now –NH-CH2-COO-, and we immobilized it on
sepharose beads using a diamino linker. The immobilized cofactor was tested with Thermotoga
maritima glycerol dehydrogenase (TmGlyDH) in two simultaneous reactions: glycerol oxidation
to dihydroxyacetone (DHA), and hydroxyacetone reduction to (R)-1,2-propanediol. TmGlyDH
could use and recycle the immobilized cofactor for more than 1,800 turnovers to allow
measurable product accumulation (2).
Ottolina and colleagues (1990) used a similar N6 linkage to modify NAD (6). They
studied the effects of strong cationic and anionic substituents as well as that of a large polymer
(PEG, 20 kDa) on the properties of the cofactor . The eleven enzymes tested generally showed
increased Km and decreased Vmax with the modified NADs. Enzyme activity with PEG-NAD
varied from as high as 66% of the activity with native NAD to undetectably low. A strong
anionic substituent led to a stronger reduction in activity (8-43% of native NAD) than a strong
cationic substituent (4-93% of native NAD). The large differences in activity of the various
enzymes with the NAD analogs were likely caused by the different shapes and chemistries of
these enzymes’ NAD binding sites.
NAD(P)-dependent dehydrogenases share a common NAD-binding domain and Rossman
fold (7). However, the nature and positioning of residues around the NAD binding site can vary
significantly from one enzyme to the next, causing variations in enzymes’ affinity for the
cofactor, determining whether NADP or NAD is the preferred cofactor, and changing which
areas of the cofactor are solvent-accessible. Here we test five NAD-dependent dehydrogenases
for activity with Sepharose-bead immobilized N6-CM-NAD to address two questions. Is our
immobilization method suitable for use with a broad set of enzymes? And can structural
information on the NAD-binding site and cofactor solvent accessibility be used to predict

86

activity with immobilized N6-CM-NAD? We selected four commercially available NADdependent dehydrogenases whose crystal structure in complex with NAD is known, as well as T.
maritima mannitol dehydrogenase (TmMtDH), which we had studied previously (11).

Table 3.1: Enzymes used in this study.
Protein Data Bank
Enzyme

Reaction catalyzed

code for enzyme
NAD complex

TmGlyDH

Glycerol + NAD+

TmMtDH

Mannitol + NAD+

Saccharomyces cerevisiae alcohol

Ethanol + NAD+

Dihydroxyacetone + NADH

Fructose + NADH

Acetaldehyde + NADH

1KQ3a

None

2HCY

dehydrogenase (yADH)

Rabbit muscle type XI L-lactate

L-Lactate + NAD+

Pyruvate + NADH

3H3F

dehydrogenase (rLDH)

Bovine liver type III L-glutamate

L-Glutamate + H2O + NAD+

dehydrogenase (bGDH)

NH3 + NADH

Leuconostoc mesenteroides glucose- D-Glucose 6-phosphate + NAD+
6-phosphate dehydrogenase

2-Oxoglutarate +

6-Phospho-D-

3MW9

1H94

glucono-1,5-lactone + NADH

(LmG6PDH)
a

Structure without NAD. NAD was imported into the structure as described (2).

3.3 RESULTS AND DISCUSSION
The Sigma-Aldrich website was searched for commercially available dehydrogenases and NADdependent oxidoreductases. The list of enzymes generated was then cross-referenced against the
87

Protein Data Bank (PDB). Of the ten enzymes that had a crystal structure in complex with NAD,
four (Table 3.1) were selected and purchased from Sigma-Aldrich for our study. TmMtDH was
added to the study because it was already available as a purified, recombinant enzyme in our
laboratory (11). To allow comparison of its active site, we built a 3D model of TmMtDH in
complex with NAD (Figure 3.1).
The adenine moiety binding portion of the NAD binding pocket of the five enzymes and
of TmGlyDH was visualized (Figure 3.1). The residues and pocket structure around the adenine
varied greatly among the enzymes. The configuration of NAD is similar in TmMtDH, yADH,
rLDH, bGDH, and LmG6PDH, with the adenine’s N6-amine pointing toward the solvent. In
TmGlyDH, NAD’s adenine moiety is flipped, with the N6-amine pointing towards the enzyme’s
surface. Since NAD’s N6-amine is less solvent exposed in TmGlyDH than in the other enzymes,
we expected TmGlyDH to have the lowest activity with N6-CM-NAD. The specific activity of
the mesophilic enzymes with NAD and N6-CM-NAD and the Vmax values with the cofactor and
analog of TmGlyDH and TmMtDH were compared (Table 3.2). TmMtDH’s Vmax with N6-CMNAD was almost 70% of that with NAD and the Km increased only 1.8 fold. In contrast,
TmGlyDH’s Vmax with N6-CM-NAD was only 2% of that with NAD and the Km increased 15
fold. With the exception of LmG6PDH, the N6-amine solvent-accessible area (Table 3.2)
correlated well (R2 = 0.95 for linear regression, not shown) with the percent activity with N6CM-NAD (Table 2) (R2 = 0.61when LmG6PDH is included). In contrast, the type of residues
(i.e., charged, polar, or non-polar) surrounding the N6-amine (Table 3.2) showed no relationship
to activity with N6-CM-NAD (Table 3.2).

88

Figure 3.1: Zoom view of enzyme-bound NAD (8-iodo-NAD for yADH). A. TmMtDH, B.
TmGlyDH, C. yADH, D. LmG6PDH, E. rLDH, F. bGDH. PDB numbers are listed in Table 1.
Several 3D models of TmMtDH were generated using Modeller software8 and the online
homology modeling server I-TASSER.9-11. The different modeling approaches used single and
multiple templates that each showed over 25% identity and below 10% gaps in alignments with
TmMtDH. Models were analyzed using the scoring methods DOPE, DFIRE, and OPUS.12-14
The best model was produced by I-TASSER using the structures of the silverleaf whitefly
sorbitol dehydrogenase (PDB # 1E3J), human sorbitol dehydrogenase (PDB # 1PL8), Sulfolobus
solfataricus glucose dehydrogenase (PDB # 2CDC), Thermus thermophilus threonine 3dehydrogenase (PDB # 2DQ4), and mouse class II alcohol dehydrogenase (PDB # 1E3I) as
templates. The structure was minimized using the CHARMM force field15 and NAD was
imported into TmMtDH’s active site using the coordinates of NAD in human sorbitol
dehydrogenase. Enzyme surfaces were visualized using The PyMOL Molecular Graphics
System, Version 1.3 Schrödinger, LLC.

A

B

C

D

E

F

89

Table 3.2: Specific activity of the selected enzymes with 100 μM NAD and N6-CM-NAD and
enzyme ranking by polarity/charge and openness of the area surrounding NAD’s N6-amine.
Specific activities were tested at 25°C by following the increase in A340nm in quartz cuvettes
containing 100 mM substrate (L-lactate, glucose-6-phosphate, L-glutamate, or ethanol), 100 μM
NAD or N6-CM-NAD in 50 mM sodium phosphate (pH 7.5). Reactions were started by adding
1.25 μg (rLDH), 1.08 μg (yADH) 0.201 μg (LmG6PDH), or 4.22 μg (bGDH) enzyme. Using the
crystal structures or 3D models of the enzymes visualized in PyMOL, amino acid side-chains
and backbone groups within 5 Å of the N6-amine were assessed for polarity and charge. The N6amine solvent-accessible area was calculated using PyMOL.
Enzyme

NAD

N6-CM-NAD % activity Polar groups near the N6-amine

N6-amine solventaccessible area (Å2)

(μmol min-1

(μmol min-1

mg-1)

mg-1)

2.2 ± 0.1a

1.5 ± 0.5a

68

1 Polar, 1 negative charge

49.9

1.42 ± 0.05

0.81 ± 0.24

57

2 Polar, 1 non-polar

43.9

260 ± 40

46 ± 4

18

1 Negative charge, 2 non-polar

41.7

yADH

38 ± 2

15.5 ± 0.9

41

1 Polar, 2 non-polar

28.8

rLDH

10 ± 1

2.5 ± 0.4

25

3 Non-polar

24.8

12.2 ± 0.8b

0.28 ± 0.2b

2

2 Polar

16.2

TmMtDH
bGDH
LmG6PDH

TmGlyDH

a -Vmax values for TmMtDH were measured at 50°C as described for TmGlyDH except that
mannitol was substituted for glycerol (2).
b-Vmax values at 50°C previously measured (2).

To test enzyme activity with Sepharose-bound N6-CM-NAD, we initially tested bGDH as
the cofactor-recycling enzyme in recycling reactions between rLDH and bGDH and between
LmG6PDH and bGDH. In these recycling reactions, rLDH and LmG6DH were set to reduce the
cofactor, and bGDH to oxidize it. Neither of the two reactions yielded any product. We could not
tell whether none of the enzymes used the immobilized N6-CM-NAD, or if the recycling enzyme
(bGDH) was causing the problem. Note that bGDH showed very low specific activity with NAD
in the conditions tested (Table 3.2). For these reasons, we repeated the recycling reactions with
TmGlyDH as the cofactor-recycling enzyme. Even though TmGlyDH is poorly active at 25°C,
we knew it to be active with Sepharose-bound N6-CM-NAD (2). Recycling reactions were set
90

between TmGlyDH and each of yADH, rLDH, bGDH, and TmMtDH, where yADH, rLDH,
bGDH, and TmMtDH oxidized the cofactor and TmGlyDH reduced it while oxidizing glycerol
to DHA. Reaction progress was monitored by measuring DHA accumulation with highperformance liquid chromatography. Because LmG6PDH’s only commercially available
substrate is glucose-6-phosphate, the recycling reaction between LmG6PDH and TmGlyDH was
set with TmGlyDH producing NAD+ during DHA reduction to glycerol, and reaction progress
was monitored by measuring glycerol accumulation by high-performance liquid
chromatography. No product accumulation was observed in the LmG6PDH-TmGlyDH reaction
(not shown). Even though LmG6PDH has the highest specific activity with soluble N6-CM-NAD
of the six enzymes tested (Table 3.2), LmG6PDH could be inactive with immobilized N6-CMNAD. Because TmGlyDH is inhibited by its products (not shown), glycerol added as part of the
LmG6PDH suspension (initial glycerol concentration in the reaction estimated at 138 mM) could
also have inhibited glycerol production by TmGlyDH, preventing cofactor recycling. Product
accumulation was observed for the recycling reactions involving yADH, rLDH, bGDH, and
TmMtDH (Figure 3.2). Although the solvent-accessible area of the N6-amine correlated well
with activity for soluble N6-CM-NAD (Table 3.2), no such relationship was apparent after
immobilization on Sepharose, with only small differences in DHA accumulation observed
between the different reactions. Regardless, our results indicate that many different enzymes can
use Sepharose-immobilized N6-CM-NAD as their cofactor.
Note that we used TmGlyDH, the enzyme with the least activity with soluble N6-CM-NAD
(Table 3.2), as the cofactor-recycling enzyme because it was previously shown to be active with
the immobilized cofactor. TmMtDH, bGDH, and yADH showed the highest relative activities
with soluble N6-CM-NAD (Table 3.2). bGDH used in combination with TmMtDH as the

91

cofactor-recycling enzyme produced 2.6 times more product than in combination with
TmGlyDH (data not shown), suggesting that poor activity of TmGlyDH with immobilized N6CM-NAD could have limited product accumulation. In contrast, the yADH/TmMtDH
combination produced only 20% as much product as the yADH/TmGlyDH combination. The
reason behind this poor yield is unknown, since the two enzymes were not inhibited by each
other’s substrate (data not shown).

3.4 CONCLUSION
Our work demonstrates that N6-CM-NAD immobilized onto a large particle, such as
Sepharose, can be used by a variety of enzymes. NAD’s N6-amine solvent-accessible area in an
enzymeNAD complex is a good indicator of whether that enzyme will be active with N6-CMNAD. The possibility of one enzyme’s substrate inhibiting the enzyme used in combination
highlights the need for an alternative cofactor regeneration method. Electrochemical regeneration
could be a good alternative but requires efficient methods for cofactor and enzyme
immobilization on electrodes. N6-CM-NAD could be used to create a bioelectronics linkage
similar to that previously developed by Hassler and Worden (4). They developed a renewable
bioelectronics interface with the enzyme and cofactor immobilized on an electrode surface. The
immobilization method used relied on a boronate linkage to NADP’s cis-diol. The main
limitation of the boronate linkage was that this linkage is relatively weak with a dissociation
constant (Kd) of 5.9x10-3, and boronate shows high affinity for other cis-diols found on the
substrates of many dehydrogenases (1). The enzyme substrates would displace the cofactor from
the electrode surface. The N6-immobilized NAD tested here would allow a new linkage to be
produced that is stable in the presence of the enzyme substrates and products.

92

3.5 ACKNOWLEDGEMENTS
We thank Phillip G. Gross for contributing to the kinetic studies of TmGlyDH. We thank Dr.
Srinivasa Gopal for helping with CHARMM minimization and DOPE, DFIRE and OPUS
analysis of the TmMtDH model, and Dr. Kaillathe Padmanabhan for his help with PyMOL and
solvent-accessible area determination. This work was supported by grant # CBET 0756703 from
the National Science Foundation, grant # 2008-35504-04611 from the United States Department
of Agriculture Cooperative State Research, Education, and Extension Service National
Research Initiative, grant # 2010-04061 from the United States Department of Agriculture
National Institute for Food and Agriculture’s Sustainable Bioenergy Research Program, and by
Michigan State University start up funds.

93

REFERENCES

94

REFERENCES

1.

Bailon, P., G. K. Ehrlich, W.-J. Fung, and W. Berthold (ed.). 2000. Methods in
molecular biology: affinity chromatography methods and protocols, vol. 147. Humana
Press Inc, Totowa NJ.

2.

Beauchamp, J., P. G. Gross, and C. Vieille. 2014. Characterization of Thermotoga
maritima glycerol dehydrogenase for the enzymatic production of dihydroxyacetone.
Appl Microbiol Biotechnol 98:7039–7050.

3.

Brooks, B. R., C. L. Brooks, A. D. Mackerell, L. Nilsson, R. J. Petrella, B. Roux, Y.
Won, G. Archontis, C. Bartels, S. Boresch, A. Caflisch, L. Caves, Q. Cui, A. R.
Dinner, M. Feig, S. Fischer, J. Gao, M. Hodoscek, W. Im, K. Kuczera, T. Lazaridis,
J. Ma, V. Ovchinnikov, E. Paci, R. W. Pastor, C. B. Post, J. Z. Pu, M. Schaefer, B.
Tidor, R. M. Venable, H. L. Woodcock, X. Wu, W. Yang, D. M. York, and M.
Karplus. 2009. CHARMM: The biomolecular simulation program. J. Comput. Chem.
30:1545–1614.

4.

Hassler, B. L., N. Kohli, J. G. Zeikus, I. Lee, and R. M. Worden. 2007. Renewable
dehydrogenase-based interfaces for bioelectronic applications. Langmuir 23:7127–7133.

5.

Lu, M., A. D. Dousis, and J. Ma. 2008. OPUS-PSP: an orientation-dependent statistical
all-atom potential derived from side-chain packing. J. Mol. Biol. 376:288–301.

6.

Ottolina, G., G. Carrea, S. Riva, and A. Bückmann. 1990. Coenzymatic properties of
low molecular-weight and macromolecular N6-derivatives of NAD+ and NADP+ with
dehydrogenases of interest for organic synthesis. Enzyme Microb. Technol. 12:596–602.

7.

Rao, S. T., and M. G. Rossmann. 1973. Comparison of super-secondary structures in
proteins. J. Mol. Biol. 76:241–256.

8.

Reed, J. L., T. D. Vo, C. H. Schilling, and B. O. Palsson. 2003. An expanded genomescale model of Escherichia coli K-12 (iJR904 GSM/GPR). Genome Biol. 4:R54.

9.

Sali, A., and T. L. Blundell. 1993. Comparative protein modeling by satisfaction of
spatial restraints. J. Mol. Biol. 234:779–815.

10.

Shen, M. Y., and A. Sali. 2006. Statistical potential for assessment and prediction of
protein structures. Protein Sci 15:2507–2524.

11.

Song, S. H., N. Ahluwalia, Y. Leduc, L. T. Delbaere, and C. Vieille. 2008.
Thermotoga maritima TM0298 is a highly thermostable mannitol dehydrogenase. Appl.
Microbiol. Biotechnol. 81:485–495.

12.

Wu, S., J. Skolnick, and Y. Zhang. 2007. Ab initio modeling of small proteins by
iterative TASSER simulations. BMC Biol. 5:17.

13.

Zhang, Y. 2008. I-TASSER server for protein 3D structure prediction. BMC Bioinf.
9:40.

14.

Zhang, Y. 2007. Template-based modeling and free modeling by I-TASSER in CASP7.
Proteins 69 Suppl 8:108–117.
95

15.

Zhou, H., and Y. Zhou. 2002. Distance-scaled, finite ideal-gas reference state improves
structure-derived potentials of mean force for structure selection and stability prediction.
Protein Sci. 11:2714–2726.

96

CHAPTER 4. Inactivation of Thermotoga maritima glycerol dehydrogenase by its product,
dihydroxyacetone

97

4.1 ABSTRACT
Thermotoga maritima glycerol dehydrogenase (TmGlyDH), which oxidizes glycerol into
dihydroxyacetone (DHA) in an NAD-dependent reaction, is susceptible to inactivation by DHA.
DHA reacts readily with primary amines on the enzyme surface through Maillard reactions.
TmGlyDH’s Arg and Lys residues modified by DHA were identified in tryptic and chymotryptic
digests of TmGlyDH by mass spectrometry. To make TmGlyDH a better catalyst, the residues
susceptible to modification by DHA were systematically mutagenized, and the mutant enzymes
were screened for increased stability in the presence of DHA. Wild-type TmGlyDH lost all
activity after 4 h in the presence of DHA at 50°C, but mutants K104Q and K361Q showed
increased stability, maintaining 90% of activity after 5 h in DHA at 50°C.

4.2 INTRODUCTION
We previously characterized recombinant Thermotoga maritima glycerol dehydrogenase
(TmGlyDH) (2). TmGlyDH is a multimeric enzyme whose crystal structure was solved as part of
a structural proteomics project (3, 15), but its activity and stability properties had not been
studied. Only the coordinates of one TmGlyDH monomer are deposited in the PDB by the
authors of the crystal structure (3), so it is unclear whether the native enzyme is a tetramer as
reported after the first crystallization (15), or an octamer as predicted by Pisa software and
deposited in the PDB (7). The homologous glycerol dehydrogenase from Bacillus
stearothermophilus shares 48% sequence identity with TmGlyDH and was reported to be a
homooctamer (13). TmGlyDH has the potential to be used as a catalyst for dihydroxyacetone
(DHA) production from glycerol or to produce R-1,2-propanediol at high enantiomeric excess
from hydroxyacetone. While characterizing the enzyme, we observed that TmGlyDH was
98

inactivated by its product dihydroxyacetone (2). Dihydroxyacetone is the only FDA-approved
sunless tanning reagent. It is able to turn skin brown through Maillard reactions with the primary
amines in the Lys and Arg residues in proteins on the surface of the skin (Figure 4.1) (6).
Figure 4.1: Maillard reaction scheme.

Rearrangement
products

Protein glycation has been primarily studied because it leads to the formation of
advanced glycation endproducts that are involved in the aging process as well as diabetes and
diabetes related diseases (11, 12). DHA has also been observed to glycate human serum albumin
(14). The extent of glycation by DHA increased with temperature (5 sites at 23°C, 9 at 37°C, and
14 at 41°C; at 30 mM DHA and pH 7.4), pH (5 sites at pH 5, 9 at pH 7.4, and 20 at pH 8.0; at 30
mM DHA and 37°C), and substrate concentration (14).
Glycation has also been observed to inactivate enzymes. The Kcat/Km ratio of the rat liver
aldehyde reductase decreased by 32% after glycation (16). During the enzymatic production of
biologically active oligosaccharides, glycation was shown to inactivate Penicillium citrinum 1,2α-mannosidase and almond α-mannosidase, and higher molecular weight enzyme species formed
by glycation were observed using Matrix-assisted laser desorption/ionization-time of flight mass
spectrometry (MALDI-TOF) (10). After a two-week reaction at 55°C, 40-60% of enzyme
activity was lost. In reactions containing Maillard reaction inhibitors, the yield increased 27-53%
compared to the control (10). However, adding inhibitors to synthetic reactions increases
reaction costs and may also increase downstream purification costs.
99

Since TmGlyDH is inactivated by its product DHA, and DHA is known to readily react
with amino groups through Maillard reactions (6), we expect that the main contributor to enzyme
inactivation is glycation by DHA. We determined which Lys and Arg residues are susceptible to
modification by DHA. These residues were targeted for mutagenesis to increase TmGlyDH’s
resistance to inactivation by DHA.

4.3 MATERIALS AND METHODS
4.3.1 Materials
TmGlyDH was purified, enzyme activity assays were performed, and mutants were
constructed as described (2). Sequencing-grade trypsin and chymotrypsin were purchased from
Promega (Madison, WI). DHA, iodoacetamide (IAA), and trifluoroacetic acid (TFA) were
purchased from Sigma-Aldrich (St. Louis, MO). Monosodium and disodium phosphate were
from J.T Baker (Avantor Performance Materials, Inc., Center Valley, PA). Ultrapure urea was
from Invitrogen (Carlsbad, CA). 190 proof pure ethanol was from Koptec (King of Prussia, PA).
Ammonium bicarbonate was from Jade scientific (Canton, MI).

4.3.2 Enzyme activity assays
Enzyme activity was measured at 50°C and 80°C by monitoring the absorbance of
NADH at 340 nm (ε340 = 6,220 M-1 cm-1) with a Varian Cary 300 UV/Vis spectrophotometer
(Palo Alto, CA) equipped with a Peltier heating system. The final assay mixture typically
contained 6.7 µg enzyme, 200 mM DHA, 200 µM NADH, and 50 mM sodium phosphate buffer
(pH 7.45 at 50°C and pH 7.4 at 80°C). The reaction was started by adding 20 µL of 10 mM

100

NADH to 980 µL of the other components or adding 40 µL of enzyme (0.168 mg/mL) to 960 µL
of other components in a heated quartz cuvette. All activity measurements at 50°C were
performed in triplicate, and all activity measurements at 80°C were performed in quadruplicate.

4.3.3 Two-hour enzyme inactivation assays
For 2-h heat inactivations at 50°C, 44 µL of a 0.168 mg/mL solution (7.39 µg) of enzyme
(wild-type or mutant) was added to 814 µL of 65 mM sodium phosphate buffer (pH 7.45), and
incubated in an Eppendorf tube at 50°C for 2 h. The remaining activity was measured by adding
200 µL of pre-warmed 1 M DHA to the 780-µl enzyme solution, and starting the reaction by
adding 20 µL of 10 mM NADH.
For 2-h inactivations by DHA at 50°C, 44 µL of a 0.168 mg/mL solution of enzyme was
added to 1,034 µL of 50 mM sodium phosphate buffer (pH 7.45) containing 200 mM DHA, and
incubated in an Eppendorf tube at 50°C for 2 h. Activity was measured after transferring 980 µL
of the enzyme-DHA solution to a heated cuvette, and the reaction was started by adding 20 µL of
10 mM NADH. A two tailed student’s t-test was used to determine if results were significant.

4.3.4 Time course enzyme inactivation assays
For time course heat inactivation assays at 50°C, TmGlyDH (58.6 μg) was incubated at
50°C for 7 h in 6.8 mL of 65 mM sodium phosphate buffer (pH 7.45 at 50°C) in a polypropylene
tube. At each time point, 780 µL of the enzyme solution was added to 200 µL of 1 M DHA
preheated at 50°C in a quartz cuvette, and the reaction was started by adding 20 µL of 10 mM

101

NADH. Non-linear curve fitting of all inactivation time courses was performed using Origin 8.0
(Originlab, Northampton, MA).
For time course inactivations by DHA at 50°C, TmGlyDH (58.6 μg) was incubated at
50°C for 7 h in 8.3 mL of 51 mM sodium phosphate buffer (pH 7.45 at 50°C) containing 200
mM DHA in a polypropylene tube. At each time point, 980 µL of the enzyme solution was added
to a quartz cuvette preheated at 50°C, and the reaction was started by adding 20 µL of 10 mM
NADH.
For time course heat inactivation assays at 80°C, TmGlyDH (6.7 μg) was incubated at
80°C for 0, 4, 8, 12, 16, and 20 min in 780 μL of a preheated solution of 64 mM sodium
phosphate buffer (pH 7.4 at 80°C) in a quartz cuvette. After incubation, 20 µL of 10 mM NADH
was added, and the reaction was started by adding 200 µL of preheated (80°C) 1 M DHA.
Time course inactivations by DHA at 80°C were set up as described for the time course
heat inactivation assays at 80°C, except that the preheated DHA solution was added at the same
time as the enzyme, and after incubation the reaction was started by adding NADH.

4.3.5 Trypsin and chymotrysin digests of TmGlyDH
Prior to digesting the enzyme, 1 nmol TmGlyDH was added to 300 μL of water. TmGlyDH was
precipitated by diluting 5-fold with -20°C acetone and cooled at -20°C for 60 min. After
centrifugation at 13,000 × g for 10 min at room temperature, the supernatant was decanted and
the pellet was air dried for 30 min at room temperature. The pellet was dissolved in 100 μL of 8
M urea containing 400 mM ammonium bicarbonate, and incubated at 50°C for 30 min. The
solution was cooled to room temperature and 12 μL of 100 mM iodoacetamide (IAA) was added.
The solution was incubated at room temperature for 20 min in the dark, and diluted 8-fold with
102

50 mM ammonium bicarbonate. The digestion was started by adding 1 μg trypsin per nmol of
TmGlyDH. The digestion was incubated overnight at 37°C, and stopped by adding
trifluoroacetic acid (TFA) to a final concentration of 0.5% w/w. The digestion by chymotrypsin
was as described above for trypsin. For the chymotrypsin/trypsin double digest, trypsin was
added after the chymotrypsin digestion and incubated overnight at 37°C; the reaction was
stopped by the addition of TFA. To ensure that hydrophobic peptides were not sticking to the
sides of the Eppendorf tube, another set of digestions was performed followed by rinsing the
tubes used for the digestion with ethanol. After the acetone precipitation step, the reaction was
prepared using half the amount of all reagents (500 μL total). After the digestion, the tube was
rinsed with 500 mL ethanol.

4.3.6 Identification of peptides by electrospray ionization time-of-flight mass spectrometry
(ESI)-TOF
Immediately following digestion, peptides were separated by HPLC (Waters Corporation,
Milford MA) using a C18 column (Beta basic-18, particle size 5 µm, dimensions 100 mm × 2.1
mm, Thermo scientific, Waltham, MA) at 0.25 mL/min at 30°C using a gradient from 98% A
(0.1% formic acid) to 98% B (acetonitrile or methanol) over 10 min (acetonitrile) or 23 min
(methanol) followed by a constant flow at 98% B for 2 min. Peptide masses were analyzed using
ESI-TOF (Q-tof ultima, Waters Corporation, Milford MA). The molecular masses of all
expected peptides, as well as those of the peptides modified by TCA or IAA, were calculated and
searched for in the mass spectrum.

103

4.3.7 Identification of the TmGlyDH residues modified by DHA
To identify which residues were modified by DHA, TmGlyDH was incubated with DHA
prior to being digested with trypsin and chymotrypsin. Five nmol of TmGlyDH (207 μg) were
added to 300 μL of 200 μM DHA and incubated at 50°C for 5, 15, and 60 min. After acetone
precipitation, the enzyme was digested as described above. Peptides were identified by ESI-TOF
as described above. The previously observed peptides were searched for with and without the
addition of the molecular mass of DHA. The Arg and Lys residues susceptible to modification by
DHA were located in the TmGlyDH protein structure (PDB no. 1KQ3) and visualized using the
PyMOL Molecular Graphics System, Version 1.3 (Schrödinger, LLC). The macromolecular
assembly generated by Pisa software (7) was downloaded from the PDB and used to visualize
TmGlyDH as an octamer in PylMol.

4.4 RESULTS
4.4.1 TmGlyDH inactivation by DHA at 50°C
During the previously reported kinetic characterization of TmGlyDH, we observed that
TmDlyDH becomes inactivated by DHA (2). We set up a time course inactivation experiment to
determine how much faster the enzyme inactivates in the presence of DHA (Figure 4.2).
TmGlyDH quickly lost activity in the presence of DHA at 50°C, losing almost all activity after 4
h. In the absence of DHA, the enzyme retained over 90% activity after 7 h at 50°C. To engineer
TmGlyDH for increased resistance to inactivation by DHA, the next step was to determine which
residues get modified by DHA.

104

Figure 4.2: Time course of TmGlyDH inactivation at 50°C with and without DHA. ( ): heat
inactivation, exponential fit ( ); ( ): Inactivation by DHA, exponential fit (---)

4.4.2 Identification of the TmGlyDH residues modified by DHA
To identify the amino acids susceptible to modification by DHA, we needed to observe
all the TmGlyDH Arg and Lys residues represented in the TmGlyDH fragments after digestion
with trypsin. The peptides were analyzed using HPLC electrospray ionization time of flight
(HPLC-ESI-TOF) mass spectrometry. To facilitate observing large hydrophobic peptides, a
second set of digestions was prepared including an ethanol rinsing step to recover the most
hydrophobic fragments. Because not all Arg and Lys residues were represented in the peptides
identified in tryptic digests, the enzyme was also digested with chymotrypsin and a combination
of trypsin plus chymotrypsin.
Peptides corresponding to 23 of the 40 total Lys + Arg residues were identified in the
trypsin digest. Adding the ethanol wash after the trypsin digestion yielded peptides for ten of the
40 total Lys + Arg residues, one of them not seen in the trypsin-only digest. Peptides
corresponding to 26 of the 40 total Lys + Arg residues were identified in the chymotrypsin
digest, eleven of them not found in the trypsin-only and trypsin-ethanol treatments. The
chymotrypsin-ethanol treatment yielded peptides for 29 of the 40 total Lys + Arg residues, two
105

of them not found previously. The chymotrypsin/trypsin double digest yielded fragments
corresponding to only four of the 40 total Lys + Arg residues, but three of them had not been
found in the single digests. Combining results from all of the digests, peptides corresponding to
every Lys and Arg residue were observed (Table S4.1).
The digestions and analysis were repeated after preincubating the enzyme with DHA for
5, 15, and 60 min at 50°C. Fifteen different Lys and Arg residues were found to be modified by
DHA (Table 4.1). The residues modified by DHA were clustered together in five different
regions on the surface of the protein (Figure 4.3). Five of the 15 total Arg and ten of the 25 total
Lys residues become modified by DHA. Of the seven Arg and four Lys residues at subunit
interfaces, only two (R143 and R217) become modified by DHA. Of the remaining, solventexposed 21 Lys and 7 Arg residues, 10 Lys and 3 Arg become modified by DHA.

Table 4.1: TmGlyDH residues found to be modified by DHA. Redundant information from
different digestion conditions is not shown. Negative sign indicates no modification, positive
sign indicates modification, asterisk indicates the peak disappeared, implying modification.
Residue Experimental conditions Incubation Time (min)
5
15
60
K60
Trypsin
*
*
K104
Trypsin
+
+
K106
Chymotrypsin-Ethanol
+
+
K107
Chymotrypsin-Ethanol
+
+
R143
Chymotrypsin-Ethanol
+
+
+
+
K157
Chymotrypsin-Ethanol
R161
Chymotrypsin-Ethanol
+
+
K181
Chymotrypsin-Ethanol
*
*
K183
Chymotrypsin-Ethanol
*
*
K216
Chymotrypsin
+
+
+
R217
Chymotrypsin
+
+
+
K222
Chymotrypsin
+
+
+
R358
Chymotrypsin-Ethanol
+
+
+
R360
Chymotrypsin-Ethanol
+
+
+
K361
Chymotrypsin-Ethanol
+
+
+

106

Figure 4.3: Crystal structure of TmGlyDH showing the residues modified by DHA. Left panel – ribbon diagram of one monomer
showing the residues modified by DHA in stick representation. Right panel – ribbon diagram of one monomer and surfaces of
surrounding subunits. All Lys and Arg residues of one monomer are shown in stick representation. Cyan – ribbon structure of
TmGlyDH, Yellow – Lys and Arg residues modified by DHA, Red – all other Lys and Arg residues, Orange – active site Zn, Gray –
surfaces of surrounding subunits.

107

4.4.3 Screening TmGlyDH mutants for increased stability to DHA
Based on the hypothesis that modification by DHA of several of the 15 Arg and Lys
identified leads to enzyme inactivation, mutations of these residues should make the enzyme less
susceptible to inactivation by DHA. To test this premise, 14 of these 15 Lys and Arg residues
were mutated individually to Gln. We chose Gln for the substitution because Gln is the amino
acid most similar in size and hydrophilicity to Lys and it is able to form hydrogen bonds. K157
had previously been mutated to Gly in previous work (2). All the mutants maintained more than
60% of the wild type activity, except R143 and K183Q which only had 39% and 46% of the
wild-type activity respectively.

Specific activity (µmol min-1 mg-1)

Figure 4.4: Screening of TmGlyDH mutants for increased resistance to inactivation by
DHA at 50°C. ( ): Initial activity at 50°C; ( ): residual activity after a 2-h incubation at 50°C;
( ): residual activity after a 2-h incubation with 200 mM DHA at 50°C; *: Statistically different
from the wild-type enzyme’s thermostability at 50°C (p ≤ 0.05); **: Statistically different from
the wild-type enzyme’s stability in the presence of DHA at 50°C (p ≤ 0.05).
18
16
14
12

*
**

10

**

8
6

**

*

**

**
**

4
2
0

Mutant enzymes were first screened for stability to DHA by testing activity remaining
after a 2-h incubation with DHA at 50°C (Figure 4.4). Stability in the absence of DHA was also
108

measured to highlight any change in thermostability. Six of the mutant enzymes (K60Q, K104Q,
K107Q, K157G, R161Q, and R361Q) showed significantly (p ≤ 0.05) increased stability
compared to wild-type TmGlyDH in the presence of DHA.

Because temperature can have a large impact on enzyme inactivation, mutant enzymes
that were significantly more stable than wild-type TmGlyDH at 50°C with DHA were also
screened for increased resistance to inactivation by DHA at 80°C. In contrast to the 50°C results,
mutants K104Q and K107Q were not significantly more stable than wild-type TmGlyDH with
DHA at 80°C. Mutant K157G was the most stable, instead of K361Q. (Figure 4.5).

Figure 4.5: Residual activity of TmGlyDH and its mutants after a 16-min incubation with
DHA at 80°C. ( ): Initial activity at 80°C; ( ): residual activity after a 16-min incubation at
80°C; ( ): residual activity after a 16-min incubation with 200 mM DHA at 80°C; *: Statistically
different from the wild-type enzyme’s thermostability at 80°C (p ≤ 0.05); **: Statistically
different from the wild-type enzyme’s stability in the presence of DHA at 80°C (p ≤ 0.05).
Specific activity (µmol min-1 mg-1)

60
50
40
30

*

*
20

*
10

*

*

0

109

**

*
**

**

*

4.4.4 Time course analysis of mutants’ stability to DHA
The three mutant enzymes showing the most increased resistance to inactivation by DHA
at 50°C (K104G, K157G, and K361Q) and two of the most stable mutants at 80°C (K157G and
R161Q) were further tested in a 7-h time course inactivation assay at 50°C. Data points were fit
to an exponential decay model (eq. 1), but the fit was generally poor, so decay constants could
not be calculated (Figure 4.6).
Equation 4.1

𝑣
= 𝑒 −𝑘𝑡
𝑣0
In general, the mutant enzymes were less thermostable than wild-type TmGlyDH, but
K104Q and K361Q TmGlyDHs were stabilized by the presence of DHA. In the presence of
DHA, wild-type TmGlyDH lost half of its activity between 1 and 2 h. Mutant K104Q was less
thermostable than wild-type TmGlyDH, but DHA temporarily prevented its thermoinactivation,
and its activity was only reduced by half after 7 h. The K157G and R161Q mutants were less
thermostable than wild-type TmGlyDH, and both lost half of their activity after less than an hour
in the presence of DHA. Mutant K361Q was less thermostable than wild-type TmGlyDH, but
was the most resistant to inactivation by DHA; it remained almost 100% active after 5 h in the
presence of DHA, and lost less than half its activity over the 7-h inactivation time course.
The four mutant TmGlyDHs that showed increased stability to DHA at both 80°C and
50°C were selected for a time course inactivation study at 80°C. Enzymes were incubated with
and without DHA at 80°C, and activity was tested after 0, 4, 8, 12, 16, and 20 min of incubation.
The inactivation time course data points were fit to an exponential decay model (eq. 1) where
relative activity was plotted against incubation time and the exponential decay constant K was
110

calculated. The fits to the inactivation time courses at 80°C (Figure 4.7) (adjusted R2 ranged
from 0.66 to 0.98) were much better than the fits to the 50°C time course data, so decay
constants are reported in Figure 4.8. The R161Q mutant had the smallest decay constant in the
presence of DHA at 80°C, but the K157G mutant retained higher activity after 16 min of
incubation with DHA at 80°C.
Figure 4.6: Time courses of TmGlyDH and mutants’ inactivations at 50°C with and
without DHA. ( ): heat inactivation, exponential fit ( ); ( ): Inactivation by DHA, exponential
fit (---).
WT

K104Q

K157G

R161Q

111

Figure 4.6 (cont’d)

K361Q

Figure 4.7: Inactivation time courses of TmGlyDH and its mutants at 80°C with and
without DHA. ( ): heat inactivation, exponential fit ( ); ( ): Inactivation by DHA, exponential
fit (---). (A) Wild-type TmGlyDH; (B) TmGlyDH K60Q, (C) TmGlyDH K157G, (D) TmGlyDH
R161Q, (E) TmGlyDH K361Q, (F) TmGlyDH K157G/K361Q, (G) TmGlyDH K157G/R161Q.
WT

K60Q

112

Figure 4.7 (cont’d)

K157G

R161Q

K361Q

K157G/K361Q

K157G/R161Q

113

Figure 4.8: Specific activities and decay constants of wild-type TmGlyDH and its mutants
at 80°C with and without DHA. Decay constants with standard error of TmGlyDH mutants
incubated with DHA (dark gray) or in buffer only (light gray) at 80°C; specific activity at 80°C
(Δ). *: Statistically different from the wild-type enzyme’s thermostability at 80°C (p ≤ 0.05), **:
Statistically different from the wild-type enzyme’s stability with DHA at 80°C (p ≤ 0.05).
70

0.18

60

Decay constant

0.16

50

0.14
0.12
0.1
0.08
0.06
0.04

40

*
**

*

*

*
**

**
**

* **

30
20

**
10

0.02
0

Specific activity (μmol min-1 mg-1)

0.2

0

The mechanisms of inactivation by DHA at 50°C and 80°C appear to be different as
evidenced by the different patterns of the time course inactivations. Surprisingly, the K104Q and
K107Q mutants that were more resistant to inactivation by DHA at 50°C, were not particularly
so at 80°C. We prepared a K157G/K361Q double mutant and tested it in both screens at 50°C
and 80°C and in a time course inactivation at 80°C. The double mutant was less stable than the
individual mutants at 80°C (Figures 4.7 and 4.8), and it was not significantly different from wildtype TmGlyDH at 50°C (Figure 4.5).

114

4.5 DISCUSSION
4.5.1 The importance of TmGlyDH stability
TmGlyDH has been previously characterized to determine if it is suitable for use in a
catalytic reactor to produce DHA from glycerol (2). To use glycerol dehydrogenase as a catalyst,
NAD+ must be supplied to the reaction in stoichiometric amounts or it must be regenerated.
Glutamate dehydrogenase and NADH oxidase have been used to regenerate oxidized cofactor,
but their use is limited to the production of high value products because of the cost associated
with using two enzymes for one reaction (17). Recent advances in electrochemical oxidation of
NADH could make electrochemical regeneration of oxidized cofactors a lower cost alternative to
enzymatic regeneration (9). Apart from supplying the cofactor to the reaction, the other main
cost of enzymatic catalysis is the cost of producing and purifying the enzyme. Optimally the
enzyme will be usable for many batch reactions, or for use in a flow-through reactor for many
turnovers. In this study, we show that wild-type TmGlyDH is unstable in the presence of DHA,
and therefore, wild-type TmGlyDH would need to be replaced often if used in a catalytic reactor.
Reducing TmGlyDH inactivation by DHA will help address this issue.
4.5.2 TmGlyDH inactivation by DHA is a complex process
Our inactivation results have not been highly reproducible. For example, in the 2-h DHA
inactivation screen, the K157G mutant retained more than 50% of its activity, but only showed
about 25% of its activity after 2 h in the time course inactivation by DHA at 50°C. The variation
between experiments could be explained by the complexity of the inactivation process. DHA
could react with any of the 15 Lys and Arg residues accessible to modification in any order and
at different rates. Not all modifications by DHA lead to inactivation, but for the ones that do, the
kinetics of inactivation of each residue and the order in which the different residues get modified
115

in each experiment will affect the inactivation course. We might observe reproducible
stabilization results only once enough residues are mutagenized to remove this stochasticity.
There are a number of different pathways to inactivation. When DHA modifies a given
residue, neighboring residues may be more likely to get modified. And DHA reacting at some
sites may have no effect on the enzyme at all. Also, modification at different specific sites could
have a synergistic inactivating effect. DHA modification in one subunit may also change the
inactivation process of adjacent subunits. Note that TmGlyDH has been shown to behave
cooperatively in its kinetics for glycerol and dihydroxyacetone, where substrate binding in one
subunit changes the substrate affinity of the other subunits (2). Also as the protein unfolds, it can
start aggregation processes that can destabilize neighboring subunits in the multimer (8). The
order of modification and effect on the enzyme could vary greatly between inactivation time
courses. Mutating all the Lys and Arg residues that, once modified by DHA, inactivate the
enzyme should make the kinetic data more consistent. In general, we trust the inactivation time
course data more than the 2-h screening at 50°C and 16-min at 80°C data points because of the
context of the surrounding time points.

4.5.3 DHA modification sites
The residues modified by DHA are localized in five general regions (Figure 4.3).
Because the DHA modification sites seemed to cluster together at specific locations in the
enzyme, it is possible that modification at one of the residues in each cluster leads to partial local
unfolding, allowing nearby Lys and Arg residues to be modified by DHA as well. Similar
proportions of Arg and Lys (33% and 40%, respectively) were modified by DHA, suggesting
that both types of residues are equally susceptible to modification. The mutations that increased
116

stability to DHA (K104Q, and K361Q) at 50°C targeted residues that are solvent-exposed on the
multimeric enzyme and are relatively far from the active site. At 50°C, the K361Q mutant is the
enzyme most resistant to DHA, followed by the K104Q mutant. Both K361 and K104 form salt
bridges (with E309 and E73, respectively). Mutations to Gln would disrupt these salt bridges,
which may destabilize the protein at high temperatures and allow the heat inactivation processes
to take place faster. But the effect of breaking the salt bridge may be small at 50°C because salt
bridges are more stabilizing at higher temperatures than at moderate temperatures (4).
Surprisingly, the addition of DHA actually stabilized these two mutant enzymes, allowing them
to maintain their activity for several hours in DHA at 50°C. This effect could be because the
substrate helps the enzyme retain its active conformation, modifications at other Lys or Arg
residues offset the effect of breaking the salt bridges, or the substrate allows the enzyme to form
an active oligomer different from the starting multimer (see below).
Because TmGlyDH is a multimer and the most stabilizing mutations are far from the
active site, we checked whether these mutations are at subunit interfaces. Neither K104 nor K361
are. Of the 15 amino acids modified by DHA, only R143 and R217 are at a subunit interface, but
neither mutant was more stable than the wild-type enzyme in the 2-h inactivation screen. In
general, the Lys and Arg residues located at subunit interfaces are protected from modification
by DHA: only 18% were modified versus 46% of solvent-exposed Lys and Arg residues. Of the
two least active mutants, the R143Q mutant could have lower activity than wild-type TmGlyDH
because of disruptions in the subunit interface. There was no clear reason for the K183Q mutant
to have substantially reduced activity. K183 forms hydrogen bonds with the main chain oxygens
of H335 and E337, but modeling of the K183Q mutation in TmGlyDH with PyMol suggests that
Gln could maintain those hydrogen bonds (not shown).

117

4.5.4 Stability and biphasic behavior in 50°C inactivation time-courses
The inactivation time courses at 50°C with and without DHA did not follow first-order
kinetics. As reported for urease and luciferase, non-first-order kinetics comes about when the
different reactions that lead to inactivation are interacting with each other (8). Biphasic
inactivation results from a combination of denaturation, aggregation, and coagulation. In most
models of inactivation, the initial drop in activity results from denaturation of the oligomer
reaching equilibrium, and the second phase is a result of coagulation and aggregation processes
(5, 8). (Aggregation is the formation of non-native oligomers of the folded monomer,
coagulation is the coalescence of unfolded monomers.) When peptide denaturation and multimer
denaturation are at equilibrium, inactivation becomes directly proportional to aggregation and
coagulation (8). For most of the mutants, inactivation in the absence of DHA at 50°C showed a
biphasic pattern. For the K361Q and K104Q mutants, stability to DHA at 50°C appeared to have
a period of stability, followed by inactivation. A so called grace period or induction period in
inactivation time courses has been adequately described by changes in the multimeric species (1,
8). Particularly, multimeric enzymes that dissociate into active oligomers show a grace period in
their inactivation kinetics (1, 8). The grace period is a result of the formation of active
intermediates that subsequently form inactive aggregates, and the kinetics of this process is such
that at higher temperatures, the inactivation will look more like first-order inactivation (1, 5, 8).
Because the grace period was only observed in the presence of DHA, the enzyme-substrate
complex may be able to form alternative active oligomers or DHA reacting at certain residues
forms active intermediates, while the enzyme alone is unable to do so. Although we previously
observed a brief grace period during inactivation of wild-type TmGlyDH by DHA at 50°C (2),

118

the enzyme had been cooled on ice after inactivation and then reheated to 50°C for 15 min prior
to measuring residual activity. Here we tested activity directly after inactivation.
At higher temperatures, the acceleration of the inactivation process can make the
multistep process appear to follow first-order kinetics (5, 8), which could explain why we
observed better fits to the first-order inactivation equation at 80°C. At 80°C, inactivation
processes that do not involve DHA will take place faster, reducing the relative effect of DHA on
the enzymes’ stability. The effect of modification by DHA is much more apparent at 50°C than
at 80°C, as illustrated by the larger difference between the inactivation time courses with and
without DHA at 50°C than at 80°C, indicating that thermal denaturation is the dominant process
at 80°C. At 80°C, the K157G mutant is the most stable followed by the R161Q mutant. The most
stabilizing mutations at 80°C are in (K157G) or near (R161Q) the NAD binding pocket.
Modification of the wild-type enzyme by DHA at these sites could reduce affinity for the
cofactor, thus reducing the apparent activity. If true, we would expect this effect to be small,
though, because we used concentrations of NAD nearly 10-fold the Km. If cofactor affinity is
playing a role in the inactivation, testing activity with higher concentrations of NAD would
eliminate the reduction in activity. It is possible that mutations K157G and R161Q both act
through the same mechanism to increase stability, because the K157G/R161Q double mutant
was not more stable than either of the individual mutants. The kinetics of modification at each
site may be different at 50°C and 80°C, leading to different inactivation pathways. Determining
the rate of modification at the two different temperatures could clarify the difference in
inactivation mechanism. For our purposes, the increase in stability of TmGlyDH mutants with
DHA at 50°C is more important because our proposed catalytic reactor will operate at 50°C and
DHA will accumulate during operation.

119

4.6 CONCLUSION
TmGlyDH becomes inactivated by its product DHA by DHA reacting with different Lys and Arg
residues. We identified 15 different amino acids susceptible to modification by DHA, located in
five regions of the protein. The two sites that destabilize the enzyme most in DHA at 50°C are
Lys361 and Lys104. Mutating these residues to Gln stabilized TmGlyDH in the presence of
DHA at 50°C. Although the stability increased substantially, the enzyme still lost substantial
activity after 7 h at 50°C. Combining multiple mutations may lead to increased stability of the
enzyme.

120

4.7 SUPPLEMENTARY MATERIALS

121

Table S4.1: Conditions used to detect peptides containing all Lys and Arg residues. •:
Peptide detected prior to incubation with DHA, and not modified by DHA. º: Peptide detected
prior to incubation with DHA and susceptible to modification by DHA. *: peptide detected prior
to incubation with DHA, but disappearing after incubation with DHA.

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40

Residue
R 10
R 27
R 31
K 42
K 55
R 57
K 60
R 74
K 93
K 99
K 104
K 106
K 107
K 136
R 137
R 143
K 157
R 161
K 181
K 183
R 191
R 202
K 216
R 217
K 222
K 230
K 266
K 272
K 286
R 288
K 289
K 322
K 326
K 330
K 344
K 351
R 355
R 358
R 360
K 361

Trypsin digest
Normal
With ethanol

•
•

•
•

Chymotrypsin digest
Normal
With ethanol

•
•

•
•

•
•
•
•
•
•
•

*

•
º

•
•
•

•
•

•
•

•

•
•
•
•
•
•
•
º
º
º

*

•
•

•

•
•
•
•
•

•
•
•
•
•
•
•
•

•
•

•
•

•

º
º
º

•
•
º
º
º
*
*

•
•
º
º
º

•
•
•
•
•
•

•
•
º
º
º

122

•

•

•
•
•

Chymotrypsin
then trypsin
Normal

•
•
•

REFERENCES

123

REFERENCES

1.

Atyaksheva, L. F., O. S. Pilipenko, E. S. Chukhrai, and O. M. Poltorak. 2008.
Similarity of and differences between the mechanisms of thermal inactivation of βgalactosidases of different origins. Russ. J. Phys. Chem. A 82:864–869.

2.

Beauchamp, J., P. G. Gross, and C. Vieille. 2014. Characterization of Thermotoga
maritima glycerol dehydrogenase for the enzymatic production of dihydroxyacetone.
Appl Microbiol Biotechnol 98:7039–7050.

3.

Brinen, L. S., J. M. Canaves, X. Dai, A. M. Deacon, M. A. Elsliger, S. Eshaghi, R.
Floyd, A. Godzik, C. Grittini, S. K. Grzechnik, C. Guda, L. Jaroszewski, C. Karlak,
H. E. Klock, E. Koesema, J. S. Kovarik, A. Kreusch, P. Kuhn, S. A. Lesley, D.
McMullan, T. M. McPhillips, M. A. Miller, M. D. Miller, A. Morse, K. Moy, J.
Ouyang, A. Robb, K. Rodrigues, T. L. Selby, G. Spraggon, R. C. Stevens, H. v. d.
Bedem, J. Velasquez, J. Vincent, X. Wang, B. West, G. Wolf, S. S. Taylor, K. O.
Hodgson, J. Wooley, and I. A. Wilson. 2003. Crystal structure of a zinc-containing
glycerol dehydrogenase (TM0423) from Thermotoga maritima at 1.5 Å resolution.
Proteins: Struct., Funct., Bioinf. 50:371–374.

4.

Elcock, A. H. 1998. The stability of salt bridges at high temperatures: implications for
hyperthermophilic proteins. J. Mol. Biol. 284:489–502.

5.

Illeová, V., M. Polakovič, V. r. Štefuca, P. Ačai, and M. Juma. 2003. Experimental
modelling of thermal inactivation of urease. J. Biotechnol. 105:235–243.

6.

Johnson, J. A., and R. M. Fusaro. 1994. Alteration of skin surface protein with
dihydroxyacetone: a useful application of the Maillard browning reaction, p. 114–119. In
T. P. Labuza, G. A. Reineccius, V. M. Monnier, J. O'Brien, and J. W. Baynes (ed.),
Maillard Reactions in Chemistry, Food, and Health. The Royal Society of Chemistry,
Cambridge.

7.

Krissinel, E., and K. Henrick. 2007. Inference of macromolecular assemblies from
crystalline state. J. Mol. Biol. 372:774–797.

8.

Lencki, R. W., J. Arul, and R. J. Neufeld. 1992. Effect of subunit dissociation,
denaturation, aggregation, coagulation, and decomposition on enzyme inactivation
kinetics: II. Biphasic and grace period behavior. Biotechnol. Bioeng. 40:1427–1434.

9.

Li, H., R. Li, R. M. Worden, and S. Calabrese Barton. 2014. Facilitation of high-rate
NADH electrocatalysis using electrochemically activated carbon materials. ACS Appl.
Mater. Inter. 6:6687–6696.

10.

Maitin, V., and R. A. Rastall. 2004. Enzyme glycation influences product yields during
oligosaccharide synthesis by reverse hydrolysis. J. Mol. Catal. B: Enzym. 30:195–202.

11.

Nowotny, K., T. Jung, A. Höhn, D. Weber, and T. Grune. 2015. Advanced glycation
end products and oxidative stress in type 2 diabetes mellitus. Biomolecules 5:194–222.

124

12.

Prasad, C., V. Imrhan, F. Marotta, S. Juma, and P. Vijayagopal. 2014. Lifestyle and
advanced gycation end products (AGEs) burden: its relevance to healthy aging. Aging
Dis. 5:212–217.

13.

Ruzheinikov, S. N., J. Burke, S. Sedelnikova, P. J. Baker, R. Taylor, P. A. Bullough,
N. M. Muir, M. G. Gore, and D. W. Rice. 2001. Glycerol dehydrogenase: structure,
specificity, and mechanism of a family III polyol dehydrogenase. Structure 9:789–802.

14.

Seneviratne, C., G. W. Dombi, W. Liu, and J. A. Dain. 2012. In vitro glycation of
human serum albumin by dihydroxyacetone and dihydroxyacetone phosphate. Biochem.
Biophys. Res. Commun. 417:817–823.

15.

Srinivasan, V., K. S. Ma, M. W. W. Adams, M. G. Newton, J. P. Rose, and B. C.
Wang. 2002. Towards the crystal structure of glycerol dehydrogenase from Thermotoga
maritima. Acta Crystallog. D. D58:867–869.

16.

Takahashi, M., Y.-b. Lu, T. Myint, J. Fujii, Y. Wada, and N. Taniguchi. 1995. In
vivo glycation of aldehyde reductase, a major 3-deoxyglucosone reducing enzyme:
identification of glycation sites. Biochemistry 34:1433–1438.

17.

van der Donk, W. A., and H. Zhao. 2003. Recent developments in pyridine nucleotide
regeneration. Curr. Opin. Biotechnol. 14:421–426.

125

CHAPTER 5. Conclusions and future work

126

5.1 REVIEW OF PROJECT
Dehydrogenases are potential catalysts for the production of many different compounds,
including chiral compounds and pharmaceutical precursors (17, 18). Catalytic use of
dehydrogenases has been limited due to the need to supply NAD to the reaction in stoichiometric
amounts or regenerate the cofactor. A limited number of high-value chemicals are produced by
dehydrogenases with a cofactor regeneration method. Typically the cofactor is regenerated by a
second dehydrogenase working in the opposite direction. The extra cost associated with cofactor
regeneration is the main limitation for industrial applications. Electrochemical cofactor
regeneration could be a lower cost alternative to enzyme coupled regeneration. The efficiency of
NADH electro-oxidation has been substantially increased recently, making electrochemical
regeneration a suitable alternative for reactions that require NAD+ (10, 11). Immobilizing the
enzyme and cofactor onto the electrode surface will allow reuse of the electrode for many
catalytic cycles. The goal of the collaboration project of the Vieille, Worden, and Calabrese
Barton laboratories is to develop a bioelectronic catalytic reactor that uses dehydrogenases to
catalyze chemical reactions. The long-term goal is to produce DHA from glycerol using
TmGlyDH in an electrochemical reactor. For my part of that project, I characterized and
engineered TmGlyDH. I also immobilized NAD while maintaining its coenzymic activity.

5.1.1 Review of TmGlyDH characterization
To determine appropriate reactor operating conditions and enzyme immobilization
conditions, TmGlyDH’s apparent kinetic parameters were determined for glycerol, DHA, and
NAD(H) at different pH values at 50°C. The Michaelis-Menten equation was not a good fit to
the kinetics data for glycerol and DHA because of apparent negative cooperativity. Thus kinetic
127

data for these substrates were fit to the Hill equation. The optimum pH for glycerol oxidation
was 7.9 with apparent Vmax, K0.5 for glycerol, and n of 24 ± 1.4 μmol min-1 mg-1, 39 ± 7 mM, and
0.86 ± 0.13, respectively, where n indicates the cooperativity level. TmGlyDH activity with N6CM-NAD+ was only 2% of the activity with NAD+. Single and double mutations and a triple
mutant in the enzyme’s active site were produced to increase activity with the soluble NAD
analogue; the activity of the best mutant with N6-CM-NAD+ increased over 10-fold. However,
the activity of the TmGlyDH mutants with sepharose-immobilized N6-CM-NAD was the same or
lower than that of the wild-type enzyme.
In a catalytic system, the longevity of the catalyst is critical because renewing the catalyst
is costly. We observed that TmGlyDH becomes inactivated by its product, DHA. TmGlyDH
inactivation by DHA is a result of the enzyme being modified by Maillard reactions between the
Lys and Arg residues and DHA. In Chapter 4, I used mass spectrometry to identify which amino
acids are being modified by DHA and mutated those amino acids to improve TmGlyDH stability
in the presence of DHA. The mutants’ stability to DHA was not consistent at 50°C and 80°C.
The mutant most stable to DHA at 50°C was K361Q while the mutant most stable to DHA at
80°C was K157G. Combining the K157G and K361Q mutations together did not improve
stability to DHA at 50°C or at 80°C significantly. Overall, the best mutant (K361Q) retained
initial activity and maintained activity twice as long as the wild-type enzyme in the presence of
DHA at 50°C.

5.1.2 Suitability of TmGlyDH for use in bioelectrocatalytic system
TmGlyDH promises to be a useful catalyst for the production of DHA using a
regenerated cofactor, but there are several reasons for which using TmGlyDH as a model system
128

for a bioelectronic reactor is not ideal. The kinetics of TmGlyDH do not fit standard MichaelisMenten kinetics, which will complicate mathematical modeling of the system. Modeling the
reversible kinetics of a dehydrogenase-based biofuel cell system has been demonstrated (7), but
modeling the reversible kinetics of cooperative enzymes would require creating a different,
potentially more complicated model using the reversible Hill equation (5). The reversible rate
equation for cooperative enzymes is empirical and does not provide mechanistic parameters (5).
Another limitation is that the enzyme becomes inactivated by its product, losing all activity after
4 h in the presence of DHA at 50°C. To prevent inactivation of the enzyme catalyst, DHA would
need to be continually removed from a catalytic reactor; this problem has been partially
addressed by mutating the DHA reactive Lys and Arg residues. The TmGlyDH mutants
produced have significantly increased stability to DHA at 50°C, but they lost thermostability.
The best mutants, K104Q and K361Q, retained similar initial activity to the wild-type and
maintained 90% of activity after 5 h with DHA. But K104Q and K361Q lost half their activity
after about 2 h and 5 h at 50°C, respectively, compared to the wild-type enzyme, which lost only
about 10% activity after 7 h at 50°C in the absence of DHA.

5.1.3 Suitability of N6-immobilized NAD for use in bioelectrocatalytic systems
To determine which chemical groups on NAD+ are accessible enough for binding with a
linker for immobilization, NAD+ was modeled into the crystal structure of TmGlyDH and into a
homology model of TmMtDH. Based on a review of the literature pertaining to available
chemistry to modify NAD and to the ability of enzymes to bind and use immobilized-NAD
analogues, I chose the N6-amine on the adenine moiety of NAD+ for modification. N6carboxymethyl-NAD+ (N6-CM-NAD+) was synthesized to allow binding to amines through
129

carbodiimide coupling. In Chapter 2, I demonstrated that TmGlyDH could use the immobilized
cofactor and that the length of the linker tethering the immobilized-N6-CM-NAD to the surface
had little effect on the ability of the enzyme to use the immobilized NAD.
In Chapter 3, I showed that several different dehydrogenases could use sepharoseimmobilized-N6-CM-NAD and that the enzymes’ active site structures could predict activity
with soluble N6-CM-NAD, but not with immobilized NAD. My results indicate that N6immobilized-NAD should have coenzymic activity with many dehydrogenases and that it is
suitable for use in an electrochemical system.

5.2 FUTURE WORK
5.2.1 Immobilizing NAD on an electrode
Having demonstrated that N6-immobilized NAD is usable by many different
dehydrogenases, the logical next step is to immobilize it on an electrode. The electrode needs to
be activated and modified to be a good NADH oxidation catalyst. Our collaborators have
demonstrated that activated carbon electrodes with methylene green adsorbed onto the surface
have excellent NADH oxidation activity (10). They incorporated carbon nanotubes that were
previously functionalized by carboxylation, and then electrooxidized them by cyclic
voltammetry. The carboxyl groups on the carbon nanotubes should provide active groups to
which a diamine linker can be attached. After electroactivation of the carbon nanotube-modified
electrode, methylene green can be adsorbed onto the surface and a diamino linker can be
attached through carbodiimide-promoted coupling. This surface should be tested for catalytic
NADH oxidation to confirm that the addition of a linker molecule does not decrease the NADH
oxidation rate because the long carbon chain in the diamino linkers may act as an insulator,
130

inhibiting current flow. Then the N6-CM-NAD can be immobilized to the terminal amine groups
through carbodiimide-promoted coupling.
An alternative approach to modifying the electrode is to build up layers of polyelectrolyte
(3). Because the carbon nanotube electrode has carboxyl groups, it is expected to have a negative
charge. Thus, asorbing a layer of a positively charged polymer like polyethyleneimine on the
electrode surface would provide primary amines to bind N6-CM-NAD. Multiple layers of
alternatively charged polymers could also be added, followed by NAD immobilization, although
too many layers may hold NAD too far from the catalytic surface.

5.2.2 Engineering of TmGlyDH
Directed engineering of TmGlyDH was only marginally successful for improving activity
with N6-CM-NAD and for stabilizing the enzyme to DHA. For stability to DHA, there are many
possible double mutants, or triple mutants, that could further improve stability. First, a 7-h time
course for the single mutant K107Q should be performed, since the K107Q mutant retained
almost as much activity after 2 h with DHA as the K104Q mutant did. Then combinations of
K104Q, K361Q, and K107Q (if K107Q’s increased stability to DHA is confirmed) should be
prepared. Combinations of these mutations could further improve TmGlyDH’s stability to DHA.
The approach taken to stabilize TmGlyDH to DHA is limited by only modifying residues that
react with DHA. An alternative strategy to stabilize the enzyme is to do directed evolution.
Directed evolution could introduce mutations that counteract the effects of the Lys and Arg
residues being modified by DHA instead of just mutating the specific residues that get modified.
Error-prone PCR could be used to mutate the enzyme and then mutants could be screened by
incubating them for several hours with DHA. Mutations that counteract the structure changes
131

that take place when DHA modifies the Lys and Arg residues would show increased residual
activity.

5.2.3 Adding TmGlyDH to the bioelectronic interface
After NAD is immobilized onto the electrode surface, the next step is to immobilize the
enzyme. TmGlyDH should have affinity for the electrode-immobilized NAD, and can then be
cross-linked in place using glutaraldehyde. A similar approach to the kinetic based locking on
strategy used for affinity chromatography with N6-immobilized NAD-sepharose (2, 15) (see
Chapter 1) could be employed to increase the enzyme’s affinity for electrode-immobilized NAD.
Adding a substrate analog like Tris (see Chapter 2) would allow the enzyme to form an inactive
enzyme-Tris-(electrode-immobilized NAD) complex, holding the enzyme on the electrode
surface. Then the addition of glutaraldehyde would fix the enzyme in place followed by washing
the Tris away. With the enzyme and cofactor immobilized, the electrode should be usable for
multiple catalytic cycles, thus reducing the operating costs. An added advantage to immobilizing
TmGlyDH is that immobilization of the enzyme onto the electrode surface couldfurther stabilize
TmGlyDH to DHA because the cross-linker may react with some of the DHA reactive Lys and
Arg residues, immobilization could eliminate the aggregation and coagulation processes, and
immobilization generally increases enzyme stability (13, 14).

5.3 ADDITIONAL USES OF THE BIOELECTRONIC INTERFACE
After the immobilization of NAD and TmGlyDH onto the electrode surface, the
electrochemical properties would be tested. If the bioelectronics interface shows suitable current
densities at moderate overpotentials and works for DHA production, it would be a proof of
132

concept for producing a dehydrogenase-based catalytic electrochemical reactor. The next step
would be to apply other enzymes to the interface. I have demonstrated that a number of different
dehydrogenases should be able to use N6-immobilized-NAD and therefore the electrode interface
should be adaptable to many different dehydrogenases. A working electrochemical connection
could be used for many additional applications such as fuel cells and fuel cell powered
implantable devices such as a blood glucose sensor and other biosensors (1, 3, 4, 6-9, 12, 16, 19).
The breadth of reactions catalyzed and the wide acceptance of N6-immobilized NAD by
dehydrogenases would allow the adaptation of the bioelectronic system to many novel
applications.

133

REFERENCES

134

REFERENCES

1.

Calabrese Barton, S., J. Gallaway, and P. Atanassov. 2004. Enzymatic biofuel cells
for implantable and microscale devices. Chem. Rev. 104:4867–4886.

2.

Forde, J., L. Oakey, L. Jennings, and P. Mulchahy. 2005. Fundamental differences in
bioaffinity of amino acid dehydrogenases for N6- and S6-linked immobilized cofactors
using kinetic-based enzyme capture strategies. Anal. Biochem. 338:102–112.

3.

Hassler, B. L., N. Kohli, J. G. Zeikus, I. Lee, and R. M. Worden. 2007. Renewable
dehydrogenase-based interfaces for bioelectronic applications. Langmuir 23:7127–7133.

4.

Hassler, B. L., and R. M. Worden. 2006. Versatile bioelectronic interfaces based on
heterotrifunctional linking molecules. Biosens. Bioelectron. 21:2146–2154.

5.

Hofmeyr, J.-H. S., and H. Cornish-Bowden. 1997. The reversible Hill equation: how to
incorporate cooperative enzymes into metabolic models. Comput. Appl. Biosci. 13:377–
385.

6.

Jasti, L. S., S. R. Dola, N. W. Fadnavis, U. Addepally, S. Daniels, and S.
Ponrathnam. 2014. Co-immobilized glucose oxidase and β-galactosidase on bovine
serum albumin coated allyl glycidyl ether (AGE)-ethylene glycol dimethacrylate
(EGDM) copolymer as a biosensor for lactose determination in milk. Enzyme Microb.
Technol. 64–65:67–73.

7.

Kar, P., H. Wen, H. Li, S. D. Minteer, and S. C. Barton. 2011. Simulation of multistep
enzyme-catalyzed methanol oxidation in biofuel cells. J. Electrochem. Soc. 158:B580–
B586.

8.

Kulys, J. J., U. Bilitewski, and R. D. Schmid. 1991. Reagentless biosensors for
substrates of dehydrogenases. Anal. Lett. 24:181–189.

9.

Li, G., N. Z. Ma, and Y. Wang. 2005. A new handheld biosensor for monitoring blood
ketones. Sens. Actuators, B 109:285–290.

10.

Li, H., R. Li, R. M. Worden, and S. Calabrese Barton. 2014. Facilitation of high-rate
NADH electrocatalysis using electrochemically activated carbon materials. ACS Appl.
Mater. Inter. 6:6687–6696.

11.

Li, H., H. Wen, and S. Calabrese Barton. 2012. NADH oxidation by
electropolymerized azines on carbon nanotube modified electrodes. Electroan. 24:398–
406.

12.

Lobo, M. J., A. J. Miranda, and P. Tuñón. 1997. Amperometric biosensors based on
NAD(P)-dependent dehydrogenase enzymes. Electroanalysis 9:191–202.

13.

López-Gallego, F., L. Betancor, C. Mateo, A. Hidalgo, N. Alonso-Morales, G.
Dellamora-Ortiz, J. M. Guisán, and R. Fernández-Lafuente. 2005. Enzyme
stabilization by glutaraldehyde crosslinking of adsorbed proteins on aminated supports. J.
Biotechnol. 119:70–75.

135

14.

Mateo, C., J. M. Palomo, G. Fernandez-Lorente, J. M. Guisan, and R. FernandezLafuente. 2007. Improvement of enzyme activity, stability and selectivity via
immobilization techniques. Enzyme Microb. Technol. 40:1451–1463.

15.

Oakey, L., and P. Mulcahy. 2004. Immobilized cofactor derivatives for kinetic-based
enzyme capture strategies: direct coupling of NAD(P)+. Anal. Biochem. 335:316–325.

16.

Wang, Z., M. Etienne, F. Quilès, G.-W. Kohring, and A. Walcarius. 2012. Durable
cofactor immobilization in sol-gel bio-composite thin films for reagentless biosensors and
bioreactors using dehydrogenases. Biosens. Bioelectron. 32:111–117.

17.

Weckbecker, A., H. Groger, and W. Hummel. 2010. Regeneration of nicotinamide
coenzymes: principles and applications for the synthesis of chiral compounds. Adv.
Biochem. Engin./Biotechnol. 120:195–242.

18.

Wichmann, R., and D. Vasic-Racki. 2005. Cofactor regeneration at the lab scale. Adv.
Biochem. Engin./Biotechnol. 92:225–260.

19.

Zhang, M., C. Mullens, and W. Gorski. 2007. Co-immobilization of dehydrogenases
and their cofactors in electrochemical biosensors. Anal. chem. 79:2446–2450.

136